Sitename

What if Performance Advertising is Just an Analytics Scam? - SparkToro


“What the pandemic showed is we can take marketing down to zero and still have 95% of the same traffic as the year before. So we’re not going to forget that lesson.

– Brian Chesky (via Campaign)

In 2020, AirBnB cut $542 million of performance advertising spend and saw no measurable falloff in attributable sales. They continued this ad-slashing practice in 2021, with similarly eye-popping numbers. Could it be that all those ads did nothing?

It reminds me of a famous marketing parable:

One day, Lorena, owner of Lorena’s Pizzeria, hired three capable go-getters to paper the neighborhood with promotional material. She provided each of her new marketers a stack of color-coded flyers (red, green, and white) with the pizzeria’s menu and a unique discount code. Lorena reasoned that if business went up, not only could she attribute sales (via the discounted pricing) to the promotion, she could determine how much each of her three new employees were contributing via the three discount codes.

After a month of distributing flyers, Lorena reviewed the sales numbers, called in her papering team, fired the two passing out the red and white pamphlets, and gave the green-flyer-distributing employee a massive bonus. After all, the sales data showed that green flyers had contributed almost 50% of the pizzeria’s monthly sales! Conversely, the red and white discount codes were used in fewer than 5% of orders each.

At the end of the year, Lorena’s accountant reviewed the business’ receipts and came to her with mixed news. Total transactions were up ~10%, but because the discount code was used so often, overall, revenue was flat. Lorena was shocked. How could the pizzeria’s sales be up only 10% when nearly half the transactions used the new, green, discount-coded flyers?

She had to know, so the following day, Lorena left the pizzeria in disguise and trailed the green-flyer distributor. What did she find? The marketer barely took twenty steps out of the restaurant’s door, and quietly slipped a green flyer to anyone whose footpath suggested they were on their way to the pizzeria.

The lesson is simple: advertise to those already primed to buy and you’ll see a phenomenal return on advertising investment without any sales lift.

Some are convinced that’s exactly how a majority of so-called “performance advertising” channels work. They don’t boost sales (at least not much), but they unfairly take credit for a huge percentage of sales that would have happened anyway.

Say you’ve just visited some articles recommending fancy paper notebook products (as I did earlier today). Later, you pop over to Facebook and what’s this?

Ads on my Facebook feed… Was I already going to buy these things? Or did these ads actually create new demand?

Perfect gifts for the men in my life? A planner recommended by Forbes? Oh, Facebook, you stalk me so well.

Chances are decent that I’d already had similar purchases on my mind. Facebook is neither creating nor fulfilling the demand, but since they have tracking pixels on nearly every web property in existence and a near-monopoly on modernity’s social networking addiction, they can serve up ads that, even if they don’t nudge behavior, get credit for it.

Google, Facebook, Amazon, and the ad networks that power the web are vast nests of behavioral and predictive data. The first half of their pitch is that incomprehensibly large quantities of historical data about what people visit and engage-with online plus artificial-intelligence technology (really just machine-learning, but “AI” sounds sexier) allow them to show the right ads to the right people at the right time to deliver sales.

I believe them.

These platforms, even without 3rd-party cookies (whenever those finally die out), have enough data about what billions of people search, visit, browse, and buy, to predict pregnancies before mothers know about them, to anticipate which restaurant you’ll look up after landing in a city you’ve never visited before, and to suggest you befriend genetic relations you never knew existed. They have our express permission (via all those ToS’ we signed) and legal cover to marry financial data, email data, browser data, private messages, social graph connections, ISP and device data, and a thousand other prediction-model-enhancing nuggets.

Determining whose ads to show when and where has been the pursuit of a generation’s best-educated, best-paid knowledge workers for two decades. So color me completely un-surprised that Google and Facebook can tell you with absolute certainty that 72% of people who purchased on your website saw the ads you ran on their networks 3.37X before conversion.

The trouble is whether those 3.37 views actually modified the purchase behavior of that 72%.

Seen? Yes. Modified behavior? Hmm. If Facebook and Google were really that effective at modifying behavior, wouldn’t we all be Qanon cultists by now (or at least have cast votes for Michael Bloomberg)? If we can prove live TV ads have an elasticity of 0.01 (i.e. doubling ad spend and viewership would increase sales by 1%), why would we expect scroll-past-em-faster-than-they-load web ads to work differently?

Here we come to the second-half of the performance advertising pitch: those perfectly-timed, perfectly-targeted ads will influence more people to buy your product that would have without them.

On this part, I don’t believe the ad networks.

Worse, I think these networks (Facebook, Google, and perhaps even more egregiously, the long tail of ad platforms) know that the degree to which ads impact behavior is small, but the degree to which ads can get credit for conversions is large.

Herein lies the scam. I’m not saying “no one buys because of a retargeting/display/branded search ad.” I’m saying, “somewhere between 60-99%* of the people exposed to those ads would have purchased anyway.

The ad platforms know this. Many of the ad buyers even know this. But because the platforms have no incentive to make incrementality (i.e. the additional lift in sales that a given ad campaign creates) clear, ad buyers look at their analytics and think, “I should spend more on performance marketing!

Google’s own Avinash Kaushik has written thoughtfully about not confusing attribution with incrementality :

Google Ads = easy to measure, Incrementality = hard to measure

You can probably guess how advertisers respond

But, when you get to the bottom of his post, you’ll see a methodology for measuring incrementality that’s probably effective, but so painfully challenging to execute that almost no one will bother.

The last few years, we’ve seen hundreds of stories about cutting off millions in ad spend with no discernable business impact. This Forbes piece covers many of the largest, but I’m equally interested in the stories of small and medium-sized advertisers who’ve been shocked to discover their search ads, retargeting, programmatic display ads, LinkedIn ads, Reddit ads, Facebook ads, etc. are some combination of entirely ineffective, fraudulent, and/or incrementally non-additive.

Here’s the benefits of performance advertising’s undeniably beautiful attribution:

Now let me make the case for turning off ads entirely to get data on the incrementality of marketing:

See what I mean? It’s a shitty pitch.

Technically, when someone does a Google search for “Williams Sonoma Cast Iron Skillet,” they probably would have clicked on one of the first 10 organic results, EVERY ONE OF WHICH leads to their website. But, y’know what ol’ Billy Ma’s performance marketers couldn’t then do: prove their value to their bosses.

It’s infuriating.

Equally infuriating: if Williams Sonoma doesn’t buy out that ad block, their competition might show instead, siphoning sales (or at least creating a perception of intolerable risk for marketers and executives alike).

Similarly, this Exasol ad for a BI performance dashboard loses nothing if I never purchase a subscription to their SaaS tool. Technically, it’s still a brand impression. The brand marketers can say “4,219 tech CEOs saw our ads, and we’re gaining mindshare.” But, if I ever do convert, having this ad somewhere in my browsing history also makes their performance marketing team look like geniuses.

If Williams Sonoma turns off branded search ads and Exasol shuts off display ads, who wins? Sure, their bottom lines do. And maybe a competitor will siphon away some traffic or even a sale. But, which member(s) of their organizations gets credit for the lower cost of custom acquisition and more efficient marketing engine?

Maybe, if we lived in a very different macroeconomic environment, a hard-nosed CFO could get a pat on the shoulder. But in a world with interest rates at 0% and investors clamoring for growth > profits, fuggedaboutit.

Most marketers I know believe:

  1. A majority of performance advertising, especially in areas like display, branded search, social, and retargeting deliver only a small amount of incrementality and cost an awful lot.
  2. In some sectors and with some brands, performance ads are super effective, adding massively to incremental sales. Differentiating these from the many sectors and companies where it doesn’t work so well is… difficult.
  3. The value of data proving the value of your efforts to your boss/team/client, and never even starting a conversation about incrementality, is pretty sweet (and being a performance marketer is already a damn hard gig).
  4. Getting clients, teams, and managers who happily spend on low-ROI ads to switch to potentially high-ROI but hard-to-measure channels like content, PR, email, organic search, organic social, influence marketing, etc. is like pulling teeth.

When no one’s incentivized to make the right move, the right move may as well not exist.

This incrementally-non-additive advertising problem is:

But, no matter how sophisticated your analytics and testing methodology is, if you spend money on performance ads with the goal of boosting sales, this problem affects you. It’s in the ad networks’ interest for this problem to affect you, which is why I really hope no one replies to this post with a link to some Google or Facebook-funded/published research paper about the incremental value of bidding on branded terms or being seen in a 2-second “scroll past” view.


One final thought experiment. Remember the 2014 LEGO Movie?

It’s a giant piece of very expensive (~$65M) content marketing. But, it made LEGO a frickin’ fortune. Revenue grew ~11% that year, while net profits went up ~12% (at their size, when you grow sales by double digits, margin typically goes down). LEGO’s marketing team could reasonably have spent that $65M (it’s unclear how much of the film they financed, but for this argument’s sake, let’s presume a lot) on advertising. The company’s estimated total budget for ads is <$100M/year, so $65M would have been a massive increase.

But, instead, they took a high-risk, high-reward bet on content, and won big.

My contention is that most marketers and brands could benefit from similarly risky, experimental bets on hard-to-track, hard-to-execute, organic marketing investments. Most could, with a disciplined, willing-to-fail approach, redirect 2/3rds or more of their digital ad spend and, over time, win bigger boosts to demand with lower cost.

Maybe you think my numbers are off, but I bet no one reading this believes Google and Facebook (nor any of the third-party ad platforms) deserve 100% of the conversion credit to which they say their ads have contributed.

My advice: if you have the power to invest any percent of your digital ad spend in other, more serendipitous, hard-to-measure channels, take it. If you have the ability to give your marketing team buy-in for that spend, approve it. If you have the patience and discipline to focus on profitability over unprofitable growth, grab it. You’ll be amazed at just how much growth you can invest in once you find channels that don’t cost $0.99 for every $1.00 you make.

Via Roy Steves (of StatBid, an e-commerce marketing agency)

At the very least, consider auditing your ad spend with whatever level of sophistication you can afford. Depending on budget, you might find more net profit cutting extraneous spend than via all your other marketing activities combined.

* My previously uncited estimate of “60-99%” drew some criticism after publication. For clarity, that’s the range from the linked-to case studies from Uber, AirBnB, P&G, eBay, Bloomberg, Chase, as well as the anonymous examples from an app-maker, music service, medical device brand, and e-commerce shop.