Why we regret putting ‘outcomes’ in our tagline

At a recent industry event, a founder who I’d lightly mentored approached me with glee in his eyes.

“Ali, I know you will appreciate this. We launched the product a few months back, onboarded our first customers and it works. The product works! It is driving outcomes, and our customers are happy.”

I gave him that insider smile. The kind of smile a parent gives to someone who just had a baby. Welcome to the club. I think founders are akin to parents; you simply can’t imagine what it’s like until you become one. Welcome to the club.

“That’s amazing. Isn’t it the biggest relief and moment of pride to find out your product actually works?” I beamed, remembering how I waited with peak anxiety to see if Chalice’s product would actually live up to its promises once we let it loose in the wild.

“Yes, exactly. You won’t believe the click-through rates.”

Oh, lord.

Like when a new parent shows off their ugly baby, there is nothing to say to an enthused, proud founder with a product that is optimizing clicks except to be gracious.

“I’m thrilled for you.”

And I was. He loves his baby. Customers seem to also like his baby.

I just wish he wouldn’t have named his baby Outcomes.

But that’s the thing, in marketing, we bastardize words. Companies co-opt the English language until phrases describe such a pupu platter of random shit that the words become meaningless. 

When we started Chalice, we thought we were so smart using the tagline “AI for Outcomes.” And we’ve been crestfallen to see it get so dumbed down.

It seemed so pure. It literally means the final result, consequence or effect of an action, event or process. We used it for top-line sales growth, market share gains, household penetration and customer value. To be honest, we still do. We haven’t found a better term to replace it.

But I’ve come to accept that the Age of Outcomes was dead on arrival.

One of our early clients was a consumer good that was sold mostly in convenience stores, with a little e-commerce on the side. The exec sponsor who brought us in was wooed by our outcomes narrative, but when we got to the day-to-day activation team, they were only here for last-“touch” attribution (for my extended family who read my articles but have no ad tech context: This means the last ad to be “seen” is the one that gets 100% credit for the sale).

Our customer success lead, a former trader, had a few suggestions to implement on top of our custom algorithm: Target ad positions at the bottom of the page, because the ad that is last to load on a page gets the “last touch” credit. No matter that most of these ads are never actually seen because users don’t scroll down that far.

Serve most ads in the last hours of the day. Because customer systems sync at the end of the day, you can get credit for a sale that has already happened. For example, Graciela buys the product at 3 p.m. We serve an ad to her at 8 p.m. We get last-touch attribution for the sale, even though she didn’t “see” the ad until five hours after she bought the product.

Override the algorithm’s decisioning on bidding for value and maximumly lower the bids so you essentially impression-bomb the internet, playing the odds that simply getting a lot of cheap impressions will catch people who were going to buy the product anyway.

There would be plenty more shenanigans if we’d had the marketing budget.

I was dumbfounded. That’s like being hired by a fast-food joint to dress up as a chicken to get people into the store and get paid for anyone who uses a coupon you hand out. Obviously, the best way to make money is to stand at the cash register and pass out coupons. But it’s not what you are hired to do.

I said no to all of it. Let the algorithm sort it out.

After a few weeks of begging the customer to measure our campaign on something that included offline sales impact (again, 80% of sales happening in store) and to stay live even on our own last-touch merits, showing them charts of their CPA starting too high but then dropping like a rock … we got turned off. C’est la vie.

Except, of course, it isn’t just life. It’s an incentive system. We refused to game the metric, and we lost the business to companies that were willing to. (Or, more generously, whose technology had been trained for years to become exceptionally good at claiming credit.) The metric said they were creating more value, so according to the metric, they were.

There is a name for this: Goodhart’s law. (Not those ad tech royalty Goodharts. Jonah and Noah founded Moat, although Moat is a real-world study of the law itself in action. The actual law is named after British economist Charles Goodhart who described how government monetary policies collapse when you try to manipulate specific financial indicators.) The version of the law everyone actually quotes came from anthropologist Marilyn Strathern, who managed to improve the economist’s original in a single sentence: “When a measure becomes a target, it ceases to be a good measure.” (Naturally, the law still carries Goodhart’s name, not hers.)

Ad tech, and marketing overall, has managed to apply Goodhart’s law not only to our metrics, but to our language.

We don’t just Goodhart metrics.

We Goodhart words.

That’s what happened to “outcomes.”

Outcomes began as a way to distinguish business value from the proxy metrics we’d spent decades gaming. Now agencies and vendors proudly sell “outcomes businesses” built around guaranteeing those same proxies. 

The pitch goes something like this: We’ll guarantee you a $100 CPA. The outcomes business buys enough media to produce an attributed conversion for $60 and keeps the $40 spread. The client gets certainty. The agency or vendor assumes some risk and makes a healthy margin. Everyone may walk away happy.

It may be a perfectly fine commercial arrangement, but calling it an outcomes business doesn’t suddenly make the advertising causal. It doesn’t establish that the sale wouldn’t have happened anyway or that the marketing created any incremental value for the brand. It mostly hides the margin behind a new label.

So yes, I’d love for us to stop bastardizing perfectly useful words.

We won’t.

There’s too much money in sophisticated language, especially when it can be hiding behind an easy button.

Instead, brands need to get better at distinguishing between companies that create value and companies that are simply very good at collecting credit for value they didn’t create.

Start by asking your agencies and technology partners a few impolite, but remarkably easy-to-answer questions.

Show me a report of ad delivery by hour of day, right next to sales by hour of day. Are ads mysteriously piling up late in the day? How many ads received credit for sales that had already happened?

Show me where my ads loaded on the page. Were they actually likely to be seen? Pair that with viewability, attention or video-completion metrics. None of those are perfect. All of them are harder to fake than an attribution report by itself.

Then ask the harder question: Can you demonstrate that sales, customer value, household penetration, brand share or whatever metric my business actually cares about increased beyond what would have happened anyway?

Those first reports are diagnostics, not proof. But they’ll tell you whether you’re looking at a business that’s trying to influence customers or one that’s exceptionally good at collecting credit for outcomes it didn’t create.

Not every campaign needs a doctoral dissertation on incrementality. Marketers have deadlines, budgets and bonuses, and sometimes a CPA really is the metric that has to get delivered. But we should at least be honest about which game we’re playing.

A click is a click. An attributed sale is an attributed sale. A guarantee is a guarantee. All can be useful. Calling all of them “outcomes” hasn’t made advertising more sophisticated. It has only made our language less so.

And I don’t think the answer is inventing another buzzword. Whatever we come up with will be on every OOH installation at Cannes by next summer. The answer is demanding that businesses explain what value they created, how they know they created it and what would have happened if they’d never existed.

Which brings me back to my founder friend and his brand-new product.

I still meant it when I congratulated him. Founders learn. Products improve. Beauty is in the eye of the beholder. Let’s just not be beholden to putting an “outcomes” onesie on every ugly baby.