Scaling spend without scaling clarity.
Like most growth-stage CPG brands, this company was expanding media across retail and digital channels — but flying partially blind. Their existing measurement couldn’t reliably identify which dollars drove incremental revenue versus those merely correlating with it. With mounting budget pressure and expanding retail distribution, they needed one clear answer: where should the next dollar go?
Five months. Compounding evidence.
YoY comparison, August – December 2025. The brand didn’t just spend more — they spent smarter. Spend scaled 2×. Incremental returns scaled 3×.
Total Revenue
+114%
$37M → $79M · Revenue more than doubled
iROAS improvement
+40%
2.75 → 3.85 · Every dollar working harder
Incremental Revenue
+192%
$9.4M → $27.5M · Causal lift, not correlation
| METRIC | BEFORE | AFTER | CHANGE |
|---|---|---|---|
| Total RevenueAll channels combined | $37M | $79M | ↑ 113.9% |
| Retail RevenueBrick-and-mortar + retail media | $15M | $39M | ↑ 160% |
| Incremental RevenueRevenue attributable only to media | $9.4M | $27.5M | 192.1% |
| Retail Incremental RevenueIncremental from retail channel only | $2.7M | $10.1M | ↑ 277% |
| Total Media SpendAcross all paid channels | $3.4M | $7.1M | ↑ 108.6% |
| iROAS — All ChannelsIncremental revenue per $ of media spend | 2.75 | 3.85 | ↑ 40% |
| Retail iROASRetail channel incremental return | 0.78 | 1.42 | ↑ 80.9% |
Three things that set this apart.
- 01
Speed to optimization
LiftLab’s forecasting output was actionable from day one — no months-long modeling lag before the brand could act. Budget reallocation began in week one, not quarter two.
- 02
Full-funnel visibility across brand and performance
The team measured both brand and performance impact in a single model. This let them justify retail investment on the P&L with confidence — not gut feel, not platform dashboards that overclaim credit.
- 03
Incrementality as the only north star
iROAS — not platform-reported ROAS — drove every budget decision. iROAS measures only the additional sales generated by media spend, stripping out halo and organic effects. That discipline is what separates a 40% efficiency gain from correlation dressed as causation.
How we measured this.
We publish experiment results without client names because the measurement evidence should stand on its own. Here is what stood behind these numbers.
