• by Trevor Monteiro , 11 hours ago

We're celebrating 3x ROAS while revenue tanks. If that doesn't tell you something's broken, nothing will.

It's Q4. Dashboard's green. ROAS up 20% year-over-year. Every retargeting campaign is "performing."

And your penetration is flat. You're reaching the same 20% of category buyers over and over, while 80% have never heard of you.

Binet & Field's analysis shows it clearly: campaigns measured over 6 months appear to favor activation, but measured over 3 years, the optimal balance shifts dramatically to 60% brand building. We're not measuring wrong. We're optimizing backward.

Byron Sharp and Ehrenberg-Bass proved it: brands grow by reaching more people, more often. Light buyers and non-buyers vastly outnumber your heavy users.

We're Celebrating Last-Click Like Planet Fitness Celebrates Pizza Friday

You know those people who join a gym Jan.1st and spend three weeks perfecting their workout tracking app? Logging every rep. Color-coding spreadsheets. 100% tracking accuracy.

Zero weight lost. Zero muscle built. Zero actual progress.

That's what we're doing with ROAS. We're not measuring whether we built the brand. We're measuring whether we showed up in the last click — optimizing attributed conversions while the actual business stays weak.

The Protein Bar Paradox

Imagine you've been lifting heavy for six months. Real strength built. Then right before you max out on bench press, someone hands you a protein bar. And you credit the protein bar for your strength.

That's last-click attribution.

The retargeting ad that caught someone right before they bought gets all the credit. The brand campaign that made them aware three months ago? Zero ROAS.

Protein bars help recovery after you do the work. But skip the workouts and just eat protein bars? You're just adding weight, not building strength.

That retargeting campaign with 8x ROAS? Holdout tests show 70-80% would have converted anyway. Your incremental ROAS is maybe 2x. That brand campaign with "disappointing" 1.2x attributed ROAS? MMM over three years shows 4x by building mental availability.

You still need those protein bars. Short-term wins matter for maintaining momentum, for showing stakeholders progress while the real work compounds. The problem isn't measuring short-term results. It's mistaking them for the actual strategy.

 Measuring What Actually Works

Marketing mix modeling looks at years of sales data against all your marketing activity and external factors to show true incrementality. It's not handicapped by last-touch attribution or identity graph gaps that make people look like new customers every time they switch devices.

Modern MMM gets to insights in weeks. Updates continuously. Combined with geo tests, holdout studies, and brand tracking, you get closer to truth than any dashboard of attributed ROAS.

Yes, it's more complex than this. MMM needs good data hygiene. Reach means nothing without active attention. Category context matters. But if you're measuring wrong, you'll kill everything else before it has time to work.

Stop Celebrating. Start Testing.

Pull your top-performing retargeting campaign. The one with 8x ROAS that everyone celebrates. Run a holdout test. See how many people convert without ever seeing the ad.

Then look at that brand campaign you're about to cut. The one reaching new category buyers, building mental availability, doing the actual work.

Your dashboard will look worse before it looks better. But three years from now, your brand will be stronger, your penetration will grow, and you'll look back at those green dashboards the same way you'd look at someone who spent a year perfecting their workout tracker while never actually lifting.