Intel Brief: The 10% Shift That Unlocked $200K in Revenue*

A B2B SaaS client was hemorrhaging money on Facebook ads with a 0.8x ROAS. Their agency blamed "market conditions" and iOS updates. Within 48 hours of our audit, we identified the real culprit: their attribution model was completely broken.

The Problem: Attribution Theater

The client's previous agency had set up a sophisticated multi-touch attribution model that looked impressive in presentations. But it was fundamentally flawed. They were giving equal credit to every touchpoint, which meant their brand search campaigns (which naturally have high conversion rates) were getting the same attribution weight as their cold prospecting campaigns.

This created a vicious cycle: The model showed brand search was "performing well," so they kept increasing budget there. Meanwhile, their actual customer acquisition campaigns were being starved of budget and labeled as "underperforming."

The Real Issue: Incrementality Blindness

We ran a simple incrementality test: we paused their brand search campaigns for two weeks. Revenue barely moved. Why? Because people searching for their brand name were going to convert anyway—they were already in-market and aware of the product.

The brand search campaigns weren't driving new customers; they were just intercepting existing demand. But the attribution model was giving them full credit for conversions that would have happened regardless.

The Solution: Budget Reallocation Based on True Incrementality

We shifted 10% of their total ad budget away from brand search and into cold prospecting campaigns with proper audience targeting. This wasn't a massive overhaul—just a strategic reallocation based on what was actually driving incremental revenue.

The key was focusing on campaigns that created new demand rather than just capturing existing demand. We also implemented proper conversion tracking that distinguished between brand-aware and cold traffic.

The Results: $200K in Additional Revenue

Within 60 days, that 10% budget shift generated an additional $200,000 in revenue. ROAS improved from 0.8x to 3.2x on their prospecting campaigns. More importantly, they were now acquiring genuinely new customers rather than just paying to intercept people who were already going to buy.

The lesson: Attribution models are only useful if they measure incrementality. Otherwise, you're just creating expensive theater that makes bad decisions look data-driven.