Channel scoring for consumer apps: the four-dimensional model that beats CAC-only thinking

When CAC won't fall, the temptation is to throw money at whichever channel has the lowest CAC last week. That is exactly the wrong frame. CAC is a price paid for an asset (the customer). The asset has a value (LTV). And both vary by channel in ways that aren't visible until they're scored properly.

The standard failure mode

In consumer apps — gaming, services, fintech — channel mixes drift toward whichever source produced the cheapest install last quarter. Affiliate marketing tends to carry meaningful spend but with opaque mix: nobody can clearly say which affiliates send high-LTV users versus pure install volume. Installs scale, unit economics don't.

The problem isn't the channels. The problem is the scoring.

A four-dimensional channel score

Every channel gets scored on four weighted dimensions:

• QoQ growth rate — is the channel still scaling or already plateauing?
• LTV trajectory — is per-channel LTV improving, flat or degrading quarter over quarter?
• ARPU — what's the average revenue per user from this channel, calculated at day 60 and day 180?
• CAC — the obvious one, but always last in the order.

Scored honestly, the matrix usually produces surprises:

• Google often drives more growth than its CAC suggests — LTV trajectory is strong because audience intent is high.
• Affiliates tend to carry both ends of the bell curve — some deliver the highest-value users in the whole account; others deliver installs that churn inside 7 days at no margin.
• Organic is usually under-invested — flat growth there is rarely a platform problem; it's a missing ASO program.

What to do with the score

Affiliate framework rebuild

Build it around a user-engagement model. Score each affiliate on the LTV of users they deliver, not their raw install volume. Pause the bottom quartile. Scale the top quartile with direct incentive alignment.

Bottom-of-funnel Google optimisation

Sharper creative and tighter targeting on BOF Google campaigns. A few points of CAC reduction tends to come from this alone.

Meta LTV audiences

Move Meta targeting away from broad lookalikes toward LTV-based custom audiences. The audience base gets higher quality, ad costs stabilise, and CAC tends to drop materially for the cohorts that lookalike-modelling alone wasn't reaching.

ASO program

Usually the biggest single CAC win. Rebuilt rich media images, a description rewritten against a primary keyword set, on-page and off-page SEO across landing pages and the app store. Organic install share tends to lift double digits and overall CAC drops sharply — organic is structurally cheaper than any paid channel.

RFM and in-app monetisation

RFM segmentation lets the CRM team design nudges that lift LTV. Coordinate with product to ship features and in-app offers for the loyalty cohort, which tends to lift ARPU 5–10% on top.

The discipline behind the framework

If channels get scored only on CAC, the account will pay the lowest CAC for the lowest-LTV traffic, indefinitely. Score on the full quadruple — growth, LTV trajectory, ARPU, CAC — and pause the ones that look cheap but produce churners. The biggest wins are often in channels that have been under-invested for the wrong reasons.