Paid media that answers to your margins

Paid media management for an ecommerce brand means running acquisition across paid social, paid search and shopping as one budget, then judging the whole thing on blended results. The honest measure is contribution margin after cost of goods, shipping and ad spend, reconciled to the bank account, not the ROAS each platform reports for itself.

Why platform reported ROAS looks better than your bank account

Every platform grades its own homework. Meta measures the revenue Meta believes it caused, inside Meta’s attribution window. Google measures the conversions Google’s tag records, including branded searches from people who already knew your name and were going to buy anyway. TikTok does the same for TikTok. Add the three dashboards together and the reported revenue usually exceeds the revenue that actually hit the bank, sometimes by a lot.

None of this is deception. Each platform is answering a narrow question about itself, and no platform is responsible for the whole picture. The problem starts when those self graded numbers become the decision metric. A media buyer optimizing to platform reported ROAS will move budget toward whichever channel claims the most credit, and the channel that claims the most credit is usually the one with the loosest attribution settings, not the one buying the most new customers.

Platform reported ROAS is the conversion value a platform attributes to itself divided by the spend inside that platform. It is useful for comparing two ad sets against each other inside the same account on the same day. It is not useful for deciding how much a channel is worth to the business.

Blended ROAS is total revenue from every source divided by total paid media spend across every platform. It ignores attribution entirely, which is exactly its value. It cannot be inflated by a window change, and it moves when the business moves.

MER (marketing efficiency ratio) is total revenue divided by total marketing spend. Many teams use it interchangeably with blended ROAS. The distinction worth keeping is scope: blended ROAS usually counts media spend only, while MER often includes everything marketing costs (management fees, tooling, affiliate and influencer payouts, retention channels). MER is also frequently reported as its inverse, spend as a percentage of revenue, so a brand running at 3.4x may describe itself as sitting at 29%.

Contribution margin is what is left after cost of goods, shipping and fulfillment, payment processing, discounts and returns, and ad spend. It is the only number on this list that tells you whether the growth was worth having.

Getting to a blended number you can trust is a measurement problem before it is a media problem. If the pixel is firing twice, if server side events are not deduplicated, or if new and returning customers are not separated, then every decision downstream inherits the error. That work sits in the tracking and attribution rebuild and it usually happens first.

What paid media management covers here

Four workstreams, run as one plan rather than four vendors emailing each other.

Paid social, full funnel

Meta is usually the core of an ecommerce media plan, but this is not a Meta and Google shop. TikTok and the rest of the ecosystem earn budget when the audience and the economics justify it. The work covers account structure and consolidation, prospecting and retargeting balance, audience and exclusion hygiene, offer and landing page alignment, and the unit economics underneath it all: average order value, new customer acquisition cost, contribution margin per order, and repeat purchase behavior. A campaign that looks efficient on order value can be unprofitable on margin if it is selling your worst products.

Paid search and shopping

Brand defense first, because it is the most argued about line in the account. The question is not whether to bid on your own name, it is what share of that traffic you would keep for free, who is bidding against you (competitors, unauthorized resellers, affiliates arbitraging your terms), and what that defense is worth in margin. Beyond brand: non brand and category capture, shopping feed optimization (titles, attributes, imagery, product level bidding, excluding out of stock and low margin variants), and governance on automated campaign types so they are not quietly absorbing branded demand and reporting it as new.

Cross channel budget allocation

Spend follows margin. Budget is reviewed weekly against blended performance and reallocated across platforms rather than defended per platform. Channels do not own their budgets. The business owns the budget and channels earn it.

Creative pipeline integration

Media plans fail when creative arrives late and untested. The media calendar and the testing calendar are the same calendar: hypotheses, one variable per variant, documented kill criteria and scale criteria. The performance creative testing system runs alongside the media, with production staying with your own creators and editors. Plaid Testing runs the system that tells them what to make next.

How budget decisions get made against contribution margin

1. Establish a baseline you can defend. Reconcile platform reported numbers to real revenue and to the bank account. Post purchase survey data and attribution tooling stay directional inputs, not verdicts.
2. Set the margin floor. Calculate contribution margin per order after cost of goods, shipping, processing, discounts and returns, then translate that into the blended ROAS or MER the business needs to clear at the growth rate ownership actually wants.
3. Choose the trade deliberately. Scale almost always costs efficiency. The only useful question is how much efficiency, for how much volume, and whether the result still clears the floor. That is a decision for the owner, taken with real numbers, not a surprise discovered in a quarterly review.
4. Reallocate weekly on blended results. Not on which dashboard is claiming the most credit this week.
5. Kill and scale on written criteria. Thresholds agreed in advance, applied consistently, so nothing gets turned off because of one slow Tuesday or scaled because of one lucky weekend.

The practical output is that cheap traffic and profitable traffic stop being confused with each other. Plenty of brands find their highest ROAS products are among their least profitable, and that the campaigns worth defending are not the ones the platform ranks first.

If you want that math run on your own numbers before you commit to anything, that is what the free 30 minute growth audit is for. Three specific fixes, yours to implement with or without us.

What the reporting cadence looks like

Three rhythms, so nothing is discovered late.

• Daily snapshot. Spend, blended revenue, blended ROAS and MER, new customer acquisition cost, and anything that moved enough to explain. Sent to you, not parked in a dashboard nobody opens.
• Weekly testing readout. What was tested, what the hypothesis was, what happened, what was killed, what is being scaled and why.
• Monthly strategy session. Margin performance against plan, channel mix, creative direction, inventory and promotional calendar, and the plan for the month ahead.

You own the ad accounts, the pixels, the data and the creative. You see the same numbers on the same day they are seen internally, including the weeks that go badly. Daily transparency is only meaningful if it survives a bad week.

What happens in the first 30 days

The first month is a rebuild, not a heroic performance chart. Weeks one and two go to measurement: verifying pixel and server side event setup, deduplicating conversions, separating new from returning customers, and reconstructing reporting until platform numbers, store numbers and bank numbers can be explained against each other. Weeks two and three go to structure: consolidation, exclusions, budget architecture, feed cleanup, and the brand defense decision. Weeks three and four bring the creative pipeline and the reporting cadence live.

By day 30 you should have a media plan tied to a margin floor, a testing calendar tied to that plan, and numbers you can quote to a board without adding a footnote about attribution. Scale comes after the foundation, because scale on a broken foundation just buys the mistake faster.

Is Plaid Testing a paid media agency?

No, and it is worth being straight about that, because most people searching for a paid media agency want the outcome rather than the org chart. Plaid Testing is an operator model. Jason Lu is a Meta Business Partner who presented “Meet Moby 2” on Triple Whale’s Customer Education Series and has spoken on panel at The Whalies. He is the person who builds the plan, and he is the person in the account.

This is not a criticism of any particular agency, and plenty of them do good work. It is a criticism of a structure. In the standard model the senior person who wins the account is rarely the person operating it a quarter later, the reporting layer sits between you and the raw data, and the commercial incentive quietly favors numbers that make the retainer look safe. Remove the layers and the incentive changes with them. There is no long term lock in and no retainer required to start.

Where a broader mandate is needed, above the ad account rather than inside it, that becomes a fractional CMO engagement covering strategy, channel oversight, hiring and vendor decisions. Most brands start with one and grow into both.

Who this fits, and who it does not

It fits ecommerce brands investing $50K or more per month in paid media, most often in apparel and athletic wear, fashion and accessories, or wellness and supplements. It fits founders who already suspect their platform numbers are flattering, who carry real inventory and margin complexity, and who are willing to change account structure and measurement rather than just asking for more ads.

It does not fit brands under roughly $50K per month, where the honest advice is that senior media management does not yet pay for itself and the money is better spent on offer, creative volume and retention. It does not fit anyone who wants their existing ROAS number defended rather than examined, anyone unwilling to share cost of goods, or brands still searching for product market fit. Saying so before an engagement starts is cheaper for everyone than discovering it in month three.

What this looks like in real accounts

An athletic apparel brand grew sales 35.7% to $9.27M from January to June 2026 while paid media investment rose 42%, holding blended ROAS at 3.36x across the period. That is the clearest example of the whole thesis: efficiency was deliberately held flat rather than improved, because the margin still cleared the floor and the volume was worth more than the ratio. Managed against platform reported ROAS, the same six months would have looked like a reason to pull back.

A fashion apparel brand grew sales 249% year over year with net profit up 205%, finishing at 41% net margin with MER at 29% (paid media absorbing 29 cents of every revenue dollar). A womens fashion brand grew sales 99% while new customer acquisition cost fell 21% and new customer ROAS rose 58%, lifting net profit 136%. Net margin there still sits around 3%, and that is stated plainly on purpose: a brand can double sales, improve every efficiency metric it tracks and still run thin. That is precisely why margin, not ROAS, is the unit of reporting.

Every figure traces to platform exports held on file. Brands are described by category rather than named because the numbers belong to them. The longer versions, including what was changed and in what order, are in the full case studies.

Get your numbers looked at first

The first step is not a proposal. It is a free 30 minute growth audit: a review of your tracking, a read on your account structure, and three specific fixes you can implement whether or not anything else follows. No retainer is required to start and nothing is pitched on the call. Book your free growth audit and bring the numbers you do not trust.