# We optimize everything except the question

**Written by JD Prater — 3x VP of Marketing.** This piece is part of *Open Folders*, which sits at the intersection of marketing science, consulting research, and practitioner experience, opening the parts of the discipline most marketers never got taught, so they can lead the whole of it, not just the slice they came up in.
Newsletter: https://jdprater.substack.com
Website: https://jdprater.com
Original post: https://jdprater.substack.com/p/we-optimize-everything-except-the

By 2009, Snickers was losing.

A decade of sliding market share. A Super Bowl campaign two years earlier that had attracted the wrong kind of attention. A brand that had lost its footing in a category full of nearly identical products competing on price and shelf placement.

In that moment, most marketing teams reach for the tools they trust. Tighten the targeting. Double down on loyalty. Prove ROI. Show the CFO something defensible.

Mars and BBDO did something different. They went looking for a human truth.

When you're hungry, you're not yourself. A feeling everyone has had and never seen named. They built a campaign around it. Launched it at the Super Bowl with Betty White, an 88-year-old actress playing a college football player.

In year one, global sales grew 15.9%. Market share increased in 56 of 58 markets. Over two years: $376 million in incremental sales.

They didn't report which channels drove it. They reported what the business did.

## Would you have made that call?

Ask yourself honestly.

A brand down 10% in market share, coming off a reputational stumble, proposing a bold emotional campaign with no guaranteed attribution path. The pressure would have been immense to prove short-term results first. To measure your way back to safety.

Most teams would have. Most still do.

I once proposed a billboard media buy to reach a broader audience we weren't capturing through digital. The idea wasn't complicated. Get in front of people before they were in the market. Build some presence while the competition was invisible.

It got killed before it got started. Too hard to measure in GA. Too risky. Probably a waste of ad spend.

We ran another retargeting campaign instead.

## The dashboard felt like progress.

CFOs loved it. Boards demanded it. And we stopped asking harder questions.

In the late 1990s and through the 2000s, something significant happened to marketing. Google built an auction model. Facebook built targeting infrastructure that could reach almost anyone by almost any attribute. Analytics platforms started attributing conversions to the ads that touched them last. And marketers trusted them.

For the first time, you could see exactly what your money was doing in near real-time.

A generation of marketers was trained inside it, learning that measurability and rigor were the same thing.

They aren't.

This is worth naming plainly. What followed wasn't a debate between brand believers and performance believers. It was a timing problem dressed up as a values debate. Performance marketing measures what happens now. Brand investment determines what's possible later. When your tools can only see one of those, you don't make a balanced decision. You make a partial one, repeatedly, until the compounding cost shows up in a number you can't explain.

The frameworks built by Kotler, Aaker, and Keller. Decades of rigorous thinking about how brands create and sustain economic value. The brand science that survived lived inside agencies. In-house marketers never got near it. Nobody dismissed it. Nobody debated it. It just never came up. The next generation learned digital marketing, performance marketing, and channel execution. They got a HubSpot certification and a dashboard.

## We learned on the job. The job only showed us part of it.

Decades of rigorous evidence on how brands grow, how advertising works, and how marketing creates long-term business value exists. The Ehrenberg-Bass Institute. The IPA Databank. WARC effectiveness studies. Binet and Field's analysis of hundreds of campaigns across categories and markets. Peter Field's work on what separates campaigns that move business from campaigns that move metrics.

It's all there. It's all accessible. Most in-house marketers have never encountered any of it.

We learned performance marketing. We learned digital channels. We learned conversion funnels and attribution models and how to read a platform dashboard. And we got good at it. The problem wasn't that we learned the wrong things. It's that we learned a sliver of the full picture and nobody told us that's all it was.

Ask the average product marketing manager who Byron Sharp is. Or Les Binet. Or Peter Field. Most won't know. That's not a criticism. It's an indictment of how the industry trains its people.

An Ipsos study of 1,226 marketers found that only 35% met a basic benchmark of foundational marketing knowledge. Two in three didn't. The areas of weakest performance weren't the classic frameworks. They were the concepts that link investment to long-term growth: excess share of voice, penetration, media neutrality, distinctive brand assets. The things that explain how brands actually compound advantage over time.

The brand science that survived migrated into agencies, consultancies, and academia. In-house teams, especially the generation that came up in the last ten to fifteen years, learned marketing through exposure. Through doing. Through whatever their manager knew, which was usually whatever their manager's manager knew. Nobody pointed them toward the research. It just never came up.

So when budgets get cut and someone asks why brand investment matters, most marketing leaders reach for instinct rather than evidence. When the CFO wants proof, the room goes quiet. When the death spiral starts, nobody recognizes it because nobody was taught what it looks like.

The knowledge gap doesn't stay in the classroom. It shows up in the budget meeting.

## The meeting that never happened.

Most companies never made a conscious decision about their measurement infrastructure. They inherited it.

In nearly every company I've worked at, I inherited the same setup. GA. Platform dashboards. A last-click attribution model nobody had consciously chosen. It had just accumulated, tool by tool, campaign by campaign, until it became the official version of reality. I rarely questioned it early in my career. I just learned the reports and got to work.

It took me until my first 90 days at Ting to do something different. I bought an MMM. Not because I had all the answers, but because I'd started to understand what we couldn't see without it. That decision changed how I thought about measurement permanently.

Most teams never get there. The default holds.

GA came bundled with the website along with last-click attribution. Ad platform dashboards came bundled with the media buy. Free, fast, and telling a coherent story.

Nobody held a leadership meeting to ask: how important is it to our business to get measurement right? What would it cost to actually know if we're creating demand or just capturing it? What people, process, and tooling would that require?

That meeting never happened. The default filled the gap.

And when you can't distinguish between demand you created and demand that already existed, you can't make a sound capital allocation decision. You can only make a confident-looking one.

Marketers got held accountable for answers they were never given the tools to find. Business growth attribution. Brand equity trends. Incremental revenue from marketing investment. Return on invested capital. Instead we report channel ROAS. Weekly CPL. Last-click conversion rate. Metrics the platform was delighted to provide, because optimizing toward them tends to mean spending more on the platform.

The 2026 Haus Decision Confidence Index surveyed 500 senior marketing and finance leaders, all managing budgets above $11 million annually. Only 49% said they measure what actually drives growth and business outcomes. The other 51% measure what's expected, visible, or easy to access. And 74% have killed or scaled back a marketing idea because they couldn't confidently measure its impact.

Organizational design failures. They just show up in measurement. The tool problem and the behavior problem have the same root: nobody made the strategic decision about what measurement was actually for.

Good ideas are being killed before they can be tested. The infrastructure was never built to defend them.

## The most important decision in marketing was never made.

We tell marketers to be data-driven. It's in every job description, every performance review, every board presentation about marketing's evolving role.

Then we hand them a free dashboard and call it infrastructure.

Finance has an audit trail by law. Sales has a pipeline by necessity. Product has analytics because shipping without them is guesswork. Marketing got a free tool and a mandate to prove ROI.

The free tools measure what the free tools were built to measure.

GA tells you where traffic came from. Facebook Ads Manager tells you what Facebook wants you to believe your ads did. Google Search Console tells you which queries brought people to your site.

All of it digital. All of it last-click. All of it pointing backward at the conversion, never forward at the category.

Analytic Partners ROI Genome found that high-performing organizations anchor measurement to commercial goals first, then assemble the data needed to answer them. Most organizations do the opposite. They inherit the tools, then build their goals around what the tools can see.

So the dashboard becomes the strategy. And the strategy becomes what the dashboard can defend.

Performance dashboards were built to help marketers make faster decisions. They accidentally became the primary artifact marketing uses to justify itself to finance. And finance, quite reasonably, takes them at face value.

I've been asked by a CFO to show the revenue contribution of brand marketing. I had awareness data. Share of voice numbers. A ROAS figure from the platform. None of it answered the question he was actually asking. I've also been asked what would happen to revenue if we cut the marketing budget in half. I gave the honest answer. I didn't know. Neither meeting went well.

The tools I had were never designed to answer those questions. I built my strategy around what they could see.

None of it tells you what caused someone to search for your brand. Whether the category is growing or shrinking. Whether brand equity is rising or eroding. Whether the base sales trend is moving in your favor or quietly working against you. Whether your media is generating incremental demand or harvesting demand that would have arrived anyway.

That level of measurement requires econometrics, incrementality testing, geo-holdouts, brand tracking, and contribution analysis. A deliberate investment in understanding the full picture, not just the part the platform can see.

Most marketing leaders never made that case internally. Some didn't know how. Some inherited a measurement culture and never questioned it. The result is the same: a function measuring the portion of its work that's easiest to measure. Capturing in-market demand. While the harder work — creating future demand — goes untracked, undefended, and first on the list when budgets get cut.

A tools problem on the surface. Underneath it, an education problem that became a leadership problem that became a growth problem.

Bigger brands with bigger budgets figured this out eventually, usually after the death spiral had already started. The argument here is simpler: you don't need to be big to start measuring the full impact of your work. You need to start before you think you need to. Because by the time the numbers make the case for you, you're already behind.

Real measurement requires people, process, purpose-built tooling, and leaders willing to make the internal case for all three, even when it's uncomfortable.

Each quarter you optimize for a portion of the whole, the harder it becomes to course correct. The measurement failure compounds the same way the brand failure does. Quietly, then visibly, then expensively.

Snickers tracked effectiveness through single-source sales data across key markets. They asked whether people bought more Snickers. The answer either justified the next investment or it didn't. No expensive attribution model required.

## Your measurement system has a point of view.

The measurement system you use doesn't just report on the strategy. Over time it becomes the strategy. It shapes what briefs get written, what ideas survive budget review, what questions a marketing leader learns to ask.

Snickers in 2009 asked: how do we regain market share?

Most of us are still asking: what does the free tool say?

There are no right answers to wrong questions.

*— Ursula K. Le Guin*

*Still thinking, JD*