Cash over Clicks

What this is

None of us have a lot of time to read detailed articles on tying advertising first principles to outcomes. I don’t have a great deal of time to write them either. So here’s the first in a series of short, plain‑English ways to judge digital ad spend using cash outcomes. No jargon. No glory metrics.

The Advertising First principle

Revenue and margin beat activity. If a dollar of ad spend can’t be tied to near‑term cash and medium‑term profit, it’s a bad bet.

The problem (in one minute)

  • Platforms celebrate clicks and views. Banks don’t.

  • Multiple channels claim the same sale.

  • Fees and middlemen chip away before an ad even loads.

  • “Viewable” can still mean “barely noticed.” Result: healthy dashboards, weak cash.

The fix: follow the money

Trace each dollar from invoice → impression → session → order → cash. Then answer four questions:

  1. Incrementality: What would sales have been without this spend?

  2. Payback: How many days until the exposed cohort earns back the ad dollar (after costs)?

  3. Margin: What contribution is left after media, fees, and fulfilment?

  4. Risk: How much runs on unverifiable supply or low‑quality sites/apps?

How we run the check

  • Define the window: Agree a standard payback horizon (e.g., 90 days for ecommerce; longer for subscription).

  • Cohorts, not channels: Group orders by week of exposure, then measure their cash curve over time.

  • Throttle by marginal ROI: Increase spend only while the next dollar pays back within the window.

  • Reconcile to the ledger: Match platform totals to invoices and your GL every month.

What you’ll see

  • A one‑page Cash Panel: working media %, cohort payback days, incremental ROAS, and fee stack.

  • A Supply‑Path Map: your dollar’s route to the screen, with each hop and fee.

  • A Spend Curve: where marginal ROI drops below your hurdle.

Some decision rules

  • Pause any line where payback > 90 days or iROAS < 1.0 for two straight weeks.

  • Shift budget away from paths with working media < 70% until renegotiated.

  • Require a 10–20% control in any scale test to prove lift.

What we verify

  • Data lineage: Every metric has a source and timestamp; numbers are reproducible.

  • Working media rate: Share of spend that buys actual media after all fees.

  • Overlap/duplication: Cross‑channel reach is deduped; you’re not paying for the same person twice.

  • Incrementality design: Holdouts or geo/time splits are documented and repeatable.

  • Contract fit: Terms reward verified outcomes, not just delivery.

The controls CFO controls should require

  1. Ledger reconciliation: Monthly tie‑out of platform spend to invoices and GL, signed by agency and finance.

  2. Cohort payback table: Standard report showing cumulative contribution to payback, published fortnightly.

  3. Working media threshold: Fee stack disclosure per path; target ≥70% working media, variances explained.

  4. Incrementality plan: Pre‑registered test design with pass/fail rules before scaling.

  5. Metric change control: Any change to definitions or windows requires written approval from finance.

Next up: Spend the Next Dollar, Not the Last One—how to budget using marginal returns instead of averages.

 

Next
Next

When “Optimisation” Isn’t