Case #021: The scale that broke at $5k/day

Case #021: The scale that broke at $5k/day

Everything worked at $1,200 a day. A nutra offer, native traffic, clean 68% ROI for three straight weeks. The temptation was obvious: pour gasoline on it.

We scaled the daily budget from $1,200 to $5,000 over four days. Classic mistake, made slowly enough to feel responsible.

Day 1 of the scale: $1,800 spent, ROI held at 64%. Encouraging. Day 2: $3,100 spent, ROI dipped to 41%. Day 3: $4,600 spent, ROI cratered to 9%. The same campaign that printed money was now barely breaking even.

The instinct was to blame the algorithm. The real story was in the placements report. At $1,200/day the traffic source fed us its premium inventory. At $5,000 it ran out of good placements and started filling the budget with garbage — junk apps, accidental clicks, sub-publishers with 80% bot rates.

We pulled back to $2,400 and added a placement blacklist of 340 sub-IDs that converted at zero. Then we scaled again, this time slower and with the filter active.

The second attempt:

— $2,400/day for a week, then $3,600
— blended ROI of 52% across $26,000 spent
— never touched the junk inventory again

The lesson: scaling doesn't dilute your campaign, it dilutes your traffic quality — blacklist before you spend, not after.