How New Search Changes The Web
Retraining the consumer, chatbot optimization, the rise of the Smart Page.
Welcome to People vs Algorithms #68.
I look for patterns in media, business and culture. My POV is informed by 30 years of leadership in media and advertising businesses.
Sometimes it’s nice to read in the browser.
Determined not to be “Chegged” by Microsoft or OpenAI, Google came out swinging this week at their annual developer conference, Google I/O. If your stated mission is to “organize the world's information and make it universally accessible and useful,” you are now, by definition, an AI company. Google needed to show leadership. They also needed to show how their reliable search cash machine will withstand an uncertain AI future. Anyone invested in the future of the web waited eagerly for a sign of what’s to come. Most got what was expected. The web is about to change. A lot.
Happily for Google, the market reacted to the announcements with enthusiasm. Following the event, their market cap was up $131 billion, as much as 9%. The well choreographed presentation demonstrated a few things. First, OpenAI will not easily walk away with chat leadership. Google’s ChatGPT competitor, Bard, was made available to everyone for free, challenging OpenAI’s $20 monthly subscription model. Bard’s improved core model (PaLM 2), its connection to Google’s search index and formidable data sets (Youtube, flight and shopping data, to name a few) give it immediate functional advantages over ChatGPT. Questions of margin brought on by enormous compute costs are less important than preserving competitive position through this inflection point. Anyone who controls share in the chatbot war will be rewarded handsomely. Seems like a good bet.
Google has hard earned depth in AI and is showing its ability to commercialize it across a suite of products faster than you can say “ok Google.” AI will quickly be woven throughout the Google suite of products — Duet (Google’s answer to Microsoft CoPilot), Cloud, Gmail, Maps, Workspace, Android. Right now, a helpful little button is sitting next to my cursor inside Google Docs is eagerly waiting to help me write this note.
Beyond Google’s newfound nimbleness and sudden willingness to launch imperfect AI tech, the announcements reflect how important distribution advantages will be to AI’s winners. AI in my text editor will trump marginally better OpenAI answers the next tab over. Google is in a very envious position.
While everyone was excited about the AI-ification of everything Google, not to mention a new folding phone that doubles as your personal living room TV, the most important announcement by a mile was the changes unfolding inside of Google Search. Everything on the internet is downstream from search. It impacts us all. An entirely new construct, entitled, “Google Generative Search”, only available in Google Labs for now and rolling out to Chrome desktop and the Google App in the US in a couple of months, shows just how much the internet is about to change.
Much has been said about how AI-driven search changes will bring more of the query response into the SERP (search engine results page), choking Google’s spigot for publishers and others that rely on it for traffic oxygen. This seems obvious. But AI will do much more than slow traffic flow. Here are five changes that come to mind:
1. A new podium for links
Per my comment above… The I/O demo revealed just how profoundly true SERP compression will become. This is a real estate question and more. The AI response now occupies oceanfront position on the results page. Links are relegated below (see the screenshot). Three results are listed inside of the “chat” result as reference sources to a sample travel question (yellow box). These are obviously gold. Winning a spot in this list will be extraordinarily coveted.
Importantly, what it will take to win here will differ from what it takes to win in SERP today…
2. From answering a question to providing authoritative reference source
Old SEO (search engine optimization) meant showing up in the search page as a reliable and useful answer to a consumer question. Enter “What is the best, affordable hotel in Barcelona?” and Google would point to a host of articles providing answers to the question. But, what if much of the context is provided by AI inside Google’s search page? It seems reasonable that the way they value a link next to the response will evolve.
Watch the algorithm start to reward different kinds of content and providers. Trusted brands become ever more important. Authoritative data sources, rankings and research providers grow in importance as does voice and point-of-view that cannot be absorbed or mimicked by AI. The new search link is a “citation.” Citations are not answers. They are references. This will force new thinking in how you optimize data and content for chat-driven results page. Chatbot optimization will introduce new rules to search optimization.
But the SKU will persist…
3. SKUs trump Pages
A SKU is a transactional unit that the AI cannot demolish. It represents a unique entry in a database connected to an object, real or intangible — something that can be purchased mostly. Traditionally the SKU is a commercial object; a bike, mattress, a flight or hotel room. But I also think of it as people and places, like a listing of a divorce lawyer at a law office in your neighborhood. A video player is a SKU of another kind. It is a discrete thing that cannot be de-contextualized by AI.
People use search to sort through billions of entries to locate the right SKUs. Today, millions of pages of content provided by thousands of publishers provide connective tissue between the search box and underlying SKUs. This affiliate layer provides money-making context for products and services and is the performance advertising substrate that pays for much of the internet. AI responses threaten to cut out content providers and provide navigation directly to the SKU.
Anybody that can find a place for their SKU in the search result is advantaged. More broadly considered, data ownership of SKUs, people, scientific data, weather… anything that becomes a valuable and actionable reference point next to content does very well in a new world. Google will still have to point to you. Media ought to think about how this impacts data and content strategy.
4. The web page get Smart
Think about chat and web pages along a continuum (see last week’s post). Chat is a real time response to a question or request. Chat pulls content and utility from everywhere and assembles neatly into personalized response. Pages are a static (or mostly static) organization of information and functionality that exist in anticipation of a user request.
In some crazy utopian future, an omniscient AGI chatbot might have God-like powers and you would never need pages. Although we have God and we still have the bible. Perhaps you always need both.
The point is this… Bing, and now Google, show us how chat and pages work together. Information on a page provides context and a canvas for the chat experience. Chat makes pages smarter, interactive and more functional than dumb HTML pages of old.
I think of these hybrids as Smart Pages. A Smart Page is one in which the interplay of AI powered chat, mixed media interactive results and contextual content provide a deeper, multi-dimensional experience for a consumer.
Google’s new Smart search page brings personalized content to the static search results experience. The same logic applies in countless use cases. In a multi-step insurance journey, a Smart Page would assist with product selection and configuration in a way that made the process simpler. Smart Pages will dimensionalize media pages with summaries, comparisons, quizzes, extended explanations, additional reference sources and the like.
Google trained consumers on how to find things on the web, navigating a vast universe of links and pages. The average time spent on a web page is less than a minute, suggesting our web journey has become fragmented and inefficient, largely because it is shaped by ad servers that need to render ads page by page. Ad incentives favor page sprawl.
Expectations will slowly change, starting with our search starting points. Google’s new search interface will introduce a fresh way of moving between chat and content without leaving the page. At first interaction will be largely text based, but will grow to encompass all media types. The way we think about web pages will evolve. They will feel much more app-like and personalized. Search/click/page/repeat will evolve into much simpler AI-assisted journeys. How we integrate advertising will shift by necessity.
Affordable open source AI models will power every online experience. The most important will be powered by valuable, proprietary data sets. Expect this to happen very quickly.
5. Banners fade to irrelevance
The back and forth interface dynamic inside of a Smart Page will not be well suited to the IAB boxes that defined the first generation of advertising on the web. Expect advertising to look more like what you see on Google — keyword positions inside of chat responses priced by performance. Affiliate models will pervade. Page-based impression metrics that have anchored media models inside of old-fashion ad servers will quickly fade. Sponsorship will become more common as advertisers look to place their brands next presenters of compelling content/functional experiences. These will avoid the historic wrestle for pixel real estate that defined so much of the banner-based models to now.
All of this is tremendously healthy for the open web. The refactoring of the search-link-page model will force a wave of innovation around useful functional experiences that tie together data, funnels and media. Value will attribute to those that have differentiated data sets to power AI models, unique POV that resists AI compression and the SKU’s that are indivisible in this emergent AI pyramid of value.
Have a great weekend. Thank you Moms…/ Troy
ON THE PODCAST
The battle for engagement has long been a losing battle. The arrival of AI, and the prospect of limitless synthetic content, means media companies on the hunt for engagement are bringing a butter knife to a gun fight. Perhaps it’s time to consider ambient media, the humbler request for people’s partial attention but done in such a way as to establish durable human connection.
Speaking of duets
Two of the greatest.
"Girl from the North Country" (occasionally known as "Girl of the North Country") is a song written by Bob Dylan. It was recorded at Columbia Recording Studios in New York City in April 1963, and released the following month as the second track on Dylan's second studio album, The Freewheelin' Bob Dylan. Dylan re-recorded the song as a duet with Johnny Cash in February 1969. That recording became the opening track on Nashville Skyline, Dylan's ninth studio album.
Rolling Stone magazine ranked the song 30th on a list of the "100 Greatest Bob Dylan Songs". In an article accompanying the list, Rolling Stones guitarist Keith Richards wrote: "While the British Invasion was going on, Bob Dylan was the man who really pulled the American point of view back into focus. At the same time, he had been drawing on Anglo-Celtic folk songs, and that's certainly true of "Girl From the North Country". It's got all the elements of beautiful folk writing without being pretentious. In the lyrics and the melody, there is an absence of Bob's later cutting edge. There's none of that resentment. He recorded it again later with Johnny Cash, but I don't think it's a duo song. Bob got it right the first time".
"Renegade" is a song by American rapper Jay-Z featuring fellow American rapper Eminem, who produced the song with Luis Resto. Written by all three, the song appears as the 12th track on the former's sixth album The Blueprint. It was originally a collaboration between Eminem and Royce da 5'9" as part of the Bad Meets Evil series, but Royce was later replaced by Jay-Z. The original can be found on mixtapes and has been leaked onto the internet.
The Jay-Z version of the song, released in 2001 as featured on The Blueprint, was later included as a bonus track on the deluxe edition of Eminem's 2005 greatest hits album, Curtain Call: The Hits.