Ecommerce
You can meet your ROAS target and still lose money on every sale.
Paid acquisition managed to contribution margin and breakeven efficiency, not the return platforms report.
4.2x
Blended return on advertising spend
Ecommerce clients, last twelve months
The challenge
Ecommerce is the one sector where measurement suffers from over-counting instead of under-counting. Total the revenue claimed by Meta, Google, TikTok and your email platform and the sum regularly exceeds what your store actually earned. Each platform is accurate according to its own attribution model — and those models overlap.
Beneath that lies a simpler problem. A return-on-ad-spend target set without reference to contribution margin is just an arbitrary figure. Breakeven blended efficiency equals one divided by your contribution margin — with a thirty per cent margin you need about 3.3 times blended return just to break even. Many brands operate to a target below their own breakeven and call it success, simply because no one did the maths.
Key factors
What we plan for
Set the target from your margin
We calculate breakeven marketing efficiency from your true contribution margin after product costs, shipping, fulfilment, payment fees and returns — then set channel targets above it with a deliberate safety buffer. It takes an afternoon and is often the most valuable thing we do in the first month, because it replaces an inherited ROAS target with a figure that has a clear rationale.
Check what your conversion values really include
Reported conversion values often include money that was never yours to keep. Sales tax is the usual offender: wherever prices are shown tax-inclusive, standard store-to-platform integrations — Shopify into Google and YouTube is the typical example — pass the gross amount straight through, and nothing later removes it. At a twenty per cent rate, a reported 4.0 return becomes a real 3.33, which for a brand with a thirty per cent margin is the difference between healthy profit and simply breaking even. Shipping revenue, post-purchase discounts and unrecorded returns distort figures in smaller ways. Cross-border selling amplifies the issue, as rates vary by destination and the blended number conceals which markets are driving it. We check all of this before optimising towards any of it.
Put limits on campaigns that grade their own work
Performance Max and Advantage+ Shopping both gravitate towards your warmest, cheapest traffic and then report the results as new-customer performance. Performance Max provides no search-term visibility outside Search inventory, so brand cannibalisation is genuinely hidden inside the interface. We apply brand exclusions and campaign-level negatives, limit existing-customer share on Advantage+, and measure brand query capture externally where the platform will not reveal it.
Plan peak season around stock, not efficiency
During peak season the real limit is nearly always inventory and fulfilment capacity, not media efficiency. We build the peak plan around stock levels and dispatch capacity, with pacing rules for the build-up, the peak itself and the discount-driven period that follows — including moments where we intentionally slow spend because the warehouse cannot keep up. Selling in several markets means several peaks that do not align: shopping calendars, public holidays and courier cut-off dates all vary, so one global pacing plan will overspend in one market while underfunding another.
Services
Services we deliver here
PPC Advertising
Paid search and shopping managed for margin, not platform-reported ROAS.
Affiliate Marketing
Partner programmes restructured around incremental sales rather than last-click credit.
Performance Marketing
Budget allocated across channels by measured contribution, not by what each platform claims.
Customer Acquisition
Acquisition economics modelled to a payback period, with media purchased to match.
Campaign Management
Profit-focused campaign planning, pacing and reviews run as a steady routine.
Common questions
Our return on ad spend is already strong. Why should we change anything?
Because return on ad spend is measured against revenue, while suppliers are paid from margin. The first question is what your breakeven blended efficiency really is, based on your contribution margin. If your target is below it, the account is losing money efficiently. If it is comfortably above, we will confirm the account is healthy and focus our efforts elsewhere.
Is server-side tracking really necessary?
If you spend significantly on Meta, yes — and the reason is bidding quality, not reporting. Low event match quality weakens the signal the algorithm learns from, which appears as poorer performance rather than as gaps in a report. It is unglamorous infrastructure work with a direct impact on results.
Should we reduce brand search spend?
Test it rather than assume. Some branded paid conversions would have come through organic results anyway and some would not, and the balance varies widely by category, competitor bidding and how your organic listing appears. A geo holdout running four to six weeks answers the question for your brand specifically. Two-week tests are where most of these questions get the wrong answer.
How do you approach peak season?
Planning begins well ahead of peak, not in the final weeks. Media costs rise as peak approaches, promotional periods now stretch over weeks rather than a single weekend, and the real constraint is usually stock and dispatch rather than budget. We pace spend against inventory, market by market where you sell in several, and agree in advance where we will slow down.
Begin with an audit.
The audit has a clear scope and a clear output, and it stands on its own — no obligation to continue. If it shows your current setup is working well, that is a perfectly valid result and we will tell you so.