Poised for a Pivot: Google Moving from “AOL v.2020” to “Amazon v.2025”
What We Learned by Answering the Questions from Congress that Google Wouldn’t
In December 2018, Google’s Sundar Pichai testified to Congress about several aspects of Google’s business. Although it wasn’t quite the high-profile media cluster-fest that Mark Zuckerberg’s testimony was, there were certainly some similarities.
One common theme: Congressional bafflement about how the Interwebs work. Senator Orrin Hatch asked Zuckerberg how Facebook remains free (“Senator, we run ads”). Representative Steve King asked Pichai about notifications on his granddaughter’s iPhone. Pichai shifted uncomfortably in his chair before responding: “Congressman, the iPhone is made by a different company. And so, you know, I mean…”
Politicians are easy targets, so in fairness, I should point out that many good questions were asked, highlighting another commonality: both Zuckerberg and Pichai declined to give precise answers to some key questions. I’m not shaming either of them. Leaders of publicly-traded companies have responsibilities to their shareholders and their businesses – responsibilities that often include legally keeping some information confidential so that competitors can’t use it. That’s Business 101.
Still… Inquiring Minds Want to Know
But I personally remain intensely curious about fundamental questions Congress asked – questions that cut to the core about how to thrive in today’s evolving digital eCommerce industry and ecosystem, as well as Google’s strategy for the future.
So along with my Jumpshot colleague Steve Levay, our team analyzed 24 billion searches conducted by consumers in the US since January 2016. This particular analysis focused on searches on desktop computers as well as Android mobile phones; voice search, app search, and iOS devices were not quantitatively included, although we do address their impact in our interpretation.
To aid in that interpretation, we partnered up with the two most knowledgeable people in the world of consumer search: Britney Muller (Senior SEO Scientist for Moz), and Rand Fishkin (Founder of Moz, and more recently, SparkToro). Rand wrote up a great summary of this work on his blog here. They have been studying search engine optimization since, well, before it was called search engine optimization, and they write and speak widely on all things search. As a quartet, we set out to answer the Congressional questions that Google wouldn’t. The four of us recently held a webinar, Google Uncensored: How Hidden Trends in Queries, Clicks, and Consumer Behavior Are Redefining Search where we tackled this very topic and took a deep dive into never-before-seen online retail market data.
Unanswered Question #1: How Much Search Is There?
There were two particularly crucial questions that Congress asked in their written follow-up questions. The first: “What is the average number of search engine results per day for the last ten calendar years 2008–2019?” Google demurred. Their answer begins, “While we do not have detailed public metrics, we can confirm that every year, there are trillions of searches on Google…” before quickly pivoting to a discussion of their mission statement.
Jumpshot data suggest that on Google proper there are, on a typical day in the US, about 789 million desktop searches, and 925 million Android searches — figures that rise 25-30% if you include searches on Google’s extended family of Google Images, Google Maps and YouTube.
The number of Google web searches is actually trending down – about 10-15% since January 2016. It’s hard to quantify if that reflects a true decline in consumer engagement with Google per se, versus a shift of some search activity to the Google app and Google Home devices. My (admittedly subjective) take is not that Google is being used less, or is becoming less relevant in people’s lives. On Android phones, for example, our data show that use of the Google app is strong, growing, and well ahead of other search apps. In May 2019, 82% of Android phones were used to access the Google app, up from 77% just since January 2019; for other search apps, this figure is less than 5%. Google app users access it four times a day, twice the rate of other search apps.
Which leads to the next question: What is Google’s share of search? Again, Google declined to give hard numbers (“We work in a highly competitive environment…”). We find that, on desktop, Google gets 61% share of searches, up from 52% in January 2016; on Android, Google gets 76%, up from 65%. The share of searches for Google’s extended family is about 90% on desktop, and 95% on Android, both figures essentially unchanged over the past two-and-a-half years.
Recall that Google’s raw volume of searches is down — so if their share is stable or rising, then it follows that competitors are losing even more volume. And that’s precisely what we found. Bing’s search volume is down about 60% since January 2016, and Yahoo’s is down about 45% (and both are tiny compared to Google, getting less than 5% of Google’s search volume).
Two search competitors stand out. The first is Amazon, which has surpassed legacy search players Bing and Yahoo to be a distant #2 behind Google. Amazon gets 38 million desktop searches a day and 18 million on Android. Think about that. Amazon search volume is 5% of what Google gets on desktop, and 2% of what Google gets on Android — and that’s enough for second place.
Of course, the nature of search on Amazon is different from a search on Google. Search on Amazon is more product-related and lower-funnel. Search terms on Google are more likely to be branded and reflect that consumers often use Google simply to get to their most heavily-used sites.
Top 10 Keywords for 2H 2018
|Top search terms on Google||Top search terms on Amazon|
Even within the more specific area of product search, Amazon and Google get qualitatively different kinds of search. Specifically, Google gets more branded search, while Amazon gets more unbranded searches focused on generic descriptors in utilitarian categories. For example, Amazon gets most of the search volume for “headphones” and “earbuds,” while Google gets more searches related to “Sony headphones,” “Bose noise-canceling headphones,” or “Airpods.” In women’s fashion, the top Amazon searches are generic terms like “leggings for women,” while the top searches in the category on Google are brand names such as Forever 21 and H&M.
Another search player worth noting is the house that anonymous search built: DuckDuckGo. Again, its volume is a tiny fraction of Google’s, but its growth rate has been impressive.
Unanswered Question #2: Where Does Google Search Traffic Go?
In the words of Congress:
“What percentage of those search engine result pages resulted in the user:
- Terminating their search because they were satisfied with the answer?
- Clicking on a link which led to the user a [sic] webpage not hosted on Google.com?
- Clicking on a link when led to the user being sent to a secondary page on Google.com, any of Google’s top-level domains, or generating another search engine result on Google?
The phrasing of Option A opened a loophole for Google to avoid answering the question they weren’t going to answer anyway. Google’s response: “Google doesn’t necessarily know why users terminate a search. Some users may be dissatisfied by the answer. Others may have found what they needed…” By the end of the paragraph, they are talking about their mission statement again. Fair enough, but if we drop the phrase “because they were satisfied by the answer,” then the question becomes very answerable.
Our data clearly show that “zero-click searches” are on the rise. On desktops, 35% of searches don’t result in a click, up from 31% in January 2016; on Android phones, 62% of searches don’t result in a click, up from 56% in January 2016. Why? As Google pointed out, we don’t definitively know the “why” (as is often the case with behavioral data). But it is reasonable to assume that Congress was right in implying that many searches don’t result in a click because users got the information they wanted. (Paid click-through rates are also rising, but that’s a subject for another post).
Certainly, consumers are “getting better” at search. Today consumers ask more specific questions in search, often using natural language that provides clues to their true interests and intent (and their place in the purchase funnel). More generally, searches that are personal, immediate and local are all on the rise. In a sense, consumers no longer search for “car insurance,” but rather “What is the best kind of car insurance for me?” Better questions, coupled with better AI to infer intent, lead to more satisfying results.
Zero-click searches are likely also rising because Google is getting better at delivering search results that meet consumer needs. Certainly, search engine result pages (SERPs) look much “busier” than they did a few years ago. Our colleagues at Moz track this closely, finding that the number of features on a typical SERP are up sharply. Featured Snippets, Knowledge Panels and local/map content increasingly answer basic questions without necessitating a consumer click-through. The prevalence of “People Also Ask” panels has risen 10X over the past three years, and is now in 85% of SERPs. This reflects Google leveraging its strength (search) to keep people engaged with Google, and by extension, seeing more Google ads.
(Note that HTTPS (secure) links are now on virtually every SERP. This may reflect the growth of online shopping related links, but it’s hard to interpret because Google has encouraged webmasters to make every link a secure link).
|People Also Ask||8%||85%|
Look at this example of a Google SERP for “Golden State Warriors.” Above the fold are recent scores, the upcoming schedule, and a summary of the team’s history.
Scroll down, and you’ll see headlines from top news stories, links to videos, and more. And why would anyone leave? It’s a wonderland of content-filled goodness. Traditional links to organic search results — Google’s historical bread-and-butter — have become much less prominent.
Of course, content-rich SERPs give Google more opportunities to send traffic to its own properties. For example, about 90% of video search referrals from Google end up on YouTube; very few of them end up on Facebook, Vimeo or other video content sites. And Google is sending less traffic to other sites more generally. Today the 500 largest domains on the Internet get 38% of their traffic from Google, down from 42% in January 2016.
Google Today: AOL v.2020
It’s hard to miss the subtext of the Congressional questions — they get to the heart of issues related to monopolistic behavior. Google has been fined by the European Union for antitrust violations in the past, and narratives about the issue are heating up in the US as the 2020 Presidential election looms. So it’s easy to see why Google wouldn’t volunteer more information than necessary.
But I have a different question: what does all this data tell us about Google’s business today, and their strategy for the future?
In short, Google has become what AOL was trying to become twenty years ago.
I hear the collective “huh?”
Today’s content-rich Google SERPs looked familiar, but it took me a while to put my finger on it — they look like America OnLine’s interface from 20 years ago. Google and AOL evolved similar interfaces because they have a shared aspiration: to become an ecosystem that is content-rich, visually-appealing, and carefully-curated to meet consumer needs so well that consumers don’t have to leave for the wild, wild west of the Internet at large.
Clearly, Google is trying, and succeeding, in keeping visitors on their properties. And when visitors leave, Google is getting better at monetizing them on their way out. Not only are zero-click searches on the rise, but so are paid click-through rates. Historically paid traffic had a reputation of being lower in quality (and conversion rates) than organic traffic. Today that’s not necessarily true. Paid click-through rates are rising, and sites such as Chewy and Wayfair have driven significant eCommerce sales growth from search traffic that is 60-80% paid.
Google Tomorrow: Amazon v. 2025
Another commonality between Google today and AOL from 20 years ago: the desire to capture transactions within the ecosystem. Both are platforms, aspiring to be toll-taking conduits for known retailers and brands. Except for some Google tech devices (Home, Chromecast), neither aspired to create their own products/brands, or to become an Amazon-like retailer/brand combo.
Google being AOL from 20 years ago is kind of “peak Google” — it’s leveraging its core model as much as it can. Finding growth will mean looking elsewhere, and the eCommerce industry is the most fertile ground. Meanwhile, Amazon has made serious inroads into Google’s core business, which Google must respond to:
- A majority of lower-funnel, product-related search now happens on Amazon
- Alexa has about two-thirds of the smart speaker market and is better poised to monetize it
- Amazon has emerged as a distant but fast-growing #3 in the digital advertising market (behind Google and Facebook)
So look for Google to move down-funnel — more directly competing with Amazon for lower-funnel search traffic, and for transactions. Expect Google, to begin with, categories — like travel — which are outside of Amazon’s utilitarian strengths.
While Google was cagey in their responses to Congress about search, they have been quite blunt about their eCommerce ambitions. NetElixir CEO Udayan Bose recently wrote an article in MediaPost, summarizing a conference presentation by Allan Thygesen, President, Google Americas. In part, Bose wrote:
“Google is making a strong move to… win back parts of the funnel that it owned just a few years back… Google plans to own the search-shop-buy funnel… and corner the market on conversions… Google’s drive for funnel domination is supported by its strong 8-member ecosystem (Google, Gmail, YouTube, Android, Images, Display, Shopping, Assistant)…”
That about sums it up. Google today is the peak form of the old AOL model: leverage the ecosystem, curate the experience, meet consumer needs so they don’t have a reason to leave, and better monetize them if they do leave. To find growth, they’ll increasingly go downstream, largely with retail/brand partners, to play in the eCommerce space more directly. If you want to learn more about searcher behavior and search opportunities, check out the webinar, Google Uncensored: How Hidden Trends in Queries, Clicks, and Consumer Behavior Are Redefining Search.