“Sign In with Apple” and digital advertising

Welcome to Marketing BS, where I share a weekly article dismantling a little piece of the Marketing-Industrial Complex — and sometimes I offer simple ideas that actually work.

If you enjoy this article, I invite you to subscribe to Marketing BS — the weekly newsletters feature bonus content, including follow-ups from the previous week, commentary on topical marketing news, and information about career opportunities.

Thanks for reading and keep it simple,

Edward Nevraumont

How do you like them Apples?

Last Monday (June 3, 2019), Apple kicked off their annual Worldwide Developers Conference with a number of significant announcements: the retirement of iTunes, the introduction of iPadOS, and the release of a standalone App Store for the Apple Watch. But for marketers, the biggest surprise was ‘Sign In with Apple.’ From Stratechery:


Sign In with Apple
: developers can now let Apple handle identity instead of Facebook or Google. Furthermore, users creating accounts with Sign In with Apple have the option of using a unique email address per service, breaking that key link to their data profiles, wherever they are housed. [Emphasis mine]


What is so significant about allowing unique email addresses?
Essentially, this feature could eliminate the ability to track information about users of the apps — which, of course, companies collect for marketing purposes. The impact of Apple’s new initiative would depend on the number of developers that choose to allow ‘Sign In with Apple,’ instead of the many current alternatives, such as email, Google, Facebook, etc. Apple obviously anticipated the possibility of being left out, so they have proactively taken a giant step toward popularizing their new feature. From the Apple developer guidelines:

Sign In with Apple will be available for beta testing this summer. It will be required as an option for users in apps that support third-party sign-in when it is commercially available later this year.


Basically, if an app wants to be listed in the App Store, it MUST include an option to use ‘Sign In with Apple.’ If that option is then selected by the consumer (and you have to assume a significant number of people might), it would eliminate the ability for companies to track the user for advertising purposes. Many marketers are freaking out about this new development. AdExchanger quoted three agency CEOs:

This is part of a continued overall trend where brands will need to prioritize the customer experience over a growth-at-all-costs approach… A lot of these changes will be painful for some brands in the short term, but will force them to build the proper muscle memory around data, privacy and customer experience. (Michael Katz, mParticle)

...on the margins it will reduce the clarity of the user’s identity. However, if Apple were to also deprecate the IDFA and prevent anonymous identity from being used for ads in apps, then the loss of the login would be pretty critical... If Apple removes the IDFA, everything changes. (Ari Paparo, Beeswax)

When it is also made available for website login, that will impact co-op-style solutions for cookieless identity for ad tech as well as consumer recognition tools that lift performance of cart abandonment flows for mar tech. I’m interested to see if publishers begin rejecting email addresses that end with apple.com. (Steven Francolla, AirDXP)

On the surface, this news is bad for Google and Facebook, whose current sign-in options appear ubiquitously across the internet. The ability for consumers to use ‘Sign In with Apple’ will almost certainly erode Google and Facebook’s sign-in dominance, as well as the value of their trove of information about users’ habits. Moreover, marketers will face challenges with retargeting strategies for anyone using ‘Sign In with Apple.’

But will ‘Sign In with Apple’ really be the game-changer that some people predict? This question ties back to the study I quoted in my previous newsletter:

In one of the first empirical studies of the impacts of behaviorally targeted advertising on online publishers’ revenue, researchers at the University of Minnesota, University of California, Irvine, and Carnegie Mellon University suggest publishers only get about 4% more revenue for an ad impression that has a cookie enabled than for one that doesn’t. The study tracked millions of ad transactions at a large U.S. media company over the course of one week.

That modest gain for publishers stands in contrast to the vastly larger sums advertisers are willing to pay for behaviorally targeted ads. A 2009 study by Howard Beales, a professor at George Washington University School of Business and a former director of the Bureau of Consumer Protection at the Federal Trade Commission, found advertisers are willing to pay 2.68 times more for a behaviorally targeted ad than one that wasn’t.

All this behavioral targeting isn’t really helping that much. Behavioral targeting is a marketing gimmick that advertisers are willing to pay significantly for (2.7x more per impression), despite being only marginally better at generating revenue (+4%). Apple’s development of a sign-in feature is going to reduce the effectiveness of behavioral targeting, but only on the margin. Maybe it drops from +4% to +3%. When you are already content to pay more than 2x for the privilege of a +4% increase in sales, moving to +3% change probably won’t change your behavior.

In related news, The Verge posted this information last week:

Firefox will now block thousands of web trackers by default, protecting users from many websites, analytics companies, and advertisers that want to follow their paths across the web. The change should speed up the browser and keep users’ web habits more private, while nudging advertisers toward less invasive practices.

This change is stricter than the Chrome browser, but still more permissive than Safari:

Apple’s browser blocks nearly all third-party trackers by default, rather than just known trackers collected on a blacklist. Apple also limits trackers from being used by third parties at all if you haven’t interacted with the website they originate from in a full day.

Given everything we hear about how terrible it is to invade the privacy of users by tracking them around the internet, why didn’t Firefox match Apple? It makes sense that Google — who derives some benefits from user tracking — would not block these ad trackers, but why wouldn’t Firefox block everything for a better user experience? An explanation:

The company found that blocking all cookies “leads to scenarios where some websites may not function properly,” and so it chose this partial approach to prevent “potential usability issues.”... Apple’s approach goes further to preserve privacy, but it may also mean more headaches for users. Many pieces of the web rely on cookies, a key tracking tool, to keep people logged in or serve them relevant information. By aggressively blocking cookies, Apple risks disrupting the experience on some websites.

There is a real trade-off between privacy and convenience. Many people SAY they care about privacy, but their actions indicate a ready willingness to give it up for even tiny increases in convenience. Apple’s focus on privacy is a marketing play (that doesn’t cost them much), in the same way companies talk about their environmental credentials. People like to hear about goodwill gestures, but most users will not pay more for them (in dollars or their convenience).

Why does this matter? Who wins?

Advertising spend on the internet is currently divided four ways:*

  • Google search ads (and Bing)­ — 41%

  • Facebook paid social ads (and Instagram) — 22%

  • Amazon display ads — 9%

  • Everything else — 28% (of which the largest is Verizon at 3%)

Understanding why digital advertising has become so concentrated in two — soon to be three — players requires some context.

Google holds a near-monopoly on search ads, but people need to realize that search ads are not really about marketing — they are about distribution. Search ads are not an effective way to raise awareness of your product — and especially not your product category. But search ads ARE effective when you want to be featured right at the moment when someone is considering a purchase. Google search ads don’t replace television advertisements (or at least they shouldn’t); search ads replace the paying of retailers for the right to be displayed on endcaps— in a world where there is only one endcap for each product on the planet.

Facebook ads work for two reasons. The first is the “ad unit” concept. When the digital world was transitioning from desktop to mobile, analysts (and Facebook executives) were concerned that the company could not make the shift without massive loss of revenue. Mark Zuckerberg’s team had figured out how to monetize eyeballs on large desktop monitors, but there didn’t seem to be a good location for advertisements on small smartphone screens without significantly damaging the user experience. With a stroke of genius, Facebook created “newsfeed ad units” that closely resembled the posts from real friends; they quickly became the most effective “display ad unit” by miles. Most ads on the internet are ignored — they are practically never seen. Because newsfeed ads look just like the rest of your feed, they almost NEED to be seen and read in order to be ignored. That makes them dramatically more effective than anything else in their category on the internet.

Media attention focuses on the “creepy” targeting available on Facebook, but the primary reason the company has soaked up 22% of all online spending is the newsfeed ad unit. Although Facebook’s targeting capabilities do, in fact, make the ads more effective, the ads themselves perform at such a high base rate that their targeting is extra and not essential (apart for simple targeting tools like hitting people in a specific city for local events, or advertising men’s shoes to men, or hip replacement options to seniors).

Until a few years ago, the most effective way to target on Facebook involved the formation of “interest” segments of roughly 5000 people. Advertisers would run different ad copy against these small groups and, using trial and error, deduce which groups best responded to specific ads. You might discover, for example, that your pet food ad appealed to twentysomething women in Chicago who were interested in “animals” and “People Magazine.” There was an art to choosing the right segments to test, and then a science to inform you which ones actually worked. Facebook also offered a product called “Lookalike Audiences.” You could upload your list of customers (with actual email addresses) to the Facebook ad platform and the company would match your information to their own database of users. From there, Facebook would use AI (before people were calling it AI) to find other users in their database with similar characteristics. There was one major problem: it didn’t work very well. Sophisticated marketers tried the process and then went right back to their 5000-person interest groups.

Sometime around late 2015, the situation began to change. All of a sudden, Lookalike Audiences started outperforming targeted interest groups. Why? I have never found concrete proof of any clear explanation, but I believe it was due to growth in the “Facebook Login” option. By 2015, a staggering 1.5 billion people were using Facebook’s social login to save themselves from having to create yet another user account. Merchants loved the increasing prevalence of social logins, because it reduced friction and increased sales. Today, there are two billion people using social logins, with Facebook controlling about 72% of these accounts (Google is a distant second at 19%). By owning the accounts, Facebook knows users’ purchase behaviors and histories. And from everything I know about targeting, actual spending activity is a MUCH better predictor of future behaviour than stated or inferred interests.

With the use of Facebook Login, the company was FINALLY able to utilize targeting in ways that everyone only THOUGHT they could do before. Reality caught up with expectation.

Let’s take a moment here to reflect on what DOESN’T matter. Facebook’s incredible targeting success is fuelled by their social login and knowledge of your purchase behaviour. NOT the interests you claim on your Facebook profile. NOT the cookies that follow you around the internet. NOT even the ads you have interacted with on Facebook in the past. All of that stuff might matter on the margins, but it is not the driver of the effectiveness of Facebook ads. Facebook ads are successful because the ad unit is awesome — they can target you based on your geography and gender and your historical purchases. Everything else is noise.

What about Amazon?

In Amazon’s financial statements, they don’t break out advertising specifically, but they do include a line item for “other,” which they have said, “primarily includes sales of advertising services, as well as sales related to our other service offerings.” Using “other” as a base, analysts have estimated Amazon’s advertising revenue at $1.4B in 2016, $2.8B in 2017, and then a jump to $10B in 2018, and 2019 is currently 50% higher than 2018. What the heck happened?

If Facebook targeting is so effective because they control 72% of the social logins on the internet, then the fact Amazon controls over 50% of ALL e-commerce sales, suggests they might be pretty good at it as well. They are.

Where Amazon struggles is the ad unit. If a company is selling a product on Amazon, there is no better ad unit than a placement at the top of the listings. This is, essentially, Amazon’s equivalent to Google Search Ads. Remember, though, that companies advertising via Amazon search ads are not really advertising — they are buying distribution. The top listing in an Amazon search is the company’s virtual endcap; it’s a fair thing to do and it’s easy to calculate the ROI. Amazon search ads are no more competing with Google search ads than Walgreens’ endcaps are competing with CVS’s endcaps. Yes, the two companies compete for user searches, but if you are selling your product, you want to locally maximize your distribution in both places. There is no trade-off from one to the other.

A more interesting concept involves using Amazon’s ad network to build awareness of your product. In this way, Amazon IS competing with Facebook (and Google Display Network and television and radio and print and every other marketing channel you can imagine). Amazon’s ad units are nowhere close to the effectiveness of Facebook’s newsfeed ads, but Amazon possesses a clear superpower: their targeting is much, much better. Ad units may change over time, but the data used to support their targeting? That is a fundamental advantage that’s not going away.

Where are we now?

After that lengthy digression into advertiser effectiveness, we can circle back to analyzing the impact of the changes announced last week by Firefox and Apple.

As a reminder, Firefox plans to limit the ability of advertising networks to track users across the internet. Who does this hurt? It doesn’t hurt Google search ads at all. As long as a user is logged into Amazon, it doesn’t hurt Amazon at all. It might hurt Facebook (marginally), but remember that most of Facebook’s targeting capabilities come from social logins — which won’t be affected in any way by Firefox’s initiative. The companies hurt by this new policy are the “other” (i.e., smaller) ones, responsible for 28% of advertising spend on the internet. In other words, Firefox may help accelerate a world where ALL internet ad dollars are directed toward AmaGoogBook.

What about the implementation of ‘Sign In with Apple’ as a way to capture more users? Once again, this change will have no real effect Google. Likewise, it will have no effect on Amazon (who will DEFINITELY not use Apple login for their core site). But it WILL swing the number of social logins using Facebook. And for each login that Facebook loses to Apple, their targeting capabilities will diminish. Moreover — unlike Google Search Ads — Facebook advertising spend is fungible. If Facebook ads get a little worse and Amazon ads get a little better, on the margin it will make sense for advertisers to shift some percentage of their spend from Facebook to Amazon. And of course, “other” will take another hit here too. In the battle of these Goliaths, David getting crushed is just a side effect.

Keep it simple,

Edward

* The vast majority (37% vs 4%) of this search volume goes to Google. The percentage is even higher in Europe. Google's share, in addition to search revenue, includes Google Display Network (GDN) and YouTube. Both are significant but much smaller than search. Microsoft's number includes LinkedIn which is likely fairly insignificant at this point. In some future article I will dig into LinkedIn advertising.