What we can learn from Google’s congressional hearing7 min well spent

Last Wednesday, on 7/29/20, Sundar Pichai answered to the American congress for Google’s anti-competitive behavior. Some really interesting things popped up during the hearing, e.g. Google’s business model creating a conflict of interest with websites, more context around user behavior ranking signals, and traffic espionage. Let’s dig in.

It’s not me, it’s you

Not too far into the hearing, Congressman and Chairman of the House Antitrust Subcommittee David Cicilline asks Sundar Pichai (6:17 in the video below):

Mr. Pichai, most Americans believe that when they enter a search query that what Google shows are the most relevant results. But increasingly, Google just shows what’s most profitable for Google like Google Ads or Google’s own sites. Isn’t there a fundamental conflict of interest between users who want to access the best and most relevant information and Google’s business model, which incentivizes Google to sell ads and keep users on Google’s own sites?”

It’s a key moment in Google’s hearing but also in the whole conversation around regulation and anti-trust that has been going on for years. If you want a grasp, I wrote about the various accusations from mistreated companies in The nasty side-effects of Google Search Traffic.

What Cicilline refers to is what I called “SERP Mazes“: click journeys that keep users in the search results so Google can display more ads and increase revenue.

Pichai’s response:

We show ads only for a small subset of queries, where the intent of users is highly commercial.

Queries exist on a power curve of search volume. Every beginner SEO knows about the shorthead vs longtail curve. A few queries hold the majority of search demand and that’s the subset Google shows ads for. And because Google makes most of its money from ads, this “subset” is worth more than $160b (Google’s revenue in 2019).

The core problem with Google’s anti-competitive behavior is that its business model poses a fundamental conflict with other businesses’ interests. These are more Cicilline’s words than mine and he expresses the problem perfectly. Google prefers its own assets over websites and destroys the ecosystem it built over decades.

One example is the mortgage calculator I Tweeted about in November 2019.

It has now reached the next step in the evolution: Google partners with the Consumer Financial Protection Bureau to show a mortgage calculator knowledge “app” (source) – after 4 ads. It’s truly more than just a knowledge card, which is why I call it an “app”, with FAQ, explanations, dictionary, and calculators.

Anyway, Mr. Cicilline then goes on to the death strike (at 7:32 in the video):

Our documents show that Google evolved […] to a walled garden that increasingly keeps users within its sites. Emails show that over a decade ago, Google started to fear competition from certain websites; web pages that could divert traffic and revenue from Google. These documents show that Google’s staff discussed the proliferating threat […] that these web pages posed to Google. Any loss in traffic to other sites was a loss in revenue. Memos observe that certain websites were getting ‘too much traffic’, so Google decided to put an end to that.

Mr. Cicilline at the congressional hearing on 7/29/2020 to Sundar Pichai

It was an issue for years but intensified with Google’s transformation from Search to Answer Engine that started around 2018.

Google broke this unwritten pact when it started to answer searches directly without sending traffic to other sites. Why send users to other sites when they could get the answer right in the SERPs (search engine results pages)? The new strategy reduces friction for users, keeps Google’s growth rate high, and shareholders happy.

The nasty side-effects of Google Search Traffic

It continues with Google’s immense pressure to drive bottom-line as the revenue from ads is stagnating. The evidence that is now coming out paints a much more sinister picture than we could have imagined from outside.

Traffic espionage

Internal documentation shows how Google built products to prevent competitive sites from getting too much traffic (source). It’s traffic-espionage.

Google monitored traffic to potential competitors
Google monitored traffic to potential competitors

The documents mention specifically Monster, Hotjobs, live.com, and Myspace. Growing traffic also seems to have been a big motivation to acquire Youtube.

That’s a clear abuse of market power and anti-competitive behavior.

The challenge is that no company before Google ever had access to such data. I think it’s possible that legislators find ways to regulate the usage of such data for own products.

Using clicks in search

We’ve been assuming it for years and every now and then, some devious document would surface from the depths of the web to “prove” that Google uses clicks in Search. Google representatives even confirmed that clicks were used but claimed that was the case outside of ranking algorithms. Now, documents that surfaced in the hearing show how Google used clicks and potentially even clickstream data for Search.

Ongoing investments in user signals are already a strong hint (see below).

Google documentation stating that clicks might be used in Search ranking
Google documentation stating that clicks might be used in Search ranking

But then, the documents clearly state that Google did “build out recommendations based on user data/behavior” and mentions using data from the toolbar and Orkut.

Google mentioning search results driven by user data
Google mentioning search results driven by user data

Google’s algorithm is intellectual property that generated hundreds of billions of US Dollars for Google itself, created SEO as an industry (and gave me a career), and build a vast amount of free tools for everyone to use. Yes, the value Google created is hard to debate. And yet, one man’s value can be another man’s destruction.

Platform Confluence and regulation

Aggregation theory describes the shift of power from supply to demand, growth based on zero marginal cost, and direct relationships with users. Platforms like Google, Facebook, Amazon, and others were able to become aggregators because of the internet.

The next step in the evolution of aggregators is Platform Confluence: the combination of several apps and sites owned by the same company into an ecosystem in which user signals and ads reinforce each other.

https://www.kevin-indig.com/platform-confluence/

In the recent hearings, it became clear to me that Platform Confluence navigates businesses closer to regulation. Most points that Big Tech was criticized for came down to unfair use of the huge troves of data it sits on. That’s only possible when merging the data from many platforms and leveraging it for monetization.