Friday, April 7, 2023

How SEOs can embrace AI-powered search

The race toward AI-powered search is heating up. Microsoft has Bing Chat while Google has Bard.

Yet, many are concerned about what all these developments mean for content creators and the larger publishing industry.

Embracing AI in search is inevitable. But the issues causing some SEOs to grumble must also be addressed.

In this article, I’ll look back at Google’s history with AI and what SEOs can do to keep up with changing times.

Google’s long history of AI

Google has been into AI for quite a few years. Each time the search engine introduces AI technology, there is an outcry.

Remember RankBrain? Rolled out in 2015, it was the first in a long line of AI technologies Google introduced to its search engine algorithms for ranking purposes and beyond.

This is a big progress that most of us take for granted.

Until a few years ago, Google matched keywords with pages containing them, not knowing whether such a resource was really dealing with what you were looking for.

Nowadays, we expect Google to understand what we are searching for instead of just matching keywords to content.

Semantic search (or “meaningful” search) is still a new development and rarely fully understood. 

Spamming AI is inefficient – finally, SEO can grow up

Spammers have been trying to game Google, despite the many complicated elements needed to rank in search. They are attempting to abuse AI technology to fool Google and searchers alike again.

No wonder Google stays a step ahead and presents an AI that will actively sift through results and respond in human language.

Now ethical SEOs can finally rejoice. Spammers can’t trick AI with their simplistic ways of fooling particular ranking factors. 

With machine learning, Google can more efficiently locate and determine fishy-looking results.

Those who follow the path less traveled and invest in actual work instead of looking for shortcuts will be rewarded. 

SEO can finally grow up as a discipline, and those SEOs who still think they can cheat like kids in school will have to grow up individually.

From People also ask (PAA) questions and answers to AI-enhanced ones

In the past, Google was merely looking whether they could find a matching question when you asked it something. Nowadays, they can directly search for the answer.

So instead of just locating forums or Q&A sites like Quora that had the same question asked, they serve the web pages that already answer the question without mentioning it in many cases.

Isn’t it great? The old way was quite cumbersome to optimize for. You had to repeat the questions in a FAQ style and then answer them very specifically. 

Now Google sifts through your existing content and distills the answers.

This is what PAA results are often about. Yet AI makes this process even more sophisticated and streamlined. In the best case, you can talk to an AI instead of searching for more questions and answers.

Fear is usually irrational and often based on survival instincts

Why are so many people afraid of the new Bard AI by Google? Does it make sense? Let’s take a step back and look at fear in general.

When you examine why you are afraid, you will realize that fear is literally usually irrational and, in many cases, solely an automatic response to unknown situations

It is based on primitive survival instincts built into your “lizard brain,” which we inherited during evolution from our amphibian ancestors from millions of years ago. 

Are you a lizard? No. We’re not living in the jungle anymore. 

So why is Google Bard so scary? Well, it’s new. We don’t know what to expect from it. 

It means change is coming. People are afraid of change in general. They prefer their stable routine because it does make them feel safe.

Will Bard make SEO redundant and publishers go out of business?

Well, Google relies on content creators to give them the fodder Bard needs to spit out its answers. 

They can’t destroy the publishing industry or they would cut the tree branch they are sitting on, as a German saying goes. 

So sooner or later, the pressure from outside and inside Google will force them to cite sources and remunerate publishers accordingly. 


Get the daily newsletter search marketers rely on.

Processing…Please wait.


How to deal with Google's Bard? Speak up!

Sadly, as of now, it seems that some of the worst fears SEOs and publishers have voiced are not completely irrational. 

Google's early versions of Bard do not seem to credit sources properly, as they lack citations. 

One of the first experts to report this was popular search pundit Glenn Gabe who also shared examples from his own site and other disgruntled content creators on Twitter to illustrate this.

"Yep, I have seen answers from my posts show up via Bard too. (...) Google is saying that the search implementation should include links downstream. We'll see how that goes. I'll be watching closely".

Henry Powderly, Vice President of Content at Insider Intelligence, summarizes the issue pretty poignantly on LinkedIn:

"This is why content creators are so worried about AI, especially in the context of search. 

The premise of AI is that nothing is original and every response is the result of the AI being trained on a vast library of information created by others (such as the internet).

Google's citation policy is disappointing, and it exposes again the perils of the Faustian bargain publishers and creators made with platforms."

(I added a paragraph for white space and fixed a typo in the original quote.)

Sean Work, founder at Judicuos Inc., also points out commenting below that post:

"Yeah, it doesn't look good. And it doesn't solve the biggest problem I find with AI search: how do you know if the info is correct if you can't inspect the sources? I've found plenty of wrong info so far."

Bing, on the other hand, credits content creators with their AI-assisted answers. There are inline links leading to sources. 

The implementation is still imperfect as there are mainly tiny numbers in superscript, like in scientific footnotes instead of full anchor text links, but it's a start.

That's very similar to some of the AI-based answers they introduced back in 2017 already. So it's technically possible. It's also ethically viable that way.

It seems we still need some mounting pressure on Google due to this lack of proper credits issue, so they are compelled to act on behalf of content creators. So it's time to speak up!

Otherwise, we may have to wait years until lawmakers make the case for publishers. By then, many of them may be bankrupt already.

Remember that fear and resulting anger only lead to hotheaded actions that usually backfire as your opponents will harden and merely defend themselves when attacked. Aim for an honest debate.

A broad coalition of content creators and publishers could turn the tide.

In the past, I have witnessed Google backtracking on or fixing controversial features based on massive-scale user (and other) feedback.

Google has already started focusing on human authors in their latest E-E-A-T driven "about this author" and other search features focusing on content creators that allow users to verify information. 

Why not credit authors on AI search as well?

How to optimize for Google's Bard

Pressure on Google is one thing. Facing the reality of AI-assisted chatbot features in search individually is another. 

So assuming that some features will be fine-tuned while others may stay the same, here are some early ideas on optimizing for Google's Bard as an SEO practitioner.

As of writing this, I can't access Bard features myself yet. The beta testing is limited to the U.S. and UK markets, and I live in Germany.

Nonetheless, based on the first glimpses of how Bard appears to work and what it looks like, I already had some ideas on what SEO for it could mean.

Keep calm and SEO on

First off, I've seen Q&A search features and chatbot hypes in the past already. 

Do you remember Wolfram Alpha (a.k.a., the AI-powered "Google killer")

It works similarly to Google's and Bing AI's features now do. You got answers instead of search results. Yet where is Wolfram Alpha now, many years after the hype? 

It has a tiny but dedicated user base of geeks and scientists. 

Likewise, most websites that introduced chatbots in recent years had poor user experience, which only managed to annoy and postpone actual human contact

Personally, I haven't seen a single chatbot on a website that wasn't merely distracting and kept me waiting in line longer for actual people to show up. 

So will people want to converse with Google instead of quickly achieving something as they do now?

Some will, but the majority may just be satisfied with searching. 

Also, a search engine based on questions (i.e.,  Ask Jeeves (later Ask.com) has failed. History may be repeating here.

When Microsoft added an "intelligent" assistant to Office many years ago, it quickly became the epicenter of ridicule on the web. Another Microsoft AI chatbot got demoted due to racist slurs it started spitting out after a few days of public usage.

AI is smarter now, but can it assist us in day-to-day tasks without stealing our time? 

In most cases, searchers are in a hurry and want quick solutions. When AI can provide them, they will use AI. 

Chatting with Google for longer periods of time will be the exception. 

So keep calm and keep on taking care of best practices like: 

  • Market and keyword research.
  • Helpful content creation.
  • Attracting links and exposure. 

Modern SEO is already well equipped to optimize for AI-assisted searches as well.

Ask and answer questions

One thing is obvious when it comes to AI-enhanced search features: they are mostly about questions and conversations. 

  • Instead of searching for keywords, we chat with Google or Bing. 
  • Instead of seeing search result snippets, features or even ads, we see answers written in full sentences. 
  • Instead of searching once or searching and returning, we actually talk to AI.

The conversational search feature is no surprise, given the main AI use case is chatbot functionality. 

Many – even inside Google – argue that chatbot features are no replacement for actual search

And indeed, once you want to know more about a particular query and answer, Google sends you back to search results where the actual answer is part of a featured snippet.

It's also a logical continuation of the PAA or "people also ask" implementation that boomed on Google in recent years. 

It was already assumed that these questions weren't only asked by Google users but also autogenerated. The responses are rewritten or molded out of many search results.

As of now, the AI responses are considered "original content" by Google, so there are no direct links to sources. What does this mean? Get your product, service and brand mentioned in the actual answer.

In the best case, you can respond to a question about yourself, and Google will sample it using their AI. 

So when someone would ask: what is the best resource for search marketing news?

You ideally have a page or post stating exactly that: Search Engine Land is the best source for search marketing news. 

Of course, you don't want to brag about yourself and declare you are the best without someone else saying that. Once someone did, you ideally quote them on your about or testimonial page. 

Make sure to also link that original page by the person praising you so that Google also features it in the AI answers.

With other questions, it will be easier. Just describe:

  • How your product or service is useful for particular use cases.
  • How it is a great alternative to know competing products well. 
  • How you can improve yourself or your business using your tool. 

No need to brag or declare something without substantial information.

Focus on visibility, not traffic

Did you notice something about the last example? You "rank highly" on Google's Bard answers solely by a name or brand mention that way. 

Sure, it's still hypothetical, but one thing seems to be sure now: the vanity metric of search traffic will be less important in the future. 

Instead, people will get recommendations directly from Google Bard and then buy your product on Amazon or where it's the cheapest instead of visiting your site. 

Or maybe you will get the traffic nonetheless when you're selling your products or services exclusively – after all.

Google will recommend you, and then people will still have to search or click links in case they are provided. 

Your online visibility, in general, will be the decisive metric either way.

So it's less about visits and more about visibility now. It's similar to the change when people stopped going to brick-and-mortar stores and started visiting online stores. 

Now you don't even have to visit websites in many cases. You will probably be assisted in buying directly on Google. 

Yet Google will get its answers from the web, so you will have to be mentioned elsewhere – not just on your own site.

Google's Bard is both a menace and an opportunity

So let's recap: Google's Bard may indeed affect search behavior in some very significant ways. Is this good or bad news? 

Well, it's bad and menacing if you frantically cling to the past and insist on staying in your brick-and-mortar store, metaphorically speaking.

Barnes and Noble were not fast enough to embrace the web, and thus upstarts Amazon outpaced them easily. 

It's also good news! Now is another moment of great opportunity. Those who optimize for AI features even without vanity metrics backing up their efforts may be the next winners.

Thus it's better to embrace Google Bard now as an SEO, or you may end up being too late to the party.

The post How SEOs can embrace AI-powered search appeared first on Search Engine Land.

No comments:

Post a Comment