Hi friends!

A few weeks ago, I tweeted about how incredible a business Google Search is.

Tanay Jaipuria @tanayj

Just absurd that Google search is a $150B/yr of revenue business still growing at 44% y/y. The best business ever invented

Their incredible performance prompts discussions from time-to-time around why there aren’t more attempts to chip away at it given the sheer size of the prize and how one might disrupt it, like this thread this past week.

That’s what I’ll go into this week. I’ll cover:

Quick overview of Google Search today
Risk Areas for Google and What Disruptors could look like

Overview

Google Search today is doing ~$150B run-rate in revenue.

The key drivers of the business are: number of ad impressions/clicks and the effective price per impression/click.

This can be further decomposed into the following drivers:

number of search queries
percent of queries with ads displayed
click through rate on ad impressions
the average cost per click charged

In general, query growth is relatively slow (Google already does >100B queries per month), but Google continues to show ads on more queries and optimize the click through rate and cost per click which continues to drive 30%+ revenue growth.

Per Wedbush, as of today, 80% of queries are still not monetized by ads (given not being of a commercial nature), so there continues to be a lot of room for Google to show more ads. On the flipside, it highlights that most queries are not commercial, and that the 20% or so that have more commercial intent are critical.

From a vertical perspective, while Google Search fares well across the board, Retail is typically the largest vertical. Based on their prior earnings call, we have a sense of some of the verticals performing well.

In the third quarter, Retail was again by far the largest contributor to year-on-year growth of our Ads business. Media & Entertainment, Finance, and Travel were also strong contributors.

A couple of interesting data points about Search which give us a sense of the search problem:

On a daily basis, 15 percent of queries submitted have never been seen before by Google's search engine. This highlights just how long the tail is of search (and also just how differently people phrase perhaps similar questions).
In 2020, nearly 2/3rds of queries on Google ended on Google itself i.e., without clicking on an external link (organic or ad). This highlights how over time Google is increasingly showing the results for common queries inline (in its knowledge graph) rather than directing traffic to websites. It also highlights how most searches are not valuable to Google (don’t bother trying to show ads to make money).

Risks and Disruptions

So now that we’re past that overview, let’s discuss some risks that Google faces, and what some possible approaches to disrupt it might be.

I. Too commercial and privacy concerns

Given the drivers of the business mentioned above, its perhaps not surprising that in order to drive up ad impressions and clicks on commercial searches, more and more of the real estate on these types of searches is dedicated to ads.

As an example, for the search query “health insurance”, there is only 1 organic result above the fold, as below.

In addition, in order to generate improved revenue for other businesses and have the most targeted and best performing ads, Google collects and tracks as much data they can about users, including:

Search history
All websites and apps visited
Youtube video viewing history
Location history through Google Maps

Given heightened awareness of and interest in privacy, there may be appetite for more private alternatives.

Now, Google has been investing in more private alternatives, but given the advertising model, there will always be some incentive to collect data to personalize organic results and ads and target ads.

And we've been investing in privacy-preserving technology for many years. Our focus is on supporting developers, small and large advertisers, creators, publishers so that they're able to mitigate impact to their businesses. And we really see the future of digital advertising being built on advances in privacy-preserving on-device technologies which support the free and open Internet and, obviously, a robust ads ecosystem.

So, what might disrupting Google on this basis look like?

The obvious answer is a privacy focused or non-ad-based business model.

DuckDuckGo is a free, privacy-focused search engine which is one take on that and does 100M search queries daily and has roughly a 2% share in the search market in the US. It does also have a keyword-based ad model, but does not personalize or target search results or ads other than based on the search term.

The branding for DuckDuckGo is heavily based on privacy

Another example is Neeva which is a subscription-model based search engine which was founded by former Google execs. Neeva is ad-free and privacy-focused, and allows user to customize search results to their liking.

Neeva highlights the difference between an ad-based vs non ad-based search results

II. Too Horizontal / One-size fits all

Given the sheer scale of Google, it is clear that it is trying to be the everything search engine to everyone.

While they have been building vertical-specific features, such as indexing products for Shopping, they aren’t as focused on either specific use cases or verticals as others might be able to.

There are a few risks for Google and opportunities for disruption related to this.

Vertical Searches start elsewhere

In 2014, Eric Schmidt said this about Google’s competition:

“Many people think our main competition is Bing or Yahoo. But, really, our biggest search competitor is Amazon,” Eric Schmidt, currently serving as Google’s executive chairman, told a crowd in Berlin. “They are obviously more focused on the commerce side of the equation, but, at their roots, they are answering users’ questions and searches, just as we are.”

One of the biggest risks to google is that searches, especially those with commercial intent, increasingly start on another platform. This might mean:

Product searches on Amazon/Etsy
Restaurant searches on Yelp / Doordash
Local Services searches on Angie’s List
Travel searches on Expedia/Hipmunk
And so on

This threat has been true for a while, and so far Google has dealt with it well, typically by countering with products such as Google Flights / Google Reviews, etc, or leveraging the benefit on neutrality (working with all Stores) relative to say Amazon. However, that isn’t to say that Google has lost search share in these verticals. For example, studies indicate that most product searches begin on Amazon.

Digital Channels Where US Adults Begin Their Product Search When Digitally Purchasing Products, Aug 2020 (% of respondents)

And the bigger risk still remains. If startups or other companies can create a better experience for searching within a specific vertical which are 10x superior to the largely one-sized fits all approach of Google, then there may be an opportunity.

One example of this is Deft which is building a better eCommerce search experience which allows for searching in natural language and converting to filters.

Curation over Algorithms

Another angle for disruption or chipping away at share is related to the horizontal, one-size fits all nature is by creating a more curated experience. As websites have started to game search engine optimization, the websites which rank at the top for many searches are because the algorithms like them rather than humans do.

There might be a few potential angles here.

One is as above to pick a vertical, and offer a curated approach for that vertical. One example of this is WireCutter or even The Infatuation, which although are content sites, become the place many users start their “search” for products and restaurants rather than Google.

The other is to somehow integrate people’s ability to vote and perhaps voluntarily curate results in the form of a search engine. While perhaps difficult to make work in Web2, perhaps with the advent of web3 we might see something like this. People could offer curations or updates the ranking of search results / search terms based on their expertise or what helped them during their search, and if these are helpful, they receive tokens in return.

If you don’t yet receive Tanay's newsletter in your email inbox, please join the 3,000+ subscribers who do:

Subscribe now

III. Changing Interfaces

One of the other risk that Google has had to contend with is remaining the go-to for search as the interface to search from changes.

Mobile

In the case of the Desktop to Mobile shift, Google was well positioned thanks to owning Android which reaches over 70% of mobile users. In addition, mobile is still largely the same as Desktop in terms of showing ads, and in fact likely has higher ad CTRs become the entire screen is just ads usually.

However, on mobile, they still have to pony up $5-10B per year (and rising) to remain the default search engine on iOS.

Tanay Jaipuria @tanayj

Google paid Apple an estimated ~$10B in 2020 to be the default search engine on iOS 🤑

Voice Search

With Smart Speaker penetration in the US over 50%, another interface shift that Google has to contend with is some search volume shifting to voice.

Google Voice Search Button is Dead, long live the Google Assistant!

This has given Google two problems:

Not being the default search engine for voice queries and losing search volume: Google has partially countered this with growing share of Google Home and integrating its assistant into many smart devices. However, Amazon for example uses Bing as the default search engine on Alexa.
Ads in a voice model: For voice searches, it is harder to display relevant ads. One way around this is to directly monetise the transaction (i.e., suggesting to book an event / buy a product / make a reservation after the user queries). The other way around this is a structual benefit that likely some if not most commercial queries will not shift to voice long-term, and so the voice queries might be those that didn’t have ads on web/mobile anyway.

AR/VR

Another potential interface change on the horizon might be users wearing a hardware device which provides them additional context in the real world.

While likely still a bit away from actual penetration to pose a threat to Google, again the opportunity to disrupt Google comes in that Google may not be the default on a given hardware, and that this form of searching will require a slightly different set of expertise although quite similar to voice search: understanding and recognizing real world objects, understanding human language voice queries in the context on the environment and providing a single answer which makes it difficult to insert ads.

IV. Changing Media and Data Sources

Media Sources

Google Search is great at indexing most content based websites, but arguably doesn’t do a great job with indexing some of the more user generated content and different media sources such as short-form video.

In addition, given the hyperoptimized content which Google’s algorithm caters to, many are searching for things on Reddit or Twitter instead. Likewise, TikTok is getting so good at understanding users that it is showing them things before they need to search for it, likely somewhat eating into search volume.

Blake Robbins @blakeir

For most queries, Google Search is pretty underwhelming these days. The top results are filled with ads and low-quality results. Appending queries with specific terms like: 'reddit,' 'wiki,' 'substack,' etc. dramatically increases the quality of search results.

While Google does index Reddit and Twitter at times, it doesn’t necessarily optimize for indexing these surfaces. In addition, it doesn’t really index TikTok, where a lot of people are going to these day for things like recipes and recommendations.

What might a search engine which indexes videos like TikTok and user generated websites at the atomic level (comment / post level) rather than webpage level look like? Or one where it automatically includes Tweets from the people you follow in search results.

Fragementation of Data Sources

Today, Google largely indexes the public web. But often when searching for items, the best results might be contained in internal work documents or in email.

Especially in a work context, the data sources that might contain answers to queries are numerous and search is quite fragmented. People are searching through email, Google, their docs suite separately.

While Google is well positioned in this space, there likely is an opportunity to become the default starting point for all searches in the enterprise, which is able to scan through internal documents and tools, emails and so on and the public web. It might even allow companies to set preferred websites of search results to return or allow users in the company to preference certain results based on what typically helps them find what they’re looking for in a work context.

Thanks for reading! If you liked this post, give it a heart up above to help others find it or share it with your friends.

Share

If you have any comments or thoughts, feel free to tweet at me.

If you’re not a subscriber, you can subscribe below. I write about things related to technology and business once a week on Mondays.

Subscribe now

Tanay’s Newsletter

Disrupting Google Search