Latest YouTube · How to Scale an eCommerce Brand Profitably in 2026: The Full System / Podcast · The Fashion Creative Playbook Most Brands Get Wrong / Writing · The case against MER as a primary KPI for acquisition View all →
Method

How we think about paid media.

This is the working version of our worldview. It is the same document our team uses to make decisions inside client accounts, written for operators who would rather understand the model than be sold to. Every position below is testable, falsifiable, and applied.

01 Diagnose the constraint 02 Three lenses, one P&L 03 Causality over credit 04 Concepts over variations
// Jump to Thesis Three lenses Measurement Creative Account structure Operating principles
01 / Thesis

Most brands diagnose the wrong problem.

The default story is that the platform stopped working. The reality is usually somewhere else: creative is too similar to differentiate, measurement is crediting the wrong levers, margins cannot absorb the spend ladder, the offer is not strong enough, or the account is built to hoard signal rather than learn from it. The first move on any engagement is finding the real constraint, then sequencing the work to remove it. Everything else is noise.

It is rarely

A platform problem.

Meta and Google work. They are noisy, biased, and credit themselves generously, but they work. If platform performance moves and the business does not, the problem is upstream.

It is usually

A diagnosis problem.

The wrong constraint gets a confident answer. Spend goes up, creative goes up, segmentation goes up. The number does not move. Then "Meta is broken" enters the conversation.

The fix is

Order of operations.

Diagnose first, sequence second, spend third. Almost every decline we have inherited was a diagnosis that never happened, dressed as activity that did.

02 / Three lenses

Three lenses on every account we run.

The same three diagnostics, every engagement, in the same order. Each lens has its own failure modes. Each one breaks differently. Each one needs a different fix.

Lens 01 · Technical account structure

Designed to produce insight, not just spend money.

Test whether segmentation is useful or excessive. Whether budget is concentrated enough to produce a readable signal. Whether the account is built so the system can actually decide what is working.

We askIs this built to produce insight, or just to spend?
Lens 02 · Creative velocity & diversity

Concepts that move spend, not variations that don't.

Under modern Meta, creative does the targeting. Volume alone is not enough. Real growth comes from genuine conceptual diversity across persona, angle, offer and format. Variations of one of those four are not concepts.

We askDo we need a new concept or an iteration?
Lens 03 · Data integrity & assurance

Numbers that reconcile with the P&L.

Judge channels by acquisition MER, not blended. Use incrementality, not attribution, as the upper bound. Reporting is meant to change decisions, not validate them.

We askDo these numbers survive contact with the P&L?
03 / The measurement standard

30-day LTGP : CAC is the operating metric.

Gross profit from the first purchase within 30 days divided by customer acquisition cost. It reconciles with the P&L, it cuts through platform credit games, and it gives a clear read on whether acquisition is healthy in absolute terms, not just better than last week.

Range / Below 1
< 1.0

Losing money on acquisition

First-order economics are negative. Spend is borrowing against an LTV assumption that may not arrive. Usually a margin or offer problem, not a media problem.

Range / Weak
1.0 – 2.0

Inefficient or under-pressure

Marginally profitable on first order, but constrained. Either pricing, creative, or account structure is leaving room. Investigate the constraint before scaling spend.

Range / Healthy
2.0 – 3.0

Strong and scalable

The right place for most acquisition-led brands. Profitable on first order with room to invest in growth. The goal is to stay here while spend climbs.

Range / Above 3
> 3.0

Potentially under-scaled

Acquisition is over-efficient, which usually means money is being left on the table. The brand could absorb more spend without breaking the economics.

/ Principle

Acquisition MER > blended MER

Returning customer revenue distorts the read on whether acquisition is doing its job. Separate them at the source, judge each on its own.

/ Principle

7-day click as default

1-day view attribution inflates Meta credit by 20–40% in most accounts. Use it as a sanity check, not as the optimisation signal.

/ Principle

Causality > credit

Geo-lift and in-platform incrementality tests bound what the platform can legitimately claim. Where it disagrees with reality, reality wins.

04 / The creative standard

Concepts move spend. Variations don't.

A concept is a combination of four variables: persona, angle, offer, format. Change one of those and you have a new concept. Change the headline or the colour of a button and you have an iteration. Most accounts are testing iterations and calling it creative testing. The delta in performance lives at the concept level, not below it.

Variable 01 / Persona
Who

Who you are talking to

The customer segment you are addressing. Founder vs operator. New buyer vs returning. Frustrated vs aspirational. Each persona forces a different angle, offer, and format combination.

Variable 02 / Angle
Why

The argument you are making

The reason this product, for this person, at this time. Pain point. Mechanism. Status. Identity. The angle is what you are actually trying to convince someone of, before you ask them to buy.

Variable 03 / Offer
What

The commercial proposition

What they get and on what terms. Bundle, discount, trial, financing, risk reversal. The offer is the variable most brands under-invest in, and the one that most often unlocks new spend ceilings.

Variable 04 / Format
How

The container the argument lives in

Talking head, UGC, listicle, static, animation, brand film. Format is a concept variable, not a production decision. Different formats serve different angles. Testing format diversity is testing concept diversity.

05 / The account standard

Concentration over segmentation.

Most accounts are over-segmented to the point where nothing is readable. Split budget across too many ad sets on Meta, or too many campaigns and asset groups on Google, and every call is made on noise, with too few conversions behind any one of them to tell signal from variance. This is not about chasing a learning phase. It is about giving each platform's system enough concentration to allocate budget well, and giving you enough signal to draw clean conclusions about what is actually working.

Principle

Signal per decision

The point of the account is to teach the system, and us, what works. Excess segmentation buys precision that the data cannot support. Consolidate until each meaningful decision has enough signal behind it to be confident.

Principle

Let the system allocate, then fence it

Meta, Google & TikTok all distribute budget better than manual management once they have volume, but they will spend into brand and existing demand and credit-grab if you let them. Concentrate spend, but fence the automation: separate brand from non-brand, new customers from existing, and read true incremental acquisition.

Principle

Hyper-segment only with a reason

There are cases where segmentation is justified. Geo, language, product line with materially different economics, audience with a different commercial value. These are decisions, not defaults. Most are defaults.

06 / Operating principles

How we think about every decision we make on your account.

Four mental models that sit underneath the lenses, the metrics, and the frameworks. We come back to these whenever the work gets noisy.

/ Mental model 01

  • The bottleneck changes with stage. What gets a brand from $5M to $20M is not what gets it from $50M to $100M. The constraint moves. The advice has to move with it.
  • Generic advice is almost always wrong by the time you apply it. The question is which constraint you are at, not what the playbook says.

/ Mental model 02

  • Most teams misdiagnose the problem. A platform problem is usually a creative, measurement, margin, or structural problem in disguise.
  • Time spent on the wrong problem is the most expensive line item in the business. Diagnose hard.

/ Mental model 03

  • Good strategy must reconcile with the P&L. If the ad account looks good but the business is not improving, something is wrong. Find the disconnect, not a justification.
  • Vanity metrics are confessions disguised as wins. We do not celebrate them.

/ Mental model 04

  • Better data should lead to better decisions. Reporting that does not change behaviour is overhead. The test for any dashboard is whether the next decision changes because of it.
  • If the number does not move the meeting, it should not be in the meeting.
If this matches how you think

Run the same diagnostic on your own account.

The free strategy session and audit applies this method to your account. Used by founders, finance, and marketing to make better decisions whether we work together or not.

Book a call →