A retail brand kept its revenue on 25% less marketing spend

The company was spending about $200K a year across paid social, paid search, email and SMS, creator partnerships, and local activations.

On paper, things looked good. Platform dashboards showed strong ROAS on paid social and branded search. So they kept spending more.

But the business wasn't growing to match. Revenue stayed flat even as spend went up.

The problem was how they were measuring. Everything ran on last-click attribution, which gives all the credit to whatever the customer touched right before buying. That distorted the picture:

• Upper-funnel channels looked like they weren't working
• Lower-funnel channels looked like they were doing all the heavy lifting
• Nobody was measuring what moved the needle

The budget kept getting pushed toward channels that looked good in dashboards, not the channels driving new revenue.

We built a Marketing Mix Model (MMM) to measure what each channel contributed to sales.

Instead of trusting platform-reported conversions, the model looked at historical spend and total revenue to isolate how much new revenue each channel generated.

We modeled two things that attribution tools ignore.

First, carryover. Marketing doesn't stop working the moment someone sees an ad. A campaign can drive purchases weeks later. We built decay curves for each channel to capture that.

Second, diminishing returns. Every channel hits a point where spending more stops producing more. We modeled those response curves to find where each channel started to flatten out.

Then we broke total revenue into its parts: baseline revenue (brand strength, repeat buyers, organic demand), the new revenue that marketing created, and how much each channel contributed.

That gave us the layer the reporting stack was missing. We could finally see the difference between revenue that marketing got credit for and revenue that marketing caused.

Marketing was driving less of total revenue than the dashboards suggested. But the revenue it did drive was concentrated in fewer channels than anyone expected.

Here's what we found:

• Paid social was saturated. High ROAS in the dashboard, but most of those sales would have happened anyway
• Email and SMS were the most efficient channels. Best return per dollar, but underinvested because attribution made them look weak
• Branded search was cannibalizing organic. It was capturing demand that already existed, not creating new demand

We restructured the budget based on where the next dollar would generate new revenue:

• Total spend dropped from $200K to $150K
• Money moved out of saturated paid channels and into owned channels
• Branded search spend was cut back

The results after the shift:

• Total revenue held steady
• Incremental revenue from marketing went up by 12 to 18%, modeled and validated after the change
• Every dollar spent on marketing started producing more new revenue
• The business spent less and got more out of it

Three numbers became the way they made decisions going forward. Channel contribution: how much new revenue each channel creates. Marginal ROI: what the next dollar of spend in each channel is worth. And reallocation lift: how much more revenue an optimized budget would produce.

They stopped chasing attributed revenue and started measuring what marketing does. Now they spend less, hold the same topline and get more real growth from every dollar.