As we move toward voice search, Google evolves at a fast speed, and it adds new features that in a way makes us feel how the voice-centric world might look like. In this article, I want to show you some observations done on Google’s search results and what they can tell us about what the future of search might look like.
Let’s start with where Google is now, in terms of business model and search.
Google business model today
On my blog, I covered from several perspectives the Google business model. If you look at Google financials in 2017, you realize how the tech giant from Mountain View has been branching out in several areas. However, the core of the company remains its advertising business model. Still, as of 2017, 86% of Google’s revenues came from advertising.
For a company that is getting closer to the trillion dollar valuation, one might ask, how sustainable would it be a company that has built its fortune on advertising alone? Even though there is nothing wrong with advertising in itself, the real issue is how an asymmetric business model, which gains from hidden revenue generation might not wake up the regulatory giants (like the European Union) and break them apart?
While businesses understand that Google is a big advertising page, where they can get featured by bidding on keywords. In many cases, users are not aware of the mechanisms behind Google revenue generation. As users learn more and more about their privacy, search engines — like DuckDuckGo — are gaining momentum. What if users would end up caring more about their privacy than the comfort of a search that is highly personalized?
To make things worst, an entire industry was born out of Google opacity in sharing how its algorithm works: that is called SEO (search engine optimization). This industry has been snowballing, at the point to become a multi-billion industry. On the one hand, SEO has helped Google become a better engine. In fact, with a race between SEOs trying to rank their content — by understanding how Google’s algorithms worked — Google had to evolve quickly, to avoid to be gamed by an SEO industry sustained by businesses that were trying to rank organically, without bidding on keywords.
The process of an evolutionary fight is typical in many contexts. It is also true that it has been in a way exacerbated by the fact that a commercial engine like Google has become the center of the world information. In short, when it comes to information and knowledge, one might wonder whether the business model Google picked is the right one. This question is not new. That is a question asked over and over again in the last two decades.
However, now that we’re moving toward voice search, the information might get more and more polarized and skewed toward one answer within multiple devices. Wouldn’t make more sense to ask this question more pressingly?
Let’s give a quick look at where Google is at today.
What’s next then for search?
When Google started to index the web, its mission to “organize the world’s information and make it universally accessible and useful” also made it a commodity.
In fact, by indexing and ranking billions of pages Google became a threat but also a great distribution opportunity. As new blogs joined the web, there was no way out. Either you were featured by Google, on its first page, or you would be cut out of the web.
This logic continued for years. As Google worked primarily by ranking pages based on a mechanism of reference from one page linking to another (so-called backlink); over the years Google also started to read the metadata of those pages. That is data that describes the web pages.
In other words, as the web grew Google kept collecting more and more data around a page. However, as the algorithm would not be able to process large amounts of data Google didn’t probably do much with that data; until in 2012 it finally started to roll out the Knowledge Graph. A massive database made of triples, that allowed Google to manage a vast amount of that and be able to do something with it. In short, while all the data in the past might have been useless, as there was no way it could be managed.
After the knowledge graph, in 2011, Google finally started to do something with that data. There was also another critical step. For years Google has assessed the quality of content by looking at straightforward metrics, which worked quite well. Yet, as the web grew and SEO practitioners got smarter, also Google grew in sophistication.
Today Google might be looking at more than 200 main signals, that given the possible combinations might well be about thousands of factors. Therefore, Google is becoming more and more a black box, where we all see what comes out of it. But none can really understand what is going on in between the search and the result.
One thing is critical though. Finally, websites can use a tagging called Schema.org, which is a vocabulary developed by search engines to assign meaning to pages. While Google keeps assessing the meaning of those pages by using direct and indirect signals. Finally, website owners can have their say and tell Google what their web page is about. While Google has still the option to use that data. It’s clear this is an opportunity for websites owners to create a better communication flow with Google.
That also opens up the possibility of alternative business models on the web, where content has been commoditized.
Valuable content is not a commodity: as it should have been since the start
Since Google indexed as much content it could, it also created a frictionless experience for its users. Today one of the most used words in the tech space is UX or user experience. The whole idea is about creating products that users want to consume and that are as frictionless as possible.
How the term "user experience" grew with Google (mentions of "user experience" over the years in millions of books cataloged by Google)
User experience grew with the rise of the web; there is no doubt Google was one of its most fanatic supporters. Today, having a slow website can mean being out of search. Of course, Google tells you that content comes first. In most cases, content alone won’t bring you far.
This frictionless experience became one of the most important metrics on the web. Google also has a set of key parameters to measure that. Those are the “user-centric performance metrics” and as Google puts it:
When a user navigates to a web page, they’re typically looking for visual feedback to reassure them that everything is going to work as expected.
Is it happening? Did the navigation start successfully? Has the server responded? Is it useful? Has enough content rendered that users can engage with it? Is it usable? Can users interact with the page, or is it still busy loading? Is it delightful? Are the interactions smooth and natural, free of lag and jank?
While this made total sense, it also created such a frictionless experience to give the impression that content could be commoditized.
What matters was the user experience!
Now you might argue, “isn’t great content also in line with user experience?” Well… yes and no!
In general, it is intuitive to think of great content as something that benefits from high engagement rates. If a piece of content is great, you’ll consume it all. However, to read that piece of content, you still have to get there. And guess who stands in the way? Yes, Mr. G! Google is the monopolist of search and the tool that can distribute content across the globe.
If Google won’t bring people to read your content, there is no way anyone might be able to experience it. And if none is experiencing it, how do you assess user experience on that content?
What I mean here, is to think about two kinds of user experience. One, which we might call “technical” and one which we might call “classic.” When it comes to the technical user experience that is assessed often a priori (before the user can see your page). Google is telling you (based on its user-centric performance metrics) that if your server is not responding successfully (which also means fast enough), those people won’t even get to your site.
Welcome to the Google’s dictatorship!
Once again, Google tells us that if you have great content, that comes first. However, great content without the right SEO strategy might be doomed. Also, as we’re transitioning toward mobile search, Google also starts to impose formats that many publishers don’t like, and for a good reason.
Take the polemic around AMP. It stands for accelerated mobile pages, and it is a format for mobile, which is intended to give a clean user experience. However, it also strips out from web pages important parts that might be critical for publisher navigation and monetization.
As pointed out by Kyle Schreiber:
Make no mistake. AMP is about lock-in for Google. AMP is meant to keep publishers tied to Google. Clicking on an AMP link_feels_like you never even leave the search page, and links to AMP content are displayed prominently in Google’s news carousel. This is their response to similar formats from both Facebook and Apple, both of which are designed to keep users within their respective ecosystems. However, Google’s implementation of AMP is more broad and far reaching than the Apple and Facebook equivalents. Google’s implementation of AMP is on the open web and isn’t limited to just an app like Facebook or Apple.
The Google AMP Lock-in experience
When you’re in business or the life of your organization depends upon the visibility it gets from the web, do you have a choice not to follow what Google asks?
As a publisher myself, I understand that SEO is a critical part of my strategy. If I stop using an SEO strategy, I’d probably revert to 2015. At the time I was the only one reading my blog as I didn’t get traffic from Google even though my content was good.
Initiatives like AMP also tells you that eventually, Google is trying to create a full experience, where users don’t have reason to leave the engine (which can be as well compared to Facebook lock-in) unless they have to transact, in that case, you’re allowed to do so because that is when Google makes money.
That is even clearer with the latest Google algorithm change, the so-called Medic Update.
Business models for voice: four possible scenarios
I imagined four main scenarios going toward voice search. Each of those scenarios might be more or less plausible. Except for the first scenario, the others are entirely speculative. However, it is interesting to start thinking in concrete terms how search might look like in the future.
Is the subscription economy the answer?
While the advertising business model is pretty lucrative for Google, that is not for online publishers (at least not anymore). As Google has been able to increase its ability to attract businesses to spend their digital marketing budget via AdWords, it also offered to bloggers the ability to quickly monetize their content with AdSense.
AdSense is the fastest and easiest way to monetize free content on the web, yet it’s not sustainable. The digital publishing industry has realized that and that is also why some players, like the Financial Times, for years have relied on the subscription business model.
In the past, though it was pretty hard to make the subscription business model work as Google aggressively indexed pages on the web and the more free content you made available, the more you could get exposure, thus monetize with ads.
Yet, as of now, large publishers are reverting to a paywall and a subscription-based business model. What’s interesting though is we now are that with the semantic tagging, you can point to the search engine that a web page has content which is not accessible to anyone (see the picture above, the property called isAccessibleForFree, which you can set as “No”). That means the search engine knows that even if the user is navigating the page and going back to the search results quickly, that has nothing to do with user experience but with a paywall.
This, in turn, should make easier for publishers to build a business model that doesn’t rely on advertising alone. As Google itself is differentiating its business model in new areas (from the Google Bets to the Google Other Revenues sources) why not allowing publishers — the juice of Google’s organic results — do the same?
Google is massively rolling out new features on its search results. One interesting example is the Featured Snippet Filter Bubble. That is a feature that Google offers to users to address a specific question, but also to allow them to chose among several intents. In short, a user looking for something might have several intents on that same search. Think the case above; I might be looking for a job from New York. Yet when I search “top companies to work for” I might well be looking at companies in other cities (Chicago, Atlanta, Dallas and so on).
With the Featured Snippet Filter Bubble, Google can tap into the users’ minds and offer them more and more relevant answers. Also, one thing of the filter bubble (as the name suggests) the user might well be satisfied with everything there is to see within that snippet without ever leaving the search results. That might imply people will be spending more and more time on Google search result page, which is comparable to the Facebook effect. It is interesting to see how the engagement metrics of Google are getting closer to that of Facebook:
It is important to notice this is a rough estimate that might be far from actual numbers but decent for a quick comparison
While Google doesn’t create any friction by allowing users to get in and out from the search engine, Facebook requires users to register to get access to the platform. Yet, when it comes to branding, there is no doubt that Google is among the most powerful brand in the world.
Even with a frictionless experience that is meant to bring a user (based on the intent) from place A to B on the web (in many cases now the user never leaves the engine) it still has extremely high engagement rates.
In this scenario, where Google locks in user experience at the point of meeting its search intent and allow the user to consume all the information she/he is looking for, all that is left is the transaction. Google is not a store, then it will send the user to transact on another platform. In all the other cases, Google might be able to provide an answer, to any question by mixing up snippets from several sites, with great precision. In that circumstance, the experience offered by Google might be way superior to the experience provided by a web page, where a publisher can only guess the user intent. In short, that is a war on unequal terms.
However, Google might still need snippets from several sites on the web, thus by sending traffic here and there to those sites. It will also share revenues or some sort of reward back to the web properties that provide the best snippet of information to meet specific users’ intents!
Is there an alternative to this scenario?
Blockchain + Knowledge graph = The end of Google?
If there is a word that has dominated the web in the last couple of years that is Blockchain. For those that heard about it, they probably heard it in the context of Bitcoin. However, the Blockchain that was created by Satoshi Nakamoto (none knows if that is a real person or a group of people) it was just the first Blockchain. Yet, as of now many new Blockchains protocols have sprouted up and have proved successful at several levels.
Let’s clarify first what a Blockchain is. That is simply a distributed ledger that allows people to transact without a central authority. Also, those transactions (or exchanges) between people are anonymous and cannot be tracked back.
What’s most interesting about the Blockchain is its ability to allow people to transact entirely at the bottom-up level. In other words, finally in the history of humanity that is possible to create commercial organizations at scale without the need for a central authority. What does that imply? That might suggest the end of bureaucracy, but also the end of the center of powers that retain most of the profits and create asymmetric business models.
For instance, when Facebook gathers the data of its users then it makes money by selling them advertising:
That model provides a hidden revenue generation that if we were to think at the business level is quite profitable. However, if we think of the negative externalities that model might create, one starts to wonder, is it worth to let Facebook collect all this data about its users just because it has to offer a better advertising experience?
The Blockchain alone can’t solve the issue. Yet it can allow the owners of data (be them users or publishers) to be part of the game.
Imagine the scenario of a publisher that has a knowledge graph built on top of its domain. That same publisher has now data that can make accessible to machines (like Google) so that they can use that data to provide content and information to final users. By plugging in a knowledge graph on a massive index, where transactions and licensing of that data is managed with a Blockchain mechanism might be a viable solution.
We’re building a search engine on top of a Blockchain might be a utopic endeavor (imagine that more of that a Blockchain has to handle at super fast speed to offer relevant results). Using the Blockchain instead, to plug the knowledge graphs of publishers and let them manage the way they want to share and license that data might be feasible. While letting semantic search engines (like Google) handle the billions of queries each day and build up content based on those knowledge graphs. Just as Google serves snippets to users. Those semantic engines can extract information from billions of triples comprised of those large knowledge graphs, attached to the Blockchain, which becomes the market where those publishers can transact and monetize that data.
In that scenario, there will be no single owner, no single company owning the data. Ideally, a symmetric business model where publishers make money with their data, the semantic engine offers the most relevant results based on users intents and put them in connections with the proper local online store where to make the purchase!
The apocalyptic scenario: Google AI takes it all!
When Google’s CEO, Sundar Pichai announced Google Duplex, back in May 2018, that showed the weapons, Google has at its disposal. Google Duplex (a machine learning system able to have complex conversations with humans) showed what Google is capable of as of today. Of course, Pichai explained Google Duplex as a way to solve practical problems (like booking an appointment at the hair salon). Yet it also showed the power of a mechanism that if unleashed can already take over the jobs of many call center people. Not that someone is trying to defend those jobs. These are probably the least paid jobs in the world. Yet when we look at the natural language understanding ability of Google Duplex one starts to wonder whether that same machine wouldn’t be able to do the same with content production.
Given enough content on the web to train Google AI wouldn’t that be able to produce its own content?
When and if that happens Big G won’t need anyone but its AI to meet users’ intents.
If this scenario prevails, would this be really the way information is supposed to be organized? In short, does the fact that an algorithm is able to provide more relevant information than a human justify this scenario?
Originally published at fourweekmba.com on September 19, 2018.