Google’s Project TACless: Paying the Piper (or Not)
The Monty Hall Path and the search market’s alternate universe
I hesitate to wade into another discussion about customer acquisition costs (CACs), because it really has all been said before — and well said, for that matter. But, there’s an alternative reality that’s been eating at me: The last year’s worth of quarterly earnings reports made me wonder whether or not, at this point in the game, it’s really worth the massive traffic acquisition costs (TAC) that Google pays to platforms like Apple?
Google long-ago crossed the transom in a leap from early adopters to the mass market, and now it’s probably wading deep into the laggards. Yet, Google’s TAC isn’t as much about reaching untapped late adopters as it is about continuing to serve its preexisting userbase anywhere and everywhere. And that’s what interests me the most…
Fear of the unknown
The TAC Google pays to Apple alone is expected to clock-in at an estimated $9B to $12B this year. While the actual dollar figure may even exceed that consensus estimate, the important thing to note is that its total TAC has remained steady at 23% of ad revenue for the last three years — suggesting that the agreement with Apple has transitioned from a flat fee (nominal dollar amount) to a percentage fee (royalty rate).
Regarding my question about whether or not Google’s TAC is worth it, my answer at first blush was emphatically “yes”: All things equal, paying TAC probably yields Google’s highest probability-weighted expected value. And, from the perspective of sound management principles, there’s of course a risk-aversion factor too — continuously paying to keep those users in its ecosystem is the bird-in-hand.
After all, the greatest risk to supermassive demand aggregators like Google is not displeasing advertisers, but rather displeasing users. Advertisers somewhat have to stay on the platform as long as Google has consumer attention, so an aggregator like Google should never do anything to provoke end-user demand to leave en masse.
Surely Google wouldn’t voluntarily risk such a life-threatening user exodus when it has the option to avert the possibility entirely, right? TAC gives Google the discretionary choice to resume its smooth sailing at the expense of mere gross margins, right? It’s only money in the end, right? 😋
The Monty Hall Path
Well, that all said, the most interesting consideration is how Google’s TAC decision pits two pillars of tech’s academic business strategy against one another: Aggregation Theory vs The Power of Defaults. Whenever their two paths shall meet in the wild — a fork-in-the-road so to speak — the dualism between these generally accepted theories reminds me of the paradox: “Can God create a stone he himself cannot lift?”
Let’s call this “The Monty Hall Dilemma”, a take on the classic Monty Hall Trilemma or Monty Hall Problem…
Behind door 1 (Defaults): Should Google not renew its TAC contract with Apple, thereby forfeiting its status as iOS’s structural default, the long term consequences would be dire, because consumers are lazy and will stick with the new default out of inertia — or so suggests the theory of digital “defaults”.
Behind door 2 (Aggregation): At the same time, if we all agree that network effects’ virtuous cycle helps Google remain the search Super Aggregator today, then, by definition, Google currently serves the marketplace with the superior user experience and its marketshare should be pretty secure even were it to let its TAC contract with Apple lapse — or so says the theory of digital “demand aggregation”.
Which one is it?
To answer with greater precision, there really are two distinct timeframes that need to be considered. The first is the rise of mobile, in which “The Power of Defaults” empirically trumped “Aggregation Theory”. Back then, Google averted new market disruption at the hands of the nascent mobile epoch by parlaying its surplus profits from desktop/web dominance into effectively purchasing a foothold in mobile — bought and paid for via R&D, M&A, and, of course, pseudo-partnerships like its TAC deal with Apple discussed herein. In other words, Google made sure to inject itself into the mobile marketplace as the default search provider when it could still afford to pay-for-the-privilege, because its network effects alone wouldn’t have necessarily been enough to overcome a rival in that new surface area. Consider the counterfactual, for example, were Apple to have implemented its own, integrated, proprietary Safari search engine and a beefed-up Spotlight Search feature for iOS. Indexing the open web and (relatively) closed apps for deep search would have disintermediated Google’s relationship with end users — ultimately usurping Google’s scale advantage.
In contrast, the second timeframe for consideration is here and now, when the mobile market is fully matured. Today, Google is already diffused throughout mobile thanks to that aforementioned R&D, M&A, and, of course, partnerships. It has a massive installed base that’s not just limited to 3rd party feeders, like its Apple TAC, but also owned and operated properties like Android, Google.com, the Google App, Gboard, etc. Incorporating this dominance with that mature market, Aggregation Theory should now trump The Power of Defaults. (Again, “If we all agree that network effects’ virtuous cycle helps Google remain the search Super Aggregator today, then, by definition, Google currently serves the marketplace with the superior user experience and its marketshare should be pretty secure[.]”)
In sum, during the rise of mobile, power initially accrued to the default services that new users discovered when they first arrived in the new surface area; but, now that mobile has matured, power accrues to the biggest aggregators in their respective verticals, including Google for search. Instead of framing this as The Monty Hall Dilemma — a binary choice between Door 1 and Door 2 — it’s more like The Monty Hall Path — a winding road to the Promised Land that first led Google through Door 1 (The Power of Defaults), then eventually wound its way through Door 2 (Aggregation Theory).
Thus, in theory, the script has flipped and Google should not have to keep paying Apple for the privilege… right?
Project TACless
I keep referring to an alternative universe in which Google renounced these TAC taxes. Let’s start referring to that prospect as “Project TACless”.
Digging deeper, The Monty Hall Path somewhat addresses even more granular, existential questions regarding Project TACless:
- Is Google Search still truly commoditized and/or modularized by mobile?
→ Are there “good enough” alternatives to Google? (e.g. Bing or DuckDuckGo)
→ How much of Google’s traffic is willing to seek-out the service were it reverted from a default back to a destination? - How much would removing Google as iOS’s search default detriment Apple’s highly vaunted “superior user experience”?
→ Is search even important enough to today’s mobile UX that Apple should care? - How many iOS users would manually switch their OS defaults back to Google services were Apple to select an alternative search default?
→ How many users would need to switch back for Google to break-even? (i.e. TAC savings net of revenue loss)
→ Do users even care about their OS default, given their other options throughout Google’s ecosystem? (e.g. Google App) - How many users would discretionarily switch their hardware to Google (e.g. iPhone to Pixel) were Apple to elect an alternative search default?
→ Again, how many would need to switch hardware for Google to break-even?
→ How many would need to switch hardware for Apple to cry uncle?
Neatly, those questions consider the tradeoffs for all parties involved in this market’s supply chain, starting with an evaluation Google’s strategic considerations as the producer (#1); then moving on to evaluate the Apple’s optionality as the distributor (#2); and finally covering the end user’s disposition as the consumer (#3 and #4).
Were we able to answer those questions at all, we’d first need to take a hard look at Google’s ecosystem beyond Android and Google.com — specifically “Trojan Horses” like the Google App, Google Assistant, Gboard, and the rest of the Google constellation that tries to meet users wherever they live. These represent Google’s emergency arteries available to Apple users were Project TACless to close down all other avenues like Siri and Safari.
For example, if iPhone owners have a question or need to search for something on iOS, a small number of them would use Gboard (because it can be set as a default keyboard that’s essentially everpresent on their iPhone). A handful would ask Google Home when within earshot (because smartspeakers are available almost everywhere that users don’t have their phones handy). A few more users would browse to Google.com in Safari, although many have the option to change Safari’s default search engine from Apple’s new hypothetical partner back to Google. Finally, the majority of users would open the Google App (because it’s fewer keystrokes than opening Safari to search even if you’re already in Safari).
The modal recourse therein is the Google App, which already has 61% market penetration in the US, accounting for all mobile operating systems. Similarly, the broader Google Search service in aggregate has 63% US marketshare of search volume. (A more commonly cited 90%+ marketshare statistic is Google’s worldwide take.)
These Trojan Horses are great for defending Google’s relative marketshare in search volume, and walletshare in ad spend will always follow consumer attention. Advertisers got to advertise, as we’ve seen not only throughout history — with advertising as a percent of GDP having remained within 1.0–1.5% over the last century — but also recently — with the deceleration in ad load across Google and Facebook properties having been offset by higher CPMs/CPCs/CPLs.
So, in the absence of an Apple TAC deal, it’s reasonable to assume that Google’s ecosystem is resilient enough, from a business perspective, to maintain revenue growth amidst a marginal drop in usage. Luckily, we have the Apple Maps case study as an empirical example to establish some base rates — from “Apple’s Chinese Wall of Worry”:
The Chinese Wall solution… delineates between kosher integrations by proprietary aggregators and non-kosher infringements upon developers by principals-cum-agents… Let’s use Apple Maps as our guinea pig. It’s the perfect example because Apple Maps did not — and still has not — improved the user experience relative to its predecessor, Google Maps, which was the iPhone’s default mapping app from the launch of the App Store in 2008 until Apple’s encroachment in 2012. So, not only was this an example of vertical foreclosure, but it was also one that poorly served consumers… Google Maps-as-a-service was fully integrated with iOS for the whole of iPhone’s history — up until the point where Apple Maps gatecrashed it.
In 2012, Apple rug-pulled its longtime iOS default mapping partner, Google Maps, and inserted its own proprietary Apple Maps as the iPhone default. Nonetheless, to this day, Google still maintains an even stronger marketshare in mapping than search:
And the moats around the former (e.g. mapping’s capital-intensity) are likely less durable as those of the latter (search’s aggregation flywheel), which only strengthens the validation of The Monty Hall Path provided by this Apple Maps vs Google Maps precedent.
While there’s no doubt that some users would stick with iOS’s hypothetical new search default out of inertia — whether Bing, DuckDuckGo, or another — the real question is where the tipping-point lay for Google. As Marc Andreessen said:
The problem with network effects is they unwind just as fast. And so they’re great while they last, but when they reverse, they reverse viciously. Go ask the MySpace guys how their network effect is going. Network effects can create a very strong position, for obvious reasons. But in another sense, it’s a very weak position to be in. Because if it cracks, you just unravel.
What constitutes a fatal crack in the Google Search edifice? What amount of traffic loss will flip Google’s high advertising demand inelasticity into elasticity? At what quantum of added friction does its high user demand inelasticity turn highly elastic?
From “Spotify and Ek’s Parlay”, which I’ve retrofitted to Google:
That said, consumer demand isn’t binary or linear; it’s nuanced and convex... That dynamic is due to the elasticity of [Google’s] demand. [Google] has proven that its demand curve is both highly and increasingly inelastic thanks to its users’ increasing dependence on [its ecosystem]. If you take that to the most cynical extreme (i.e. [search is] undifferentiated and perfectly substitutable among competing platforms), then personalization becomes the preeminent point of differentiation — in which case [Google’s] demand would be even more inelastic, since users’ sunk costs accrue as time passes.
Perhaps the better way to frame that concept is ‘growth accelerates exponentially on the way up, and losses accelerate exponentially on the way down’. Both start as a trickle, then gains or losses snowball as momentum builds among users who are entering or exiting the network.
Whether symptomatic of a sunk-cost fallacy or a deeply enmeshed utility, the stickiness of Google’s user experience is a real barrier-to-exit for users. Google VP Aparna Chennapragada really drove-home this point during Google’s recent I/O 2019 developer conference (emphasis mine):
What you are seeing here is text-to-speech, computer vision, the power of translate, and 20 years of language understanding from search, all coming together… The power to read is the power to buy a train ticket, to shop in a store, to follow the news. It’s the power to get things done, so we want to make this feature accessible to as many people as possible.
In other words: ‘You have a lot of history here with Google, and that history compounds into a lot of downstream products and services that facilitate your work and play; so, if given the choice, you are not going to leave behind that ecosystem and disturb those benefits.’
Again, this TAC is effectively a 23% tax on Google’s ad revenues. Rather than key-in on the absolute dollar figure “$9B to $12B”, the “23%” is a better number to focus on, because Google’s TAC has remained remarkably steady at that percentage across quarters, suggesting that its deals are structured more like royalties than flat fees — almost like a franchise fee for Apple, in particular.
With that in mind, take a step back. The combination of Google’s Trojan Horses, high demand inelasticity, and that 23% provides a pretty wide margin-for-error were Google trying to establish a breakeven for Project TACless. Yes, there’s that dangerous tipping-point where losses start accelerating exponentially; and, yes, there’s customer lifetime value (LTV) to consider above and beyond discrete revenue loss; but these adjustments to the model pale in comparison to the resilience of Google’s business were we to add Google’s owned and operated properties to the equation — including Android, Google.com, and AdWords.
If this were such a slam-dunk, we have to ask why Google hasn’t given Project TACless to go-ahead…? Of course, maybe it just isn’t a slam dunk. Or, as I mentioned in the beginning, since risk-aversion is often synonymous with sound management for mature companies — the other side of the Innovator’s Dilemma coin — maybe Google is just poised to pull-the-plug on TAC when the time is right. I can envision two possible openings:
- Mobile growth slows and/or a new epoch emerges; or
- Another one of Alphabet’s moonshots emerges as a cash cow to supplement the Google Search franchise
Until then, as they say, you don’t kill the goose that lays the golden eggs!
I know we all want certainty, absolutes, guarantees, conviction; so perhaps “it depends” is a deeply unsatisfying answer to the questions I set forth herein. However, in arriving at these conclusions, we’ve established not only the resilience of Google’s core product, but also a base-case expectation for the future of billions in cash flow between Google and Apple — and, perhaps most importantly, a new strategic framework in “The Monty Hall Path”. 🏁
Another fundamental web and mobile utility
I just didn’t want to waste everything I read, and once I had a way to save the good stuff, I naturally wanted — or in some cases needed — to share it. That was just the beginning. Now you can join the network where one person’s annotation is another person’s summarization, so you and yours can get straight to the point. Check out “Annotote”, your antidote to the information overload:
