Soundtrack:

Table of Contents

• Endtroducing Incrementality

• What Happened To This Beautiful Concept?

• How This Actually Changes Everything

• Understanding Meta’s Real(istic) Goals Here

• How Smart Meta Advertisers Will Use This Tool

• What Is An Incrementality Professional To Do?

• Give The Devil Its Due: Digital Signal Driven Bidd …

• The World of “Offline Data” is Still Largely Safe

• The Huge Gap This Does Not Fill, For Anyone

• Did Anyone Ask Robyn What She Thinks?

• What Does This Mean For The Free Internet?

Endtroducing Incrementality

Hooray! Incrementality is over! It’s solved!

What was Incrementality?

I’m surprised anyone who would be reading this didn’t get sick of hearing this word years ago, let alone need it defined for them, but to answer the question, here’s a definition I don’t hate from an article by Measured:

“Incrementality in marketing refers to the causal lift in sales driven by a given marketing activity for a given period of time.”

Basically: how much more stuff did you sell because you performed some marketing activity, than you would have sold if you did not perform said marketing activity.

Incrementality is arguably the “perfect” single advertising success metric, at least without getting into things that bring in corporate accounting and other things that are very much not advertising.

This is why it has long been a popular measurement topic for advertisers who care about actually doing a good job.

What Happened To This Beautiful Concept?

The thriving ecosystem of incrementality enthusiasts has been hit by the same comet that eventually blasts a Yucatan Crater sized hole in every fun little marketing microcosm:

It was coopted by by The Powers That Be. Specifically, Meta is launching its own in-platform Incrementality product.

There’s good news, bad news, moderate peril, and great opportunity!

The good news is we’ll no longer have to deal with clickbait LinkedIn posts where someone just says “you need to measure the INCREMENTAL impact of your marketing” like this isn’t akin to stating “I think someone should do something about this One Ring that Sauron is going to use to conquer Middle Earth.”

The bad news is that yet again we will have to wrestle with the fact that a critical concept that applies to the whole world of marketing is becoming the domain of a global megacorporation that doesn’t about us continuing to make a modest living on the outskirts of its empire.

I’ll cover the peril below, and then get to the constructive thoughts I have around where opportunity lies for people who are not The Zuck.

How This Actually Changes Everything

Look, I’m proud of most of you who have been banging the Incrementality drum for the last three or thirty odd years.

I think there’s still room for y’all to keep doing that in a productive way. I honor that you were there first, and love you for your pioneering spirit, or in some cases, mistrust you for your cynical buzzword chasing.

Pioneers and plunderers alike do, however, need to take a moment and understand that Meta attribution will be the Kleenex, or Coke, or whatever shorthand you like, for “Incrementality” as a digital ads product soon.

It’s not just that your mail carrier with a thriving Etsy business and your weird uncle with a dropshipping hustle are about to learn the word “Incrementality,” and assume that it’s an idea Meta invented.

Besides, I know that long-tail of itty bitty advertisers isn’t what my audience cares about. Here’s most of y’all when the project budget is less than $100 million:

Hey, while we’re talking about budgets and money, may I offer you something of great value, for free? A subscription to my humble newsletter:

Alright, so here is the problem that will actually curdle the milk for all you big budget fat cats:

Everyone in enterprise marketing, VC backed ephemera, B2B SaaS Demandification, and all these other spaces are going to quickly swallow the whole narrative peddled by Big Tech Platforms like they have for everything else going on 15 or so years now.

It’s won’t be only because of the usual sad reasons, like how every ad agency holding company in the world combined is valued at a fraction of Meta’s market cap, or that last I checked the N in FAANG wasn’t Nielsen, nobody respects people in advertising and we don’t respect ourselves, etc.

It’s going to happen even more easily than usual because the independent incrementality world already built a lot of its clout on credentialism from these vary companies by making a bunch of ex-(BigTechName) people its stars.

You can’t look at anyone’s list of a who’s who here and not see a lot of ex-Amazon, ex-Google, ex-Meta and such. The very notion that we had some powerful minds working in independent incrementality seems to have been built partially on the assumption that powerful marketing minds mostly come from Googlemetamazoniatok.

Well, we’re about to learn the power of people wielding those company names with no (ex-) in front of them, who still have the fiscal and cultural heft of said companies, plus their B2B marketing machines, at their back.

I’m not saying I think that great experts on marketing can’t come from these companies. I’m just saying it’s a double-edged sword we’re all too eager to wield when it’s going to swing our way.

I like Prof. dr. Koen Pauwels and his marketing newsletter a lot, you should read that before my slop! But let’s also be real, his Principal Research Scientist at Amazon resume entry pops a lot more to you than his prestigious academic positions and publications if you are, like me, North American Scum, because we’ve been deeply conditioned to think like that. We love big companies, we respect them so much, no matter what.

The good news for us Non-Xooglers…oh no, hold on, sorry, I watched myself type that word, and I’m feeling…

Ah, okay, I got it out. I feel much better. Never willfully typing that again, though. Awful.

Anyway, the good news for the rest of us is that there’s still a lot of valuable territory that neither Meta nor any likely fast-followers in that class of company can claim, and if y’all are smart about it, there’s still very much a bright future outside the parts of incrementality these leviathans are going to swallow in the next few years.

Understanding Meta’s Real(istic) Goals Here

First, let us understand our adversary, and the territory they aim to take.

Meta will say this is an aim to capture more spend, because they’re one of the INFINITE GROWTH FOREVER companies. Arguably, one of the first. Anyway, you don’t talk about “playing defense” or “mitigating risk” when you’re like that.

The truth is, though, that in many markets, Meta Ads have become a bit oversubscribed, and what this will really be is a vaccine against the increasing number of bugs trying to get under their skin and show advertisers that they’re spending a bit too much here.

We know Meta are already worried that people don’t understand the true value of their current investment. They went really hard on a partnership with Longtermist Legend Les Binet, so we know that they’re at least a little hurt that some of us don’t think Meta is the number one longterm-ism channel for brand building and Byronic Magic.

In the past five years, Meta also got a lot of mileage out of the power of their own brand they’d built with marketers, but burned some credibility along the way. Lots of advertisers spent a year or more post-iOs14.5 meltdown continuing to drop tons of money into Meta Ads, even while Meta’s own ROAS metric outputs tanked.

These brands believed in Meta’s brand promise too much to stop spending, but some of them did at least begin to seek new wisdom on where their dollars might work best…sometimes towards independent incrementality experts and tech!

There was also a major push from the TV industry, especially in the UK, to produce a lot of research arguing TV was largely underbought, largely at the expense of digital channels, and this gained a lot of traction in parts of marketingland, which also likely spooked Meta a bit.

So here’s what Meta hopes this does for them, as seen in your own life, dear media buyer or analytics professional:

Six months from now, you trot out your brilliant incrementality measurement solution to a client, and have a great meeting, and feel like you’ve found a prospect who REALLY GETS IT…and then when you mention that among the data you’ll need from them is Meta’s, and they’ll ask you:

“Why does your incrementality solution need Meta data? We already know how incremental that is.”

The thing that should make you optimistic is that Meta’s strategy here seems more inwardly-focused and, though they won’t say it, defensive, than it may appear to be. At the very least, I think they believe they can only lose to themselves, and so outside of Robyn (more on her later) they aren’t trying to swallow the whole marketing measurement world so much as defend Meta spend from it.

We can get through this happily, together, if we imagine how we’d act in the best case scenario we’ll encounter: having smart clients.

How Smart Meta Advertisers Will Use This Tool

This part of the blog is inspired by a LinkedIn post by Jeff Eynon, and I’ll summarize our consensus (as I saw it) here.

Advertisers fall into three groups as far as this tool’s release is concerned.

1. Those who don’t yet have reason, or means, to be using experimentation frameworks or technology to consistently determine attribution and apply it to paid media. Annual ad spenders in the six figure to low seven figure range, businesses with less than two years of usable sales data, etc.

2. Those who have both the need and the capacity, but aren’t doing this yet.

3. Those already engaged with an independent incrementality solution of some type.

Group One should turn this on and see if it tells them anything useful, and then pit these numbers against an independent incrementality approach as soon as they are able, giving no bias towards Meta’s numbers simply because the solution has been in place longer.

Group Two should do the same thing, essentially. Use this product’s rollout as an excuse to get your butt in gear quickly. Turn on the Meta product for context ASAP, but don’t dilly dally on arriving at your sensible independent incrementality approach.

Group Three is where we will see our thesis put to work immediately. Brands already using an incrementality approach should “bake off” this number as a directional, channel optimization only metric against what they’ve been using previously, which is likely ROAS or CPA. Don’t assume the Meta incrementality number will track closer to independent incrementality until you test it!

Gruppe Sechs is a pretty funny running joke in the Grand Theft Auto series of games about a fictional German private security corporation:

If you’re an advertiser and this helped you, I will have plenty more where that came from, so please consider lending me your ear, or, rather, your inbox:

What Is An Incrementality Professional To Do?

What an Independent Incrementalist is to do is now apparent: embrace the Meta incrementality number, but only for the sake of diminishing it to a channel-level optimization data point, a capacity in which it might be useful.

Don’t pretend it isn’t there and lose credibility. Acknowledge it, and get ahead of the conversation so you can rationally advocate more comprehensive, independent solutions.

Now, let’s cover some devilish details, top tips, and things not to do.

Give The Devil Its Due: Digital Signal Driven Bidding and Audience Optimization

Here’s where I’m going to possibly pick fights with two groups of people who may read this by saying you absolutely need to experiment with the real-time optimization element of this new offering, and those groups are:

1. Incrementality vendors / experts who think this product is entirely rubbish and the only correct thing to do is scream this at the top of their lungs that it’s terrible, until it goes away.

2. Humans, and technology companies, that push targeting and bids into Meta from the outside, whose value proposition is that they optimize based on data external to Meta Ads.

I think I have been pretty clear that you shouldn’t use Meta’s own numbers to actually gauge Meta’s incrementality as a channel, unless you have literally no other choice, and in that case you need to be careful. My whole opinion here is pro-people who are not Meta.

That being said, there’s not much I can really say to Incrementality Heads who won’t even test this.

We’ve all seen the long, sad history of what happens when critical conversion data moves into a digital ad platform and there’s a too little + too late rebellion against it.

Focus on your total incrementality solution, and not maintaining your grasp on Meta-specific incrementality optimization.

If what you mainly sell is bids, that’s fine! There’s lots of stuff you can still make hay pushing bids into. Focus on smart budget bucket development and management for Meta, which I personally still see as a big need for many business types.

The World of “Offline Data” is Still Largely Safe

There are lots of Meta advertisers that do not use the core Meta online success metrics like ROAS or CPA.

CPG brands sold mostly through retailers. Movie studios. Bulgur wheat wholesalers. Domestic propagandists for Lockheed Martin and Exxon. International propagandists for the venture funds of brutal despots. Things of that nature.

For most of these industries, the key results data is exterior to Meta, and so the key measurement solutions are much less reliant on Meta data, utilizing merchant POS data and the like.

In these cases, media optimization solutions, human or machine, that are outside of Meta still have an easy claim on being a better solution than Meta’s own incrementality optimizer.

This seems to be what most of your independent bid automation platforms and custom algorithm solutions focus on, which made tons of sense before now, and makes the most sense in the hereafter.

The Huge Gap This Does Not Fill, For Anyone

Guys, here’s the best news of all: this solution looks as ill-positioned to address the weakness of every other Incrementality-Obsessed Solution to date: the fourth dimension, time.

But wait, you cry! In the very definition you gave for Incrementality earlier, there is the phrase “for a given period of time.”

Yes, but as much as this is going to make practitioners of both “pure incrementality” and MMM hate me, I think an actual key difference in the real world is how much one does, and one doesn’t, like to play with and define that time period.

MMM tends to be quite interested in working out what a time period is, or at least ensuring the question is very carefully considered. Adstock, which is essentially this expected time period, goes into every widely used MMM I’ve seen in the past several years and calibration of the MMM is often quite focused on this.

What MMM people don’t love about this statement is they probably think there are much more important and interesting answers to the question of “how is MMM different than incrementality experimentation or longitudinal incrementality measurement?”

Well, they get off easy compared to what I will say about pure incrementality tools and the Meta Ads platform itself, which is that they mostly treat this question of time-lag-to-impact as an inconvenience to rush past as quickly as possible. The end result tends to be that they don’t think adstock for Meta, or anything, is ever very long at all.

Meta curiously dabble from time to time in studies on their ads’ longterm impact, but don’t seem to focus on that where it matters. They have defined a lot of marketing thinking with their classic 1 Day vs. 7 Day attribution view they focus on within the management platform and reporting, which also tends to show that the most key difference is not time, but if the user clicked or merely viewed the ad.

Even where time window comes in for Meta, it is often very closely tied to click vs. view-one of the most common things you’ll hear from people who analyze this data is about how click or view conversions are more or less reliable depending on when they come in, i.e. time lag is kind of a secondary lens to lay on click vs. view attribution.

Did Anyone Ask Robyn What She Thinks?

Hey, remember that whole open-source MMM product Meta launched, called Robyn?

I get the vague sense this thing actually has some adoption among advertisers who think an MMM is important to utilize but not important enough to pay for.

While one shouldn’t conflate any incrementality experiment or tool with MMM, it’s clear that lots of people believe that advertisers should use both, and there’s sensible ways to go about this.

Helping advertisers disjointedly adopting both solutions is probably a worthy Consulting Endeavor.

You wouldn’t want to put Robyn in the corner, watching you click “See Incrementality” in your little Meta Ads tabs, all alone, would you?

For any bigwigs in the measurement space who are going to have access to lots of Meta Ads data from this new incrementality tool, and the same kind of data needed to run some real Robyn analysis…

…I think it would be very, very interesting to do a real large-scale study where you thoughtfully and carefully interpret the outputs of this incrementality tool alongside what Robyn says about Meta’s place in a company’s marketing mix.

It would be a great starting point for something extraordinarily interesting.

What Does This Mean For The Free Internet?

Well, I said this question is what the whole newsletter is going to be about, though admittedly this first topic is pretty cut-and-dried.

The more today’s Meta Corporation dominates what we know as the internet, the less free and good it is, barring massive systemic changes at the company and with all of its products that I can scarcely imagine actually coming to pass soon.

So, the more it profits and is able to extend its grip, the worse things get. To the extent this product preserves or expands the ad spend they get, it’s a stinker!

How could it end up being a good thing?

Uh, well, imagine very imaginatively with me for a second. Let’s test the limits of optimism.

If Meta taking over grading its own homework inspires people to save their more serious incrementality work for other platforms, we might continue to discover that non-Meta platforms more conducive to an open internet and media environment, are also more incremental than previously understood, and we’ll invest more in them to tip the balance a bit.

It’s also possible that if the Zuck uses not just this incrementality tool, but the coming All-AI Advertising Suite he is currently ballyhooing, to chase humans away from making a living by managing advertising on the platform, an unexpected second order effect will take hold.

The sustained investment through Meta’s no good, very bad, terrible, awful post-iOs 14.5 time was due to human affinity for the platform.

We had all these Meta ads managers, and all the Meta ads focused agencies, and they couldn’t just all pivot overnight, dammit! We need to keep spending on Meta because that’s what we humans all knew how to do back then: spend on Meta!

Well, what happens if in a few years performance flags, and there’s nothing left but Independent Incrementality Machines coldly gazing at Meta investment and results? Absent soft, sentimental humans who still remember their first social media agency job and the friends they made at it, what will protect Meta from the full consequences of declining effectiveness?

Nothing.

Maybe when the Incrementali-Tron Megamind some future analysts have built coldly zeroes out the last budget line item Meta Ads had left, The Big Blue App will belch out some strange AI slop version of a scream.

“I was almost a network of weather balloons beaming internet down on the undeveloped world. You were going to live in me as a virtual avatar. Trent Reznor soundtracked a movie about my creator and it was all so beautiful, and that’s why we had to lie and tell people even the ugliest heart can create something sublime by dumping out its own emptiness into a chatbot.”

By the time it reaches any human eyes through the twisted labyrinth of social shopping intelligence agents that will be the only thing that are capable of processing raw social ads directly in the 2030’s, it will just be end up as a push notification that “The Social Network” on BluRay is 42% off at the nearest TemuTower ShopXperience Zone.

Anyhow, you made it this far, so if you haven’t already, for the love of all that’s good and just as importantly, for the hate of all that’s evil, please: