Where the Next User Acquisition Dollar Should Go

The decision in one sentence: Put the next incremental UA dollar behind the campaign with the strongest marginal cohort ROAS and enough scale headroom, not simply the lowest CPI or the highest blended ROAS.

That distinction matters because the metrics most teams review are averages. Average CPI describes what has already been bought. Blended ROAS combines customers who would have converted anyway with customers created by the next dollar. Neither metric, by itself, tells a budget owner what the next unit of spend will return.

The comparison that should drive the decision

Use a dated, cohort-matched view. Compare campaigns over the same acquisition window, normalize revenue to the same maturity point, and separate observed efficiency from the capacity to absorb more spend. The example below is illustrative and must be replaced with current account data before publication or use in a live budget decision.

Illustrative figures only. Replace with one source of truth, a fixed reporting cutoff, and the same revenue maturity window for every campaign.

Decision from the illustrative snapshot  

Snapchat’s lower illustrative CPI, positive D30 ROAS, and high scale headroom justify a place in the mix and a test of channel-specific marginal returns. Retargeting has limited headroom, while TikTok merits only a controlled test until cohort returns improve.

Why CPI cannot make the call alone

CPI is a price signal, not a value signal. A lower CPI can indicate efficient reach, but it can also reflect lower-intent inventory, weaker geographies, or users who monetize slowly. A higher-CPI campaign can still create more profit if it acquires customers with stronger retention and revenue.

The useful question is not “Which campaign buys the cheapest install?” It is “Which campaign can turn one more dollar into the most incremental, risk-adjusted value?” 

That question forces CPI, ROAS, volume, incrementality, and measurement confidence into the same decision.

The four checks before moving budget

1. Match the cohorts  Use the same acquisition dates, attribution rules, geography, platform fees, and revenue horizon. Comparing D7 ROAS for one channel with D30 ROAS for another produces a ranking, but not a decision.

2. Estimate marginal return  Ask what ROAS is likely at the next spend level. If performance deteriorated as the budget rose, the blended average overstates the value of the next dollar.

3. Confirm scale headroom  Check impression share, audience saturation, frequency, auction pressure, and creative supply. A high-return campaign with no room to grow cannot absorb the allocation.

4. Discount weak evidence  Reduce confidence when conversion volume is thin, postbacks are incomplete, holdouts disagree with attributed results, or recent creative and bid changes have not matured.
   
   
   

A decision rule teams can operationalize

Rank each eligible campaign by expected marginal ROAS, then apply three gates: the return must clear the company’s hurdle rate, scale headroom must be real, and the evidence must be strong enough to justify the size of the move. Fund the highest-ranked campaign in a measured increment, then re-read performance after the cohort reaches the chosen maturity window.

Turn the allocation into a learning loop

Move enough budget to produce a detectable result, but not so much that a wrong assumption becomes expensive. Define the increment, owner, start date, stop condition, and review date before launch. Hold creative and measurement changes constant where possible. At review, compare actual marginal performance with the forecast and either continue, cap, or reverse the move.

A practical allocation record should capture: the winning campaign, baseline and new daily spend, expected marginal CPI, expected marginal ROAS, minimum acceptable ROAS, measurement window, confidence level, and the date of the next decision.