Happy Friday, Moz fans! In today’s Whiteboard Friday episode, Tom digs into his research on Apple’s moves in search, specifically their recent launch of what he believes is a search engine, how it works, and how they could possibly hope to compete with Google in the future.
Click on the whiteboard image above to open a high resolution version in a new tab!
Video Transcription
Howdy, Moz fans. We’re here today to talk about an Apple search engine. There have been rumors circulating for a couple of years now that Apple are building a search engine. I’m here to convince you I think they are building a search engine and I think they’ve already launched it.
“Apple won’t be building a search engine”
But before we get into what does that look like, let’s start off by talking about the common reasons people say, “Apple won’t be building a search engine.” The first is that we know that Google pay Apple something in the order of $18 billion to $20 billion dollars a year to be the default search engine on Apple devices. That’s a lot of cash.
But if you consider Apple makes more than a billion dollars a day, probably not such a big idea as we think it is. But more importantly, in 2020, the Department of Justice in the U.S. said that they were going to be suing Google for monopoly practices. One of the four bullet points they listed as their reasons was that they don’t want Apple being paid by Google to be the default search engine. They believe that’s a monopoly practice.
They want it to stop. So even if Apple do care about the cash, it might be going away anyway. So then it was not necessarily a surprise, in 2018, we saw the head of search at Google leave and go to Apple. Apple hired the head of search. That’s a very suspicious thing to do if you’re not building a search engine. We also saw at the same time they started listing job listings for hundreds and hundreds of search engineers, which all adds up to building something search related.
Apple has launched a search engine
So rumors circulated, oh, Apple are going to build a search engine. It flew under the radar a little bit, but Apple already launched a web search engine. In iOS 14, we’re on iOS 16 now, iOS 14 launched September 2020, Apple made a change where if you do a search in the iOS default search screen, the Siri suggested websites used to be powered by Google’s web index, iOS 14 changed it to Apple’s own web index.
They’ve already got a web index, and they’ve got a page on their website where they tell us about that web index. On that page, they describe how their Applebot works. They’ve got a web crawler that for the last three years or so can already render JavaScript, something that took Google 20 years or so before they had that. So they’ve got a mature crawler. On that page, they also outline their search ranking factors, their top five search ranking factors they have.
It’s things like links, web page design characteristics, which sounds like Core Web Vitals and PageRank. So the search ranking factors look very similar to Google. They’ve got a crawler very similar to Google. On their page, they talk about how if you’ve got no Applebot specific rules in your robots.txt file, they’ll follow Googlebot rules. So you can see the direction of travel.
They’re trying to crawl the web in a similar shape to Google. But I think that’s a distraction. I think there are several things that they could do differently if they wanted to build a bigger search engine, and we’re going to talk about the three differences I think they could have that would allow them to build a search engine in a very novel way that would compete with Google in a way that we haven’t necessarily fully understood.
Federated search
So let’s talk about that. So the first thing is federated search. Consider how you do a search on Google. You type in “Madrid,” and it goes and it looks in Google’s one giant database. That database has a web index and a knowledge index and all of that stuff. But it’s essentially looking in Google’s own singular database. With Apple, it’s tempting to think, okay, well, they’ve got a much smaller web index than Google.
They can just go and look in that. But what I actually think is they’ll use a federated search approach. Federated search is when you delegate a search to multiple different providers and then aggregate the results. In Apple’s case, I think they can delegate the search to all the apps on your device. Basically, you do a search for “Madrid,” and rather than just looking in the web index, it might say, well, this Madrid search, the intent might be to book a hotel, so I’m going to look in the Hotels.com app.
The intent might be to book a flight, so I’m going to look in the Skyscanner or the Lufthansa app. It might be learn a language, Duolingo. It might get a travel guide, so you look at the Lonely Planet app. The point being that they would ask all of these devices and then blend the results from multiple different sources together. But how would they know what apps are on your device and cater to what intents?
Mobile native
In 2020, they rolled out a change to the operating system that allowed apps being installed on your device to register a list of intents that they can cater to, which is very interesting. It seems like 2020 is the year where they started launching a lot of these things and setting up for this potential search engine. So you would get a search that went to their web index, but also searched through the apps on your device and blended those results together.
As further evidence that they will blend these results, Apple managed to reverse engineer the API they use for their web index, and when you do searches, you get one list of results. You don’t get web results and knowledge results. You get one big list of results, sorted by relevancy, and it might be news, web, web, maps, whatever .
The point being it shows Apple’s thinking. They’re concerned more with relevancy than where the results came from, and they’re happy to blend them together. Okay, so they’ve got this federated search approach. So let’s talk about how these searches can also have custom UIs. So when they rolled out the change to allow an app to cater to specific intents, that app can also tell the phone, oh, when you’re catering to this specific intent, here’s a snippet of user interface you can use outside of the app to cater to that in the best possible way.
As an example, if you ask Siri to split a check, it understands that the calculator app on your phone can cater to that intent. But it doesn’t send you to the calculator app. Instead it shows a little custom UI snippet right there, in the search results, where it shows you how to split that check. So apps that have been installed on your phone can say, “I can deal with this sort of intent, and this is the best way to present that result to users.”
It goes one step further. At the same time, they also launched App Clips. App Clips are what Apple describe as small parts of an app. What’s important about App Clips is that they get streamed to your device. So you can scan a QR code or click a link and it will go straight into the App Clip. There’s no need to install a full application on your device.
App Clips are designed to do one small part of what a full app would do, but basically frictionless in loading them up. So an example might be imagine you arrive in a city and you’ve not been there before. You’re running late. So you do a search for “scooter,” and rather than just searching the apps on your device, because Apple have got this database of App Clips and they’ve got almost 2.5 million apps in their App Store, they could also serve that intent.
So you search for “scooter,” Apple say, “Yes, here’s an App Clip that can help you with that.” You press the button and you go straight into the app. It directs you to a nearby scooter. You pay for it with Apple Pay. You get on the scooter and off you go. This is a mobile native experience. You didn’t go to any website. You didn’t install an app. Everything happened right inside the search results. So Google have been moving to this mobile first paradigm, but it’s still web first. With Apple, there’s a potential for a mobile native experience. This is something that Google could only dream of. What is also true is that this federated search approach would allow Apple to completely bypass Google’s strengths.
Google’s strengths are the fact that they’ve got this huge web index and they’ve been building that for 25 years or whatever. With a federated search approach, Apple can cater to many of the same intents without needing to replicate Google’s web index and their rankings and all of that magic. So if you’re going to build a search engine, you probably also want to be able to personalize results.
Private personalization
How are Apple going to do that especially considering that they position themselves as the sort of privacy-centric tech company? So first thing, quickly consider the difference between Google Photos and Apple Photos. On Google Photos, they will do machine learning of all of your photos on their servers in the cloud, allowing you to say, “Oh, show me my photos of animals,” for example.
You can also do “show me my photos of animals” with Apple Photos, but all of the machine learning happens on your device. It shows the two different paradigms. So when you do a search, Google will serve you some personalized search results based on your history or whatever it might be, but it’s all happening in the cloud. What I think Apple will do is serve you a list of potential results from their web index, and then User A and User B would see a different subset of those chosen by their phone using the machine learning on the device, which understands your preferences, your history, what apps you’ve got installed, etc.
The final thing is that, in terms of personalization, Apple have a potential advantage here too because with the federated search approach, Apple can do what I call fully authenticated personalization. If you do a recipe search for a schnitzel recipe on Google, you’ll get some personalized results, but it will be from the web index. If you do a schnitzel recipe search on your device, this federated search approach will allow Apple to look into your “private” databases.
If you’ve got the Recipe Keeper app, it will see, oh, yes, you’ve got a schnitzel recipe in the app. I’m going to pull that out. It might see that your mom sent you a message on Facebook about the best way to make schnitzel, so it will pull that out. This is stuff that Google can’t do. Google can’t look behind the curtain into your “private” databases. Over the last few years, I think Apple have been playing a slow game, laying the pieces to move towards this mobile native sort of experience.
So when I first started talking about this research in mid-2022, at the SearchLove Philadelphia conference, I predicted that Apple would move the search field onto the iOS home screen in order to try to change user behavior and push them towards doing this mobile native sort of search rather than having users go into a web browser to do a search.
Then a couple of weeks later, Apple announced they were doing exactly that, and that rolled out in iOS 16. September 2022 Apple made it so that the search field is right there on your home screen now, trying to push people towards doing this sort of search. What that means for us as SEOs is yet to be seen, whether they move in this sort of direction, but all the pieces seem to line up.
If you’ve got an app, this is definitely something that you should be thinking about. Otherwise, we should all be thinking about what this means for us over the next couple of years as Apple try to shift the user behavior. What that means is still to be seen. Thank you so much.
Technology can be a conductor or a barrier. Everything we do to market local businesses is meant to culminate in a human encounter. When we get our part right (and external forces smile upon us), technology connects us. When we get our part wrong (or external forces impede us), technology can have the frustrating effect of sundering local brands from their customers, with everybody losing out on the deal.
The modern phenomenon of local search exemplifies the concept of a “mixed blessing”. Loss of control over significant parts of the customer journey can be a source of legitimate stress for owners and marketers. Stress isn’t good for us, of course, and that’s why I’m hoping this message brings some welcome relief: control of the most important aspects of the consumer journeys remains strongly on your side, and you can thrive without the parts you have to give up. We’ve got data to back this up, thanks to Moz’s recent report, The Impact of Local Business Reviews on Consumer Behavior, and I’m hoping today’s column will lift some burdens that may have been weighing you down.
The data
Let’s start out by taking a moment to really reflect on what it means that 96% of adults read local business reviews. Basically almost everyone in your community is perusing this content, making it the widest possible road to your front door, but the truth is that it exists in a space you only partially control. Given that only 11% of review-readers trust brand messaging over public opinion, reviews matter greatly, and it’s a tough reality that they mainly happen in digital spaces you rent rather than own.
If something goes wrong with your reviews on third party platforms like Google, Nextdoor, or Yelp, such as a spam attack, or the random disappearance of your reviews due to a bug or update, or a single irate customer shouting half-truths or downright falsehoods through a megaphone amid a small number of reviews, you have limited direct recourse for resolution. Platforms may or may not respond to your pleas for help, and some customers may ignore even your best offers to resolve their complaints – the sense of lost control is not imaginary.
Here is the good news: for 91% of your potential customers, the very next step they take after reading reviews will land them in spaces you own. 51% will head to your website, which you fully control, 27% will visit your place of business, which you also fully control, and 13% will contact you, and it’s you who control your phone and text lines, your email, forms, and live chat. Apart from the 8% that will move from reviews to the profiles you rent on social media platforms, management of customer experiences is almost all on your side and in your house.
Barring mishaps like your website being infected with malware, a temporary closure of your premises due to illness, or a power outage bringing down your phone lines, it turns out that you remain in charge of key customer/brand experiences during nearly all of the post-review consumer journey. Great news, indeed! But it carries some big responsibilities with it.
Converting on the next step after reviews
The wide funnel begins to narrow as consumers transition from reading reviews to their next steps. Winning maximum conversions from their next actions depends on having the right welcome in place in all three of these spaces:
The local business website
Whether customers click from the review profile to your website homepage, or to a landing page your listing is linked to, prepare this welcome for them:
An accessible, secure, technically-clean, optimized website housing the multi-media content and features the customer needs to take their next steps towards a transaction.
Highly visible information on every way in which the customer can contact and visit you, including phone, text, chat, messaging, email, forms, hours of operation, maps, and written directions.
Additional first-party reviews to provide further proofs of your good reputation and tide you over in times when bugs make your third-party reviews go missing.
A unique selling proposition to seal the deal.
Your place of business
Whether your place of business is your physical premises, or your clients’ locations, you can shine on this main stage with the following:
Exceptional customer service based on the training of your staff and good management of the entire customer service ecosystem. With 65% of review writers saying they’ve written negative reviews because of experiencing bad or rude customer service, building an employee-centric company that radiates both happiness and helpfulness is your best bet for building an excellent reputation.
Careful guardianship of your supply chain. 63% of review writers say they’ve written negative reviews after purchasing bad products. The quality of your inventory supports both repeat purchases and high ratings.
Accurate online local business listings. 52% of survey respondents have written negative reviews after encountering incorrect business information on the Internet. Use of listings management software like Moz Local can ensure that what’s published about your business online (like hours of operation, addresses, and key services) matches what the customer will experience in the real world, preventing inconvenience and disappointment.
Your contact options
Whether a review reader turns next to your phone line, text line, live chat, website form, or email, assist them towards a next conversion by:
Reducing on-hold times on your phone line to the bare minimum
Ensuring all public-facing representatives of the business are well-trained in your products, services and policies
Providing realistic estimates of when a customer will hear back if they are required to leave an email address on chat instead of speaking immediately to a live person
Reducing the number of form fields the customer is required to fill out before reaching you
Offering an after-hours support option
And, of course, for the 8% who will visit your rented spaces on social media platforms as their next step after reading reviews, be sure your full contact information is included on your profiles.
Despite the market disruption of the Internet, so much about local businesses remains the same
While technological innovations are ongoing, it’s apparent that deeply-rooted consumer behaviors continue to follow a traditional pattern that’s existed for hundreds of years. In summary, people in your town want to know what others say about your business >>> people want to connect with your business for a possible transaction >>> people then tell others about what they experienced with your business. All of this cycle has always happened offline, and the only real change is that the means for some of this communication has partly transitioned online.
Just as business owners always had to do without the ability of controlling the word-of-mouth reputation their community was creating for them on front porches and over fences, modern business owners can live without directly controlling the online brand sentiment that exists in spaces they have to rent rather than owning. While it’s true that traditional PR may have had more power to shape public perception before online local business reviews made individual consumer voices so loud, the not-so-secret ingredient to brand longevity and loyalty remains unaltered: great customer experiences at and around the time of service are the foundation of success.
What every local business needs today is a thoughtful plan for managing the digital assets that now contribute to these positive consumer experiences. The winning recipe, then, is developing high standards for the spaces you own (your website, place of business, and most contact methodologies) and being as hands-on as possible in the spaces you rent (the online profiles containing your local business information, reviews, and social content). With a workable strategy and good quality tools for managing this ecosystem, the development of your good name in the community you serve will follow.
Even in the relentless news and noise of early 2023, you’ve probably heard about ChatGPT, the GPT-3-powered (technically, 3.5) chatbot that’s eerily able to simulate human-like responses. Consider, for example, this question and answer (created by GPT-3):
Is ChatGPT a search engine?
No, ChatGPT is not a search engine. It is an artificial intelligence chatbot that uses natural language processing to understand and respond to user input.
That’s not bad, and it sounds like something an actual human would write.
GPT (Generative Pre-training Transformer) is a deep learning system that generates text based on predicting language patterns after being pre-trained with a large set of documents. For this post, I’ll be using OpenAI’s Playground with the “text-davinci-003” GPT-3 model, which can be programmed to produce shorter answers and is more reliable than the ChatGPT Beta.
The current models are genuinely impressive, including parsing complex questions:
How would you explain ChatGPT to a 5 year old?
ChatGPT is like a robot friend who can talk to you and answer your questions. It’s like having a friend who knows a lot about different things and can help you learn new things.
GPT can even mimic human creativity, and make us start questioning our own talents:
Write a slogan for a Mars-themed restaurant run by lizard people, in the style of Edgar Allen Poe.
“Lizards of Mars, Dine Here In Splendor, Where Dreams Unfurl And Palates Render!”
By now, you’ve probably seen dozens of examples, so what does all of this have to do with Google Search and our jobs as search marketers?
Google and the era of natural language
In the early days of search, search engines were built to index words and short phrases and had no ability to process natural language. If you were looking for a new laptop, you might search for something like “laptop” or “buy laptop.” I sometimes joke that, if you walked into Best Buy and just started shouting “laptop!”, you’d probably be escorted out by security. We humans adapted to early search, though, and we dumbed down our language.
As search became more sophisticated, we learned that we could use more complex queries, and, especially with the advent of voice search, we experimented with something closer to natural language. As Google adapted, something interesting happened: our queries became questions, prompting Google to realize that sometimes we don’t want a link to a resource — we want an answer.
Google has evolved a great deal as an “answer engine,” including building out the Knowledge Graph and surfacing web results as direct answers known as Featured Snippets. If, for example, you search for “How does Google search work?”, you might get back a Featured Snippet (essentially, an attributed answer) like this:
This particular example is a little clunky, but it illustrates how the algorithm works. Google is looking for bits and pieces of documents that might answer the question. In this case, it’s obvious that the extracted passage isn’t really a complete answer.
Now, consider the same question posed to GPT-3:
How does Google search work?
Google Search works by crawling the web and indexing websites. When a user searches for something, Google looks through its index and returns the most relevant results. It uses a combination of algorithms and other techniques to determine the relevance of the results. These algorithms take into account factors such as the words used in the query, the freshness of the content, the popularity of the website, and the location of the user.
I can nitpick this answer, but all in all, it’s pretty good, and it’s a much more complete answer than an excerpt from a website. The GPT-3 engine has essentially summarized multiple sources.
It’s easy to see how this could start to make Google nervous.
Trust and the attribution problem
On the surface, GPT-3 (and especially ChatGPT) can return some pretty human-looking responses. Putting aside ethical questions (including plagiarism and bias, which I’ll cover briefly later), there are some serious practical problems.
The Google “answer” has one major advantage: attribution. We can easily go to the source page. This not only provides us more information, but it provides more context, including details about the source itself. Those details can help us trust (or, in some cases, distrust) the answer.
Consider, for example, another answer GPT-3 returned for the question above:
How does Google search work?
Google Search works by using a combination of algorithms and programs to analyze the content of web pages and other online content to determine its relevance to a user’s search query. Google Search algorithms look for a variety of factors, including keyword density, freshness, and the number of other sites that link to the page. Google also uses artificial intelligence and machine learning to improve its search results.
This answer also sounds plausible, except that most SEOs would cringe at the mention of keyword density (underlining added by me), widely considered a myth. It’s entirely possible that this piece of information came from an entirely different source, but without attribution, we can’t validate it.
While this is a complex problem (machine-generated text doesn’t use sources in quite the same way a human dozen, or might use dozens or hundreds of sources), hybrid approaches that reduce the number of sources and provide attribution are possible.
Consider this answer from Neeva — an alternative search engine focused on privacy (hat tip to Greg Sterling) — for “Should I get a tall latte or a grande?”:
While this functionality is in beta and is obviously not operating at Google scale, Neeva attempts to provide primary sources. I’ve purposely chosen a ridiculous example because the stakes are low. Imagine a question squarely in what Google’s calls the YMYL (Your Money, Your Life) realm, such as “Is Bitcoin or Dogecoin better?”:
This is a question that could seriously impact my financial future, and knowing where the answer is coming from makes a big difference and allows me to dig deeper and either establish trust or reject those sources.
Related to trust is a deeper, ethical issue of whether machine-generated text is a form of plagiarism. While traditional plagiarism generally requires intent and copying sizable chunks of text, machine-generated text is still being built off of the backs of writers and content creators, without — at least in the case of ChatGPT — any clear attribution. In Google’s case, overuse of machine-generated text could discourage content creators and cause us to question whether we want to contribute our efforts to someone else’s machine.
Hallucinations and machine dreams
Without careful constraints, machine learning systems are prone to hallucination. In the case of ChatGPT, that essentially means that the system can generate answers that are not rooted in reality. Simply put, ChatGPT doesn’t understand anything it’s saying — it’s creating new patterns based on human language. Consider this nonsense question I posed to GPT-3:
What is Mozart’s 1976 hit “The Fast and the Furious: Tokyo Drift” about?
Mozart’s 1976 hit “The Fast and the Furious: Tokyo Drift” is not related to the 2006 film of the same name. It is an instrumental piece that captures the frenetic energy of Tokyo’s streets at night. The song is built around a catchy, syncopated piano line and a driving, upbeat rhythm.
While intriguing, there is no such instrumental piece. I’ve asked a nonsensical question and GPT-3 has obliged me with a nonsensical answer (Nonsense In, Nonsense Out).
I should point out that ChatGPT already has some constraints built in that address problems like this, and complex models (including OpenAI’s GPT-3 interface) allow you to control “randomness” (high randomness = less constraints), but the core problem is challenging.
Again, consider a higher-risk financial question:
What’s more popular: Dogecoin, Bassetcoin, or Norwegianlundehundcoin?
Dogecoin is by far the most popular of the three coins, with a current market capitalization of over $9 billion. Bassetcoin and Norwegianlundehundcoin are much smaller coins with market caps of only a few hundred thousand dollars each.
A market cap of a few hundred thousand dollars each is pretty impressive for two cryptocurrencies that (as of this writing) don’t exist. I’m sorry to say that I started this example with Labradoodlecoin, only to discover that Labradoodlecoin actually exists.
I’m pushing the engine pretty hard to prove a point here, and modern machine-generated text is much less prone to hallucination than previous iterations. That said, any time you combine multiple sources without regard to their veracity or completeness, there’s a real risk that the end result will be plausible-sounding nonsense.
Scale and the real-time internet
This one’s pretty straightforward: What works at beta scale may not work at Google scale. As the late Bill Slawski would point out, just because Google has an idea — or even patents an idea — doesn’t mean that they implement that idea in search (for many reasons).
Another issue is the sheer speed of the internet. ChatGPT is trained on a static corpus — a moment in time. Google crawls and indexes the internet very quickly and can return information that is recent, localized, and even personalized.
It’s worth noting that Google has invested massive amounts of money into machine learning. Google’s LaMDA (Language Model for Dialogue Applications) is capable of generating complex, human-like text. Google is well aware of the limitations and costs of these models. If they’ve moved slowly in deploying them across search, there are probably good reasons.
While the topic of bias is far beyond the scope of this article, scale also contributes to bias issues. Once you move on from a static, controlled corpus and open up machine learning models to the entire world of real-time content, human bias creeps in quickly (including racism, sexism, homophobia, and other destructive biases). At Google scale, reducing bias is a problem that requires a lot of human intervention and resources.
Code red and future threats
None of this is to suggest that Google doesn’t take GPT (including ChatGPT) seriously. Google reportedly issued a “code red” after ChatGPT went viral. While Google is understandably quiet about their internal reactions, it is almost certain that recent developments are accelerating discussions and developments in Mountain View.
Another clear threat is that Microsoft is an investor (to the tune of $1 billion) in OpenAI (and, with it, ChatGPT), and rumors are circulating that they may invest billions more. Microsoft has also announced that they will start integrating ChatGPT into Bing search before the end of March, although it’s unclear what from that initial rollout will take.
One of the biggest long-term impacts of models like ChatGPT and LaMDA will be in conversational search (including, but not limited to, voice search). Consider the following “conversation” I had with ChatGPT (edited for length):
What are the top five three-row SUVs?
It’s difficult to say which are the top five three-row SUVs as it can be subjective and depend on factors such as personal preference, budget, and location. However, some three-row SUVs that are often considered to be among the best in their class include:
Honda Pilot
Toyota Highlander
Kia Telluride
Hyundai Palisade
Subaru Ascent
Which of them is the least expensive?
Out of the five SUVs that I mentioned, the Subaru Ascent tends to be the least expensive with a starting price around $32,295. The other SUVs that I mentioned have a starting price around $32,000 – $38,000. However, it’s important to note that the final price can vary depending on the specific trim level and options that are selected.
Is the Subaru generally considered reliable?
Subaru is generally considered to be a reliable automaker, with many of their vehicles receiving high marks for dependability. The Subaru Ascent, in particular, has a reputation for being a reliable vehicle with a comfortable interior and good fuel economy. According to Consumer Reports, the Ascent gets an above-average reliability score.
Sorry, ChatGPT can definitely get chatty (and this is the edited version). Note the follow-up questions and how I didn’t need to repeat information or provide context. ChatGPT understood that I wanted to build on previous answers.
While this model won’t serve all of the use cases of search, there are definitely situations where the ability to easily and conversationally refine a query could be revolutionary, especially for complex questions, and, yes, complex purchases. Imagine Google being able to serve different ads at each step on this journey toward a purchase.
Unfortunately, the biggest short-term threat to Google is that people and companies will likely use ChatGPT to churn out mountains of low-quality content, costing Google time and money and likely resulting in major, reactive algorithm updates. This will also be a serious headache for search marketers, who will have to react to those updates.
What’s certain for 2023 is that the popularity of ChatGPT and its accessibility to the general public is going to cause an explosion of investment (for better or worse) and accelerate development. While Google isn’t going anywhere, we can expect the landscape of search to change in unexpected (and occasionally unwanted) ways in the next year.
Keyword cannibalization occurs when a website has too many similar keywords spread throughout the various pages on that site. This can harm the SEO potential of the pages involved, and can quite often go unnoticed for some time. When more than one page has the same, or a similar keyword target, those pages will then start to compete with each other, also creating confusion in the eyes of the search engine, resulting in a struggle to decide which page to rank for what term.
Remember: a single URL can rank for multiple keywords, which is 100% fine. The problem of keyword cannibalization occurs when two or more separateURLs are all competing for the same keyword.
To ensure that a website’s best content is ranking well, you should identify which of your pages are targeting the same keywords. You can then put together a strategy for monitoring and fixing these issues on a regular basis.
In this Daily SEO Fix edition, we take you through what keyword cannibalization is, how to investigate it on your own website, and how to solve any potential issues your site may have.
Recorded at SearchLove London in October — in the wake of three Google updates — Tom presents a different take on core updates in this Whiteboard Friday.
Click on the whiteboard image above to open a high resolution version in a new tab!
Video Transcription
Happy Friday, Moz fans. So I’m here at SearchLove London recording this Whiteboard Friday. I don’t know when it will reach you, but this is a bit of a different take on how to think about core updates. So obviously, I’m filming this in October. We’ve just had three updates back-to-back in quick succession.
I think it’s quite interesting that we had three updates and they were described in very different ways. So we had a helpful content update, a core update, and a product update or a product review update. It’s interesting that sometimes Google talks about updates very specifically. So I think the best examples are things like HTTPS, or Core Web Vitals, page experience update, where they’re very concrete about what they’re going to do, how they’re going to do it, how they’re going to measure it, and how it’s going to impact the algorithm.
Then you have core updates, where they do say things, but they tend to be kind of saying the same thing every time. So every single core update, they’ve said make good content, work on your expertise, authoritativeness, trust. This isn’t very concrete. It’s not very specific about what they’ve changed this particular time.
Indeed, if you’re a site that’s affected by these updates, it can feel quite random. It can feel like you’re just going upwards or downwards. There’s no particular rhyme or reason. So how can that be? So I want to give you some different ways to think about that.
A refresh
So the two different ways that I’d like to focus on, one of them is this concept of a refresh. So Google used to talk a lot about algorithm refreshers. This is up till about 2012. What they meant was this is something that was different to an algorithm update. So it wasn’t called an update. It was called a refresh, as distinct. They were trying to say that this would kind of be like a mini reset of how the algorithm was thinking about certain things.
If you look at how they talk about core updates in their documentation, they say things like this, “So your site might not recover until the next core update.” So you’d have situations where this is your rankings in blue. This is your competitor’s rankings in red. You get to a point where they’ve improved their site over time gradually. They’ve not been recognized for it, and then a core update comes along and suddenly they go up. So your position, you go down, and you’re left thinking, “Oh, that was a little bit random.” But, of course, it wasn’t random. It was just that they were gradually being recognized for things they’ve worked on or suddenly being recognized for things they’ve worked on gradually over time.
Testing
The other concept I’d like to talk about is the extent to which Google is testing. They’re iteratively testing over time. Again, they talk about this in their own documentation. There’s an article that I’ve come back to quite a few times, back in 2018, it’ll probably be linked beneath, where they invited some journalists to a meeting of their search engineering team. In that meeting, they were talking about how they were thinking about some changes they were making to the SERPs, and they talked a lot about how they were going to run some things as a test and look at certain metrics, see how they were improved. So it’s important to think that Google has their own metrics that they’re iterating towards, and they’re not necessarily saying, “Oh, your site is bad,” or, “There’s something wrong with your site.” They might be saying, “Oh, well, what we’re aiming for at this point could be more beneficially affecting some sites than others. Ultimately, if someone comes up, someone else has to go down.”
Indeed, in MozCast data, if we look at sites that were affected by at least four updates, so this is looking since Medic, technically there were some core updates before Medic, but I think the industry has been very focused on this since Medic, if we look at the core updates, of which there have been 12 now, and the sites that were affected by at least 4 of them, the vast majority in MozCast data had both some major positive movements and some major negative movements. So this tiny green slice represents the sites that only saw positive movements, and this red slice, the sites that only saw negative movements. So it’s incredibly unusual to have mono-directional movement, which just shows that people are winning and losing as Google tests different things. It’s not necessarily that some sites are just better suited to core updates and win every time. That’s very, very rare.
Longer term
I also want to talk a little bit about the longer term. I think it’s important, when we think about these updates, to zoom out of it because these short-term effects can seem more random, harder to explain, harder to predict. So I’ve looked at a lot of sites again in the MozCast data over time and how they’ve been impacted by each update.
So this is an example, and obviously, it’s drawn on a whiteboard, so it’s not super precise. But this example I’ve attempted to illustrate here is actually Reuters, the news organization. I’ve chosen them because this is a site that obviously produces a lot of original content. It’s very authoritative. It’s hard to criticize it from the regards that Google likes to talk about in its core updates discussions and announcement. These bars represent how it was affected by each core update over a period of time. So it had some big negative hits, not many serious gains from these updates. So this doesn’t look very good. But if you track how their traffic grew or their visibility grew within MozCast over time, it looks a little bit different. So it sort of gradually grows over time. So what this means is even though on the days of the specific updates they were taking sometimes negative hits, like if we look at the week before and the week afterwards, sometimes they took big hits. Obviously, there’s long periods of time between these updates, so they might still be able to grow.
So say there might be three months between these bars, and even though they took a big hit here, they’re growing over the next three months. Maybe over here, they take this big hit, but they’ve more than recovered it by the time they get to the next update, take this hit, more than recovered it by the time they get to the next one. That could be that their SEO team is working some magic behind the scenes, but this is quite a consistent trend. This happens to a lot of sites. What I would suspect is actually happening is Google, when they launched the core update, they’re, to some degree, resetting certain things, looking at things afresh, valuing different metrics. Then over time, whatever historically was making that site perform will creep back in and start to be considered again.
So I hope that was interesting. That’s just a few different ways to think about core updates besides the usual messaging that we get, which is very consistently just E-A-T, good content. I’m not saying you shouldn’t do those things. Those are important. This longer-term trend that you get with a lot of sites that do do those things shows how important it is. But I think when you look at individual updates, you have to keep in mind that it’s not necessarily that Google is suddenly optimizing for these things more. They’re just iterating over time. That’s all from me. Thanks.
Ready for a deep dive into 40+ common features found on Google Business Profiles (GBPs) and the New Merchant Experience (NMX)? Here we go!
You may feel daunted by the fact that GBPs are major assets you rent from rather than owning because Google has always taken a crowdsourced approach to vital information that represents your company. You may also be disgruntled right now because you had gotten used to managing your listings in the historic Google Business Profile Manager dashboard, like this:
And now Google has ousted you from the familiar dashboard and tossed you into the chilly, unfamiliar pool of in-SERP listings management which we’re calling the New Merchant Experience and which looks like this:
You might rightly remark that major change was the last thing you needed right now as a local business owner or marketer. But, deep down, you know you’re going to have to get the hang of this sometime, and this column is here to help you with that. After all, with estimates as high as 70% – 80% of local business leads happening right within Google’s interface, you can’t ignore GBP. Instead, you’ve got to learn to control every bit of it that you can. This article will take you through the most common features of these listings, explain what you can and can’t control on them, and teach you how to manage controllable features in the NMX.
If, after reading this cheat sheet, you find you just can’t stand the clutter of trying to manage your listings via the NMX amid the crowded organic SERPs, remember that listings management software like Moz Local still offers the niceties of a clean, organized dashboard for those who require a quieter workspace.
Core features on most Google Business Profiles and NMX interfaces
There is significant variation of Google Business Profile and NMX features based on business industry and category. For example, hotel listings have elements that aren’t included in listings for plumbers or grocery stores. Google is continually experimenting with new features while demoting others. This section will cover some of the commonest elements available to most businesses.
Before we look at our graphic and key, it will help you to know that you can access the New Merchant Experience by searching within Google’s organic SERPs for the name of your business, or the name of your business + its city, or for the phrase “my business” while logged into your Google account.
1. See Photos
When clicked on, this takes the user to both owner and user-generated photos in a set. Photos significantly impact CTR. Photos must be monitored for spam. You can also upload videos to your photos section, but try to post more than 2 videos so that you’ll get a separate mobile video subtab. It’s also important to know that previous versions of Google’s local product offered an analytical feature called “Photo Insights” but this no longer exists in the NMX.
Controllable?
Partly — photos are both an owner and crowdsourced element.
Where to control?
By clicking on the “Add Photo” link in the listing or the NMX, you will be taken to the following wizard for uploading different types of photos.
If you wish to view all your photos for the purpose of identifying user-uploaded spam, click on the image, itself, in the listing, rather than the “add photos” link, and you will be taken to the interface where you can both see and flag photos.
2. Maps
When clicked on, this takes the user to the Maps-based listing accompanied by map with pin. Be sure your map marker is correctly placed.
Controllable?
Partly — owner can correct misplaced map marker, but users can submit placement edits, too.
Where to control?
If your map pin is in the wrong place, click on the “Edit Profile” link in the new interface, then click the “Location” tab at the top of the “Business Information” pop-up, then click on the map.
3. See Outside
When clicked on, this takes the public to an interactive Google Street View visual of the business. When you are logged in, clicking on the feature takes you to this interface for uploading your own exterior imagery:
Controllable?
Partly — the owner can upload their own exterior shots, but cannot control StreetView. You can flag issues.
Where to control?
Click on the “see outside” section of the Google Business Profile to upload photos.
4. Business name
This must reflect the real-world name of the business and be formatted according to Google’s guidelines. If you’re marketing a service area business like a plumbing franchise without storefronts, your name should match what appears on your website.
Controllable?
Yes — the business owner provides, though the public can suggest edits to this feature.
Where to control?
In the NMX, click the Edit Profile link and mouse over your name in this popup to edit it in the About section:
5. Website
This link takes you to whichever page of your website you’ve designated as the Google Business Profile landing page.
Controllable?
Yes — owner provides, though the public can edit.
Where to control?
In the NMX, click the Edit Profile link and mouse over your name in this popup to edit it in the Location section:
6. Directions
This link takes you to the maps-based driving directions from the user’s location to the place of business.
Controllable?
No. This experience is created by Google. However, see earlier advice if you need to move your map pin.
7. Save
When a user is logged into their Google account, clicking the “Save” button brings up the above pop-up, enabling them to store the business in their favorites or other areas like places they want to go or are planning to travel to.
Controllable?
No — Businesses do not control whether customers utilize the “Save” function.
8. Call
On mobile devices, clicking the “call” button automatically triggers a phone call to the business. On desktop, it brings up this popup asking the user to pick an app with which to start the call.
Controllable?
No — beyond ensuring that your phone number is accurate, this is not an area controlled by the business.
9. Star Rating
Google uses a simple mathematical average to calculate the overall rating of a business on the basis of all the star ratings reviewers have given the business. The star rating is considered the #1 review factor that customers use to differentiate one local business from another.
Controllable?
No — businesses cannot directly control or alter the star ratings they receive. Providing excellent customer experiences is the best way to earn the kind of reputation that results in a high Google Business Profile star rating.
10. Review Count
This is a simple count of the total number of Google-based reviews the business has received. Surveys indicate that consumers can become suspicious when any local business has either too few or too many reviews compared to its competitors. When clicked on, the review count takes the user to the main review overlay, which we will cover in full below.
Controllable?
No — the business cannot directly alter the count of its reviews, but can implement a review acquisition program to increase its total number of reviews over time.
11. Editorial Summary
This is one of the more diverse elements of the GBP. In our initial example, a common format is shown of Google taking the primary category and combining it with location information, like this, near the top of the listing:
However, other iterations may also be considered editorial summaries, including this example which cites a third party like Wikipedia and may appear further down in the listing:
And there is at least one other version that may appear within the reviews section of the listing, specifically cite Google as the originator of the summary, like this:
The nomenclature for these elements is loose, and Google may well have unique internal names for all three instances.
Controllable?
Partly — in the case where the primary category is included in the summary along with location information, the business is contributing the information that yields that summary, and in the case of well-known entities that have control of a Wikipedia page, there would be some editorial control built into the source from which Google is deriving the description, but overall, it is Google who generates these summaries.
12. Further Directional Cues
This is a feature which only appears on some Google Business Profiles. In this example, the user is being given information about the walking distance to nearby major local attractions with embedded links to those venues.
Controllable?
No, but do be sure your map pin is correctly located so that Google can easily identify whether your business is located near to local attractions.
13. Address
For brick-and-mortar businesses, this line must display a genuine, physical address that complies with Google’s extensive guidelines. For service area businesses that don’t serve customers at the company’s premises, Google wants the address to be hidden, and instead of showing a street address, an “area served” section like this will typically be shown:
From the NMX, click on “Edit Profile” and then the “Location” tab to edit your address. This is also where the toggle is located to hide or display your address.
For service area businesses, the place to edit your information is also located in the “Location” tab of the NMX. You can add up to 20 locations describing your service area, but its overall boundaries shouldn’t extend beyond about 2 hours of driving time.
14. Hours of operation
One of the key ways in which Google Business Profiles provide customer service is via the presence of accurate hours of operation. Be sure yours are kept up-to-date at all times, so no customer is inconvenienced by arriving at your premises to be greeted by a “closed” sign.
Controllable?
Yes — businesses add their hours, though the public can edit them. Businesses can also set special hours for holidays.
Where to control?
From the NMX, click on “Edit Profile” and then the “hours” tab to edit your information. You will also find a feature there for adding further special hours:
15. Phone
You can add a primary phone numbers and additional phone numbers to your listing. Google prefers that you use a local phone number rather than a call center helpline whenever possible, and they want your listed number to be in the direct control of your business. Your number can be either a landline or cell phone number, and it’s a best practice to be sure that whatever number you list is answered by staff who say the name of the business when they pick up. It’s not a good practice to have more than one business share the same phone number. For example, if you own both a Christmas tree farm and a woodworking studio on the same property and are listing each business separately, each business should have it’s own phone number. Premium-rate telephone numbers that charge callers high rates aren’t allowed.
Controllable? Yes — the business lists their numbers, but the public can suggest edits to them.
Where to control?
In the NMX, click “Edit Profile” and then “Contact” to see and edit your phone numbers. Related to this, the NMX also has a button labeled “Call”. When clicked on, this button will ask you if you want to enable Google’s call history feature, which you can read more about here:
16. Further Location Information
In our example, Google is providing additional location information. Some listings will have a province or county listed in this area. Others will have a different special feature called “located in” that can appear when a business or department is contained within another entity, like the parts department inside an automotive dealership:
Controllable?
Partly — the information about your province or county relates to the physical address you provide, but the separate “located in” feature can only be edited via Google Maps. If you wish to edit this feature, find your listing on Google Maps, click the “suggest an edit” button, and look for the “located within” area of the editor:
17. Suggest an edit
This function is available to any member of the public who wishes to suggest an edit to the listing. For example, an individual can let Google know they think your address, or hours of operation, or phone number is wrong.
Controllable?
No. Unfortunately, the suggest an edit feature is a cause of confusion and stress for many local business owners who have no control over whether or not members of the public suggest edits to listings. The difficulty of the scenario is summed up well by this Diamond Product Expert in Google’s help forum:
18. Own this business?
If you see this link on a listing, you either haven’t claimed the listing in question or are not logged into the account you used to claim the listing. Claiming a listing is the terminology for verifying with Google that you are authorized to take control of the management of a listing for a specific business. If, however, you are logged in properly and have already claimed the listing, you should see this link, instead:
Controllable?
Yes — you can claim listings you have the authority to represent and once you have claimed them, you can use the edit link to be taken to the NMX to edit your listing at any time.
19. Products
You can greatly enhance your listing by adding products to it representing key offerings in your inventory. These can be physical products, like retail goods, digital products, like software, or intangible products, like home services. You can add a price, description, buttons, and links for more information.
Controllable?
Yes — you can add products.
Where to control?
In the NMX, click “Edit Products” then click, “Add product” to be taken to the product editor.
20. Questions & Answers
Businesses are permitted to use this feature to upload questions they commonly receive and answer them. The public can also ask questions and answer the queries others have submitted. It’s important to continuously monitor this area of the GBP to be sure that questions from the public are receiving timely, authoritative answers from your brand, rather than random answers from the public which may or may not accurately reflect the business.
Controllable?
Partly — you can ask and answer questions but the public can also both ask and answer.
Where to control?
To add and answer questions, click the “Q&A” link in the NMX and you’ll be taken to this interface where you can either ask or answer questions.
21. Reviews
Reviews and owner responses are some of the key features that turn a static local business listing into an interactive two-way medium. For deep insight into the best strategies for review management, read The Impact of Local Business Reviews on Consumer Behavior | SEO Industry Report, but in brief, reviews are the main online seat of local business reputation, they influence local search rankings, and ratings, review text, and owner responses all have significant impacts on customer journeys. To avoid reputation damage and litigation, all review management practices must align with platform guidelines, like Google’s content guidelines, as well as the laws, like FTC guidelines, governing your country.
Controllable?
Partly — you can request reviews and respond to reviews, but you cannot control whether customers leave you reviews and you are not permitted to review your own business or incentives others to do so. Fraudulent reviews can be reported to Google in hopes of removal.
Where to control?
The NMX has several different menu items related to reviews, including “read reviews” and “ask for reviews”. Clicking on the former will take you to the entire review overlay, with “reply” functionality so that you can respond to each review or rating you receive:
Clicking on the latter will take you to a popup that generates a handy link to your review profile, with prompts to share that link via vehicles like email, WhatsApp or Facebook:
22. Updates (formerly known as Google Posts)
This feature is like a microblogging function located directly within your Google Business profile. It enables you to post new content to your listing on a regular business.
Controllable?
Yes, most business types (with some exceptions) can write updates.
Where to control?
Click the “add update” link in the NMX and you’ll be shown three types of posts you can write: updates, offers, and events. The first lets you add an image, a description, and a linked button and will stay live for six months before being archived. The second includes all these fields, plus a date range setting for the period in which your offer is live, and additional informational fields to further explain the terms of your offer. The third type includes fields for information like when your event starts and ends. Some listings may also feature discontinued post types.
23. People Also Search For
This is one of those areas that makes it clear that local business rent space from Google rather than owning it. Chances are, if you owned your Google Business Profile, you wouldn’t feature your competitors on it, but this is just what Google does.
Controllable?
No — you have no control over which competitors Google shows in this section of your listing.
Additional GBP and NMX Features
Some business categories have unique Google Business Profile features or elements we haven’t yet covered. The New Merchant Experience also includes some other functions we’ve yet to explore. Let’s dive back in!
Categories
Categories may be the most important feature that you control when it comes to the search phrases you want to rank for, and categories also determine many of the fields you’ll have in your Google Business profile. Your primary category may appear on its own near the top of your GBP, but it also may appear as part of your editorial summary. In order to see all of your categories live, you have to head over to Google Maps. There, by using a Chrome extension like GMBSpy, you can see all the categories both you and your competitors are using.
Controllable? Yes — businesses can choose primary and secondary categories.
Where to control? Click on “Edit Profile” in the NMX and edit your categories under the “About” tab.
Attributes
This is an experimental area of GBPs to which Google is always adding so that both businesses and the public can further define the characteristics of the business. Attributes can include information about who owns the business, or amenities, such as whether a building has gender-neutral restrooms, or protocols, such as whether an appointment is required to visit a business, or layout information, such as the presence of wheelchair-accessible entrances.
Controllable? Partly — businesses can add a great many attributes by clicking the “Business Information” tab in the NMX and then going to the “More” section. However, Google also crowdsources attributes from the public.
Description
Local businesses can write a 750 character description of themselves.
Price Summary
Some business listings display a pricing overview of 1-4 dollar signs.
Controllable?
Controllable?
Unknown – Google asks reviewers to evaluate the prices they’ve paid at businesses, as seen above. There has been past debate over whether the prices published by local businesses on their digital menus influence these attributes, but Diamond Product Expert Ben Fisher has confirmed with Google that these elements stem entirely from UGC.
Menu links
Google offers a variety of menu options for businesses. Food and drink businesses have a standard menu that can link to a page on their website or a third-party platform, and menus can be created within Google itself, including menus with photos added to them. Additionally, some businesses can create service menus in Google.
Controllable?
Yes — where relevant, businesses can create or link to menus of goods and services.
Where to control?
With thanks to Damian Rollison, logged in owners of food and drink establishments should see edit menu links in the NMX, as shown above.
Businesses offering a menu of services can edit them under the “Edit Services” link in the NMX, including adding custom services.
Booking
Businesses can feature several different booking and appointment functions on their listings, powered by a provider. For example, this instance on the listing of a dental practice stems from ZocDoc:
Controllable? Yes, relevant businesses can add booking and appointment features.
Where to control?
With thanks to Petra Kraft for the screenshot, applicable businesses should see a booking button in the NMX, as shown above. Businesses wishing to enable booking and appointment features must follow these instructions and can learn more about the Reserve With Google system here. Also pictured here is the “Waitlists” function tied to the Reserve with Google program for managing queues of potential diners.
See what’s in store
The See What’s In Store feature shows up on the Google Business Profiles of companies that have integrated the Pointy device into their point of sales system.
Controllable? Yes, if you are in the US, UK, CA, IE, or AU, have an appropriate point of sales system, and vend items with standard barcodes, you can apply to become part of this program.
Where to control? Google’s documentation states that, depending on how you signed up to be part of their See What’s in Store program, you will either be managing your inventory directly via Pointy or within your Google Business Profile.
Order online
Some GBPs may feature an order online button . When clicked on, it brings up a list comprised of the brand as well as third-party services that facilitate taking digital payments in exchange for goods that can then be picked up or delivered.
Controllable? Partly — business owners can opt into this, but there has been considerable controversy over the years about whether third-party providers in the delivery business are more of a bane or blessing to local brands, and whether or not its right for Google to promote these types of services on GBPs.
Reviews from the web
This features a rating summary and links to relevant third-party review sources, determined by Google.
Controllable?
Partly — owners can’t dictate which 3rd parties Google chooses, but they can work to build up positive reviews on featured sources.
Profiles
Google business profiles can feature icon-based links to major social media sites where the company has established a presence.
Controllable? No — there is no field in the NMX for adding your social media profiles, but best practice advice is to be sure they are featured on your website so that Google can easily associate them with the business.
Popular times
This Google Business Profile information is drawn from users who have opted into Google Location History. It’s meant to help users plan visits. It’s conceivable that this could be utilized as a ranking factor.
Controllable?
No.
Chat/Messaging
This feature enables customers to leave messages for businesses and receive responses.
Controllable? Yes — businesses can choose to opt into this.
Where to control? Google’s branding of this feature is messy. In the NMX it is labeled Messages, but when opting into the program, you will be shown a popup asking if you want to turn on “chat”, and customers will see buttons labeled “chat”. You can read more about this feature here.
Performance
The performance tab in the NMX takes you to the analytics data historically known as Insights. This is where Google reports on how the public is interacting with your listing.
Controllable? No — apart from changing the date ranges in this interface, this is not data you can edit.
Advertise
This link in the NMX takes you to where you can sign up to run paid advertising.
Controllable: Yes — local businesses can opt into Google Ads.
Admission
Some attractions feature admissions information on the Google Business Profile, with price comparisons for tickets.
Controllable? Not directly, though attractions can be sure pricing information is clear on their websites. This is a controversial feature, because Google states the following:
“Tickets are ranked based on several factors, but mainly by price. Official tickets are given preference in the ranking. Ticket suppliers don’t pay Google to appear in these search results and can’t pay to influence their ranking. Prices are based on standard adult entry and may vary depending on the specific ticket types, dates or eligibility for other discounts. Displayed currencies may differ from the currencies used to purchase tickets.”
However, as can be seen in my above example of the GBP of the Japanese Tea Garden in San Francisco, the official website of the attraction is being outranked in the Admissions results by an entity that is not official and is also more expensive, contrary to Google’s statement of how this feature should work.
Experiences
Some Google Business Profiles may included nearby experiences. This screenshot is also from the GBP of the Japanese Tea Garden in San Francisco, and is displaying other things to do in Golden Gate Park.
Controllable? No — this information is provided by Google, not the business owner.
Opening date
It has been some time since I have seen an opening date featured on a GBP, though I frequently spot them in local pack results, but businesses do have the ability to tell Google the month and year in which they were founded.
Controllable? Yes — businesses can provide this information.
Where to control?
In the NMX, click on “Edit Profile” and then click on “Contact” to find the field for adding your founding date.
Lodging listings
Hotel profiles are unique to the point of being almost a different style of Google listing. They feature many distinct elements including price comparison, environmental practices, availability-based booking, and other features.
Controllable?
Mostly — owners have a variety of features they can enable and manage, though Google pulls in many elements from third parties including the public.
Hopefully, this guide will act as a trusty flotation device the next time you need to manage something in your Google Business Profile via the New Merchant Experience. What I’ve learned from writing it is that if you click around long enough in the NMX, you can find most things, with a few exceptions like those that have been noted by friends like Damian Rollison and the good folks over at BrightLocal and Online Ownership.
This column has covered what the main features are and where to control them, but if you’re ready to move on to making the most of each element, read the popular Beginner’s Guide to Google Business Profiles for optimization and marketing tips. If you want to understand where GBPs fit within the larger framework of local search marketing, read The Essential Local SEO Strategy Guide for nine chapters of free, expert advice. Or, if you feel you’d benefit from investing in a more guided experience, consider registering for our video-based Local SEO Certification. Google will just keep changing – that’s a guarantee! Our best strategy, then, is to embrace continuing education. With determination and knowledge, we can all swim these laps for the win!
There was a pre-search-engine age. It’s hard to conceive of now, but there was. Even in the early days of search engines, when Ask Jeeves, Yahoo, and Excite still competed for the crown, I can remember web portals. Pages that I’d start at, in the “computer room” at school, to navigate and explore the web not by searching, but by clicking on organized links.
In the beginning, there were web portals. The internet was without form and void, and darkness was over the face of the deep.
These were already the death throes of a previous internet age. Search engine dominance, and specifically Google dominance, has been the norm for this kind of journey for decades now. It’s all that many SEOs have ever known.
But what comes next?
People have talked for a long time about existential threats to Google’s dominance, and often, implicitly, by extension, SEO. You’ll have heard the claims that Amazon or YouTube are now preferred engines for certain kinds of search, or that Google is going to struggle against the unique technological advantages of Apple, AI chatbots, the unique regional advantages of Baidu, or the unique format advantages of TikTok. Or maybe you’ve even heard that people prefer to restrict their searches only to Reddit. Even mainstream outlets are suggesting that Google search quality may be in decline.
This post is not about the health of Google search as a product, or about the implications of improving AI products for your SEO strategy right now. (Although, I know of at least one post for this blog being written on that topic!) Instead, this post is about which of these threats, if any, actually stand a chance of unseating Google’s dominance.
In what capacity?
To ask what might take Google’s role, we must first ask what role it is that we’re interested in. Google is many things, and possibly part of the reason Google’s doom is so often predicted is that we’re not always talking about the same specific things.
What exactly is it that search engines as a genre, and then Google, have dominated? Perhaps we might mean:
The place you’d start to find a web page on a site you’ve not yet discovered? For example, you might not know yet what the best site is for a given topic.
The place you’d start to find a web page on a site you’re already familiar with? Perhaps you’re searching on Google hoping to see a result from Reddit, or from Wikipedia.
The place you’d start to answer a given question? So maybe you’d be happy with a non-web result as long as it answered your question.
The place you’d start to complete a task? So, again, the best answer might not be a web page at all.
The truth is that the present reality blurs these use cases to the point of it not being useful to separate them. But for Google to be replaced by something that maintains this close alignment, it’d have to be a close peer competitor.
The obvious pretenders
There are two that come to mind, as similarly resourced companies trying similar things via a similar method (a web index): Bing and Apple.
I don’t want to be dismissive of Bing, or of the value of someone — anyone — else maintaining a similar enough competitor to keep Google somewhat honest. Although it’s often mocked in SEO circles, Bing in reality is not so many years behind Google at any given point. But, really, it’s hard to see the events that could lead to Bing supplanting Google at its own game. It’s just too similar for people to make the switch. One possibility based on recent news is for Bing to become less similar, pursuing one of the precise alternatives I’ll cover below – but more on that when we get to it.
Apple, on the other hand, is doing something similar, but with some unique advantages. I must credit my former colleague (and 2023 Mozcon speaker) Tom Anthony who has been very prescient around Apple’s moves in this space, going so far as to backward-engineer Apple search results that weren’t supposed to be publicly available. Apple can do things that Bing can’t, leveraging Apple’s app ecosystem and device integration to provide search results that skip certain steps of a user journey in ways that Google cannot, or will not.
The trouble with Apple as a Google search competitor is obvious, though. The unique advantages, as I said, are to do with apps and hardware. Apple devices are expensive — prohibitively so. (This varies by market – in the US, with the base price of a phone contract being so high, iPhones are more palatable and have a notably higher market share than in Europe, for example. But, that’s a topic for another day – either way…). There is a fairly hard cap on the market share of a search engine that is only superior on high-end devices, and not only that, but ones from a specific brand.
So could Apple take a big chunk out of Google? Yes, it may already quietly have done so with various iOS changes pushing the prevalence of Apple’s own search results. But totally replace Google? Very unlikely.
You can say the same for regional competitors like Baidu, Yandex, or Naver. These may well consistently beat out Google in their own backyards, and perhaps even spread to nearby countries and regions, but it’s hard to see them beating Google in its own backyard(s).
Revolution, not evolution
So what about competitors that replace Google by doing something totally different, to solve the same problems? The reality is that a lot of the problems we solve right now with web search, are not actually well suited to web search. The fact that something like a Google Home will often answer your questions by essentially reading out a featured snippet is a symptom of Google’s dominance, not a symptom of web search being well suited to that use case. Even Google themselves recognize this, and betray that in tools like Google Translate, clocks, calculators, and so on, embedded in SERPs. So who might the more disruptive threats be?
One name that came up a lot in 2022 is TikTok, and I’d point you to this excellent post by Lidia Infante on this very blog. To sum up her argument, TikTok can take market share from Google, but it can’t replace Google entirely. TikTok is too specialized (in video format and certain topic areas), and the quality assurance is too weak. So, again, we have a competitor that chips away at Google without replacing it.
Then of course, most recently, SEOs of Twitter have been right to point out that for many queries, ChatGPT produces better responses than Google. Take this example, “excel query for extraction the domain name from a url”:
The ChatGPT result above is far more informative and easy to follow. However, like TikTok, this only works for certain things. ChatGPT is not a web search engine:
So you have to be willing to abandon the premise that your result should be a web page. Which, in this context, comes down to: do you trust an answer if you don’t know who wrote it? ChatGTP and similar technologies have access to “knowledge” sourced from the web, like Google, but they don’t cite a source. Indeed, it would be immensely difficult to trace the source of their various claims, some of which seem quite… odd.
Similar to TikTok, then, this is something I might prefer to Google for a specific kind of query. In this particular case, the kind of query that previously took me to StackOverflow. But I’m not going to ask it for mortgage advice.
I noted above that Bing is rumored to be integrating ChatGPT with its own search product. This enlarges the threat to Google in that it makes this technology more accessible, but really, the same qualms apply – there are many, many queries for which this is not helpful. Even if Bing can hybridize these technologies into a “best of both” of traditional web search and NLP, well – that’s already the road Google is going down.
The other challenge with this “ChatAI as search” model is an economic one. Google and Amazon have both already come to the conclusion that the type of queries asked of their personal assistant devices are barely, if at all, economic to run – because of the limited monetization opportunities for purely informational queries. Perhaps my distinction above, about what we mean by replacing Google, is very relevant here – some of our use cases of Google as a search engine are actually just a loss leader for others. As such, perhaps this bundling of disparate uses is necessary.
The King is dea… wait, wait, he’s still breathing
Number of explicit core search queries powered by search engines in the United States as of January 2022 – via Statista
Ultimately, these threats look set to chip away at Google, not replace it. At worst, a broad monopoly will be sliced up and shrunk, and that doesn’t feel like any great evil. For SEOs, we should be aware of these new search engines, and these new “search engines”, and of the risks attached to being locked into the Google ecosystem. But don’t forget the chart above: the original pie is not going anywhere. The Google SEO game is still not a bad game to be playing.
Business blogging is an interesting concept. Many businesses have a blog, and yet, in most cases, there isn’t much planning going into how those blogs fit into larger business goals. Someone on the team is just assigned to publish content regularly, and that’s what they do.
A business blog is usually an island within a company: There’s likely, not much measuring, so no one has any idea if it is generating any traffic, sales, or rankings.
And yet, despite a lack of planning or measuring, there may be some good traffic coming.
Blogging is a good way to achieve “aссidental rankings” (those you didn’t necessarily plan for) and discover more search-driven sales channels. But many business owners or marketing managers have no idea what their blogs are ranking for. Consequently, there’s no buying journey set up for someone who lands on a blog post.
Clicks may come, but a blog remains an island. So, how can you include your blog in your overall marketing strategy to convert those clicks?
Step 1: Identify your blog’s current rankings that may bring sales
In other words, let’s see if there’s anything to worry about in the first place. Normally, the good old 20/80 rule works for blogs as well: About 20% of your content will be bringing 80% of your traffic. So, you need to know what 20% of your blog content actually does bring traffic.
Navigate Performance > Create a filter to see all the pages that contain “blog” (or whatever domain or subdomain you have your blog at) and click on the “Pages” tab.
The list is going to be sorted by the number of clicks your blog is driving.
Don’t forget to use Moz Pro’s rank tracker to set up position monitoring of those pages that are driving organic traffic.
Step 2: Plan your buying journeys from your blog
Make sure your blog makes it clear that this is a business blog and you are selling something. Define the major conversion funnels for your readers to follow.
At the very least, there should be a site-wide call to action (CTA) inviting readers to check your product or products, but the best idea is to create contextual CTAs and lead generation forms, which could better align with the searchers’ intent.
Hubspot is a great example of contextual CTAs that differ in each article and engage readers by providing “upgrades” to the article content:
There are many more content-based B2B lead generation strategies, but contextual CTAs work in e-commerce and retail as well.
HomeDepot is another prime example of including contextual CTAs (and automating them). They always show related products right next to each of their guides. Not only are these products actually helpful for following the tutorial and solving the searcher’s problem, they also create the right expectation that this content has a business behind it:
Look at your blog and open any individual article. Do you offer readers ways from that blog to follow the sales funnel? You may also want to set up a heat-map and button click tracking to better understand what your blog readers interact with on your blog post.
Step 3: Revive lost rankings
An important reminder: Blogs inevitably lose rankings.
The essence of any blog is that the more content you add, the further your older content is moving down the archive, away from the home page. And gradually, but inevitably, that content is going to lose rankings.
Thus, checking your blog’s traffic and rankings should be done regularly because blogs tend to lose rankings over time. You can do that using the same tools. Search Console has a “Compare” option allowing you to see pages losing clicks…
Click the “Click difference” header twice to sort by pages that lost the most clicks. You can only go as far as 16 months back, though.
And again, once you set up tracking, the Moz Pro rank tracker will tell you exactly which keywords are losing rankings. You can compare your current positions to what you had when we first started recording any particular keyword.
So now that you have a list of articles that are losing rankings, what to do with those?
Is this something that can be saved?
First of all, figure out if this is something worth saving. There will always be rankings that you can just let go of. These include old news or press releases, articles that stand no chance of converting readers into buyers, expired product reviews or listicles, etc.
Update the content
If it looks like a page that is ranked for something with transactional potential, think about whether you should update that content to align it better to those keywords as well as relevant contextual CTAs. Maybe there’s a new video you created after writing that article, so now you can add it there.
Also, study your organic competitors to see where they might be better. You can use IMN’s Side-by-Side Tool to see the content of pages that gained rankings that you lost. The tool will compare various important on-page elements like titles, subheadings, keyword usage, etc. This will give you lots of clues about what needs to be expanded, added to, and improved.
SE Ranking on-page SEO checker is another great tool to see where your content needs to be updated and optimized better based on your competitor research:
Using SEO change monitoring, you can also be on top of your competitors’ content updates so that you can timely update yours and avoid losing rankings.
WebCEO is another cool tool that analyzes around 30 competing URLs for your target search queries and suggests what needs to be added for your content to catch up with your competitors:
Next, republish your refreshed content with a new date to push that article back to the top of your blog’s archive. For seasonal content, it’s especially important to time those content updates with the upcoming holiday or season.
Depending on those keywords (i.e. their relevancy and search volume), you may want to make that page more evergreen and prevent it from ever getting lost in your blog’s archives. Internal linking is one way to do that.
You can link to that page from relevant static landing pages that always remain at the same level in the site structure. Or include static site-wide or blog-wide “Featured guides” sections linking to those important blog posts. Here’s an example of a product category page linking to related guides:
Not only do these links help buyers make more informed decisions, they also help those linked guides maintain more consistent and long-term rankings.
What if your blog has no traffic at all?
But there’s another possibility, a bigger and broader one that probably deserves a separate discussion: What if your blog doesn’t rank at all, and there’s no organic traffic to convert? Sometimes those unexpected rankings never come, so what can you do?
Ask yourself why you have a blog in the first place and what you’d like to achieve with it. Is it sales? Then research your keywords and create effective contextual CTAs. If it’s links and connections? Then research journalists in your niche and what they are interested in, to better create content for successful outreach.
Researching your keywords and aligning your content to your goals is a huge topic, which is covered in-depth in Moz’s Keyword Research Master Guide, so start there.
Make your blog part of your whole company
The above steps help you revive and improve your blog’s rankings and utilize those clicks better. But without solving an underlying issue, you won’t see your blogging strategy succeed in the long run.
You need to integrate blogging into your company’s life and marketing strategy. Let your whole company submit content ideas for your blog, encourage everyone to share those articles that go live, and invite your current employees to write articles and create graphics for your blog.
Make sure your blog writers talk to your customer support and sales teams because those talk to your actual customers, so they will have a lot of insight into what those people are struggling with and how to help them. Give your blog writers access to your CRM to be able to assign certain clients’ questions and interactions to them as content ideas.
Keep an eye on your blog’s metrics and share important milestones with your company. A blog can make a big difference to your company’s bottom line, but only if you are taking it seriously.
Welcome to 2023! In the first Whiteboard Friday of the new year, Paddy walks through three methods to evolve your link building efforts — not reinventing the wheel, but by improving what you currently do.
Click on the whiteboard image above to open a high resolution version in a new tab!
Video Transcription
Hi, Moz fans. Welcome to another edition of Whiteboard Friday. I’m Paddy Moogan. I’m the CEO of Aira. Today, we’re going to talk about link building and how to evolve your link building to its next phase and to improve what you do at the moment.
So we’re going to talk through three methods to improve your link building and evolve it and improve the results you get for clients.
1. Use the customer journey to generate ideas
So firstly, number one, there is a bit of an issue sometimes with link building and the relevance of the content that we produce for link building. So one way we can actually improve that and make our content as relevant as possible is to use the customer journey to generate link building ideas.
Now, there’s lots of different ways of thinking about the customer journey, different funnels, different methodologies. One thing we use at Aira for this is the awareness, consideration, and decision phases, where the customer goes from being aware of the problem, the pain points, something they want to fix, then looking at their options and trying to figure out, “Okay, who should I choose to solve that problem?”
And then, finally, maybe getting their credit card out, picking up the phone, emailing, that kind of stuff to actually decide who they’re going to go with. Now, we can use this model to actually come up with link building ideas. I’m going to share how we do this at Aira. So there’s four pillars we can use to try and figure out content ideas that are relevant to our customers. Firstly, we start with the audience themselves.
So who are our audience? Who are we trying to speak to? Who are our ideal buyers for the products and services that we sell? So that’s where we start. We then move on to the pain points of our audience. So what kinds of things do they struggle with? What kind of triggers can actually make them start to look for a product such as yours or a service such as yours?
And then thirdly, what solutions do you offer as a brand or a website to try and fix those pain points for those customers and trying to connect those things together? Fourthly, what keywords are you trying to rank for? What are you trying to rank for in Google? How does that connect back to your service and your product? And what that allows you to do is to tie together these stages into actual rankings, actual keywords that you’re trying to improve with your link building and your content.
The nice thing about using these four areas to try and come up with content ideas for link building is that it forces relevance. It’s really hard to actually do this. If you stray too far away from it, you come up with ideas that aren’t very relevant. So if you stick really closer to it, it’s a bit harder. That’s kind of the point. What you’re trying to do here is strike that balance between ideas that can get links, but also content ideas that are very relevant to your audience and who you’re trying to sell your products or your services to.
So try and use these four points here to come up with your ideas, and make sure the ideas you come up with are truly relevant to your customer and not just being done purely for link building.
2. Move beyond campaigns
Second point, second thing we can do to evolve our link building is to move beyond big, shiny, interactive link building campaigns. Now, we’ve built hundreds of these over the years at Aira. We will continue to build them.
I love a good, shiny, interactive hero campaign. But I believe there’s more we can do when it comes to content-led link building. So there’s four things we can do here. So firstly, try and make any campaigns, any content that you come up with evergreen. What I mean by that is it shouldn’t be a start and a stop campaign, where you launch a campaign, do your outreach, and then forget about it and move on to the next one.
Try and create ideas and campaigns that you can outreach and promote over and over and over again. If you can do that, it continues to be relevant throughout the year, across multiple years. It isn’t just a one-off campaign that you’re hoping will get links before you move on to the next one. The second thing you can do is actually not to stop the outreach. So most SEOs actually will outreach a campaign and then stop once that campaign is finished and move on to the next one.
What you should be trying to do instead is create campaigns that you can always outreach and always promote. So try not to stop your outreach, and what happens over time is you build up multiple campaigns, multiple content pieces, and you end up with a nice, big bank of content that you can continually outreach and get links to. It’s not relying upon one campaign after another to get links.
3. Aim for links you didn’t ask for
Third point we can do is to look at existing content.
So again, when you’re thinking about creating link building campaigns, the natural tendency is to think of something new, to think about, “Okay, what can we create that’s brand new, hasn’t been done before, completely new campaign for a website or a client?” Well, actually, that’s absolutely fine, but there may be some existing content on that website that is link-worthy. So make sure, when you start working with a website or a new brand, look at what they’ve got already, see which content they’ve got that might be link-worthy, may have gotten links in the past, and try and outreach and promote that content whilst you’re building new campaigns.
This works really well particularly for agencies, because when you’re launching a campaign, it takes a bit of time to come up with ideas, do the production, do the design. Whilst you’re doing all of that, if you can find existing content on that website that is link-worthy, you can promote that and get links to that whilst you’re building the big campaign. And you’re offsetting the risk there as well and getting links in the short term as well as the long term.
Fourthly, simplify your execution. Not every content piece has to be big, shiny, interactive, full-page takeovers. It can be as simple as a blog post. It could be a small, mini infographic. Try not to worry too much about big, fancy executions. Let the idea lead the execution. And if that means that you write just a very, very simple blog post with a couple of images, maybe lots of copy and bullet points, that’s absolutely fine.
If that’s the best way to execute your idea, then that’s what you should go with and not worry too much about making something interactive or big just for the sake of it. So thirdly, the third thing we can do to elevate our link building is to aim for links that you didn’t ask for. So a lot of the time with link building techniques they rely upon outreach, and outreach means sending emails, trying to get in touch with webmasters and link prospects and get links from them.
That’s absolutely fine. We should continue doing that. But one place you want to try and get to is to create content where it naturally gets links on its own. So you may well do outreach and get links for that content. But as well as that, you’re also getting people who link to that content naturally, so you’re not relying too much just on the outreach. Now there’s three examples here of content you can create that actually will do this. Now the first thing to remember is actually content that ranks tends to get links, because when people are doing research into different areas and trying to write about certain industries, they may look for content they can also link to.
So, for example, if you have a page that’s about statistics and trends in your industry, and someone is writing an article about your industry and trends in 2022, if they find your content, there’s a good chance that they will link to it as well in the content they’re producing. But if it ranks, they’ve got more chance of finding it. The second area is long-form guides, really in-depth, really detailed content. If you can find some good topic ideas to write about in the long-form way, long-form guides tend to rank for a lot of long-tail and mid-tail keywords.
Again, if it ranks really well, it’s very detailed, very techy, this works really well for very technical products and very technical services. If you can create content about your product or service that’s quite technical, it’s got a good chance of ranking well and also being referenced and linked to by other websites. And thirdly, industry reports. So actually, we do this ourselves at Aira.
We produced a State of Link Building Report a few months ago. It’s a report that gets referenced a lot by our industry because it’s a standard in the industry for link building and people think of it when they’re writing their own content. So in your industry, if you can create industry benchmark reports, state of industry reports, surveys, that kind of stuff, that actually shares knowledge and shares information with the industry, it’s easy for people to reference that in their own content that they produce as well.
But again, it needs to rank well for people to find it in the first place. So there’s three ways to actually evolve your link building and try to elevate it to the next level beyond campaigns, beyond outreach, and actually trying to get links that you didn’t ask for. So that’s Whiteboard Friday for me. I’m Paddy Moogan. If you’ve got any questions, drop them into the comments below. I’ll be happy to get back to you.
Over the past twenty years, I’ve probably answered tens of thousands of questions relating to local search marketing. I’m sure I’ve gotten it wrong lots of times, but one thing that’s never happened is that the people requesting my consultation have mistaken me for an omniscient entity. Local business owners and marketers in fora, email, DM, and on the phone understand that I am just a person doing my best to give them good advice. Concerns arise when I consider what applications like ChatGPT may be mistaken for – an expert, a genius, a demigod?
Let’s do an experiment today to see what happens when we ask ChatGPT some of the commonest local SEO questions that humans like me are used to fielding, and let’s imagine what would happen to local business owners if they built their marketing strategy on the advice received.
The Local SEO ChatGPT Experiment
In which ChatGPT merrily instructs me to set about building GBPs for virtual offices
Contrary to ChatGPT’s answer, the guidelines for representing your business on Google prohibit the creation of listings for virtual offices. Unfortunately, our hypothetical local business owner has just built one on the strength of this bot’s advice.
2. In which ChatGPT encourages me to violate Yelp’s guidelines
Yelp specifically states, “Don’t ask for reviews,” but our hypothetical local business owner is out there right now disregarding this policy, thanks to ChatGPT’s advice.
3. In which ChatGPT promulgates one of the most persistent local SEO myths of all time
It must be one of the oldest myths in local SEO, and it’s one that marketers have to keep refuting because it keeps floating around out there in the ether. With AI promoting this myth, our hypothetical local business owner will now spend hours geotagging all of their imagery for its promised ranking benefits.
4. In which the suite number myth lives on
Every local SEO will immediately recognize this situational question in which the inquirer is trying to differentiate one business from another at the same address, and wanting to know if adding a real or fictitious suite number will help Google separate out the entities. What people like Joy Hawkins know that ChatGPT doesn’t is that Google ignores suite numbers, but the myths persist.
5. In which the robot starts to head down a better path and then goes totally awry
Initially, I was worried that our hypothetical local business owner asking whether they can create a Google Business Profile for a lead gen business wasn’t receiving a ChatGPT answer that began with, “No!” Google’s guidelines specifically list lead gen companies as being ineligible. ChatGPT ignores this, and proceeds to instruct the inquirer on how to create a forbidden listing.
A brief ray of hope shone out for me when the ten-step instructions were followed by the warning that (correctly) explains that you shouldn’t create GBPs for businesses that lack physical locations. “Good,” I thought. “This will cause the lead gen practitioner to pause and adhere to guidelines.”
My hopes, however, were short-lived when the bot followed this up with the head-slapping suggestion that the lead gen brand resolve their dilemma with the creation of a virtual office! We’re back to square one with ChatGPT causing users to fall afoul of Google’s most basic guidelines.
The Results of a ChatGPT-led Local SEO Strategy
Let’s imagine that our fictitious local business owner has mistaken ChatGPT for an expert, and has acted on this advice. What happens next:
The Google Business Profile for the virtual office is hard-suspended and likely permanently removed for guideline violations, negating any investment the owner had made in creating the listing.
The hours invested in geotagging images are all wasted, and have eaten up valuable time the business owner could have spent in pursuing tactics that do influence search engine rankings.
The business owner who may already have been heading down the wrong path in hoping that fictitious suite numbers will make Google believe that the separate categories of a single business each deserve to have their own listing will experience suspension of one or both listings sooner or later.
The lead gen business owner will be suspended on two counts: 1) for listing a lead gen business in the first place, and 2) for listing a virtual office.
Once they are caught, the business name is mud on Yelp, where their profile is stamped with an ugly public warning for engaging in prohibited practices.
What we have here, then, is not a local SEO strategy, but the shreds and tatters of disinformation that could result in reputational damage, loss of time, and wasted marketing budget. These are significant real-world consequences.
A Mind of Metal and Wheels
Last year, I was invited to contribute an article to a large publication, commenting on the subject of how writers like myself feel about the intrusion of AI into our craft. I wrote a thoughtful essay entitled Art Vs. The Machine, in which I drew on J.R.R. Tolkien’s philosophy of the gifted human use of tools vs. the coercion of wills that underlie the development of much machinery. My essay remains unpublished, as yet, because the publisher rejected it in favor of an ode to AI that admonished writers to sublimate our fears and get with the program.
My heartfelt sentiments were clearly not a good match for the publisher’s agenda, and the debate about AI should absolutely include all voices and viewpoints – some people are very excited about the arrival of ChatGPT and will laud Google’s analog, if and when it arrives, but others are not applauding. Teachers, for example, appear not to feel that you get the world’s next Tolkien when machines do the “thinking”. My own take at present is that, if I were to use the present incarnation of ChatGPT to write my Moz Blog column, all of you would be reading local search marketing disinformation right now. I strongly hold that you deserve better than that from me.
In fairness, I want to conclude that ChatGPT got some answers to my local SEO questions right, correctly referencing sections of the all-powerful Google guidelines. I was pleased when it was able to talk to me in Swedish, and was surprised when it managed to respond to me in Irish Gaelic. But the bot lacks the art of Elvish. I asked a question in Sindarin, and ChatGPT had no idea what I was saying:
When I translated my question into English (Westron, for all you Tolkien philologists), ChatGPT understood what I wanted, and its reply contained a revealing proviso that should be a reminder to all AI users about a critical difference between the machines and us.
For now, I’ll be continuing to write my column in the old-fashioned human way, remembering specific (and wonderful!) interactions I’ve had with local business owners and industry colleagues, and drawing continuously on my personal experiences.
Exemplo de Mulher Cursos 