Most DTC brands trying to fix creative performance by making more ads are misdiagnosing the problem. The system that judges the creative is what is broken, not the creative team. Four pillars close the gap: briefing, scoring, kill criteria, feedback loop. Get them right and the same team that ships 5 ads a month starts shipping 30 to 40 a month with CPA moving the right way.

Almost every DTC founder we sit down with has the same diagnosis for the creative problem: "we need more ads." The fix gets briefed to the creative team. The volume goes up. The CPA goes up too. Six weeks later the conversation is identical except the numbers are worse.

The diagnosis is wrong. The problem is almost never that there are not enough ads in the system. The problem is that the system judging them is broken. There is no shared definition of a winner, no kill rule for losers, no feedback loop from what shipped to what gets briefed next. More creative just produces more noise, faster.

This is the version of the conversation we have with founders running $5M to $30M DTC brands when their creative output is climbing and their performance is not. The fix sits in four pillars, not in headcount. Briefing, scoring, kill criteria, feedback loop. Get those right and the same team that was shipping 5 ads a month starts shipping 30 to 40 a month with the CPA moving the right way.

The full Creative Testing System playbook is free. This post is the version of the argument that sits underneath it.

Why "more ads" is the wrong fix

When we audit accounts where creative output is high and performance is poor, the same five workflow failures show up. None of them are about volume.

The brief is doing too much work. A single line in a Slack message ("test more men's product angles, also we need black friday creative") becomes the input to a 14-variant test. The designer guesses at concept, angle, hook, and format. Half the outputs are derivative of last month's winners. The other half are unfocused. None of it produces a clean read.

There is no shared definition of a winner. The creative team thinks CTR. The media buyer thinks ROAS. Finance thinks contribution margin. The founder thinks whichever metric was best on the agency report. Three months later, the team is testing toward three different definitions of success, which is the same as testing toward none.

Losers are not killed fast enough. Underperforming ads continue running at meaningful spend because nobody made the call. The result is wasted spend that compounds and dilutes the read on the rest of the test.

There is no feedback loop. Last month's winning ad did not produce a learning, it produced a celebration. Next month's brief looks identical to the one before, not informed by what worked. The system tests the same hypothesis 11 times without realising it.

Volume is the symptom that gets noticed. The system underneath is the structural problem.

The four pillars

The Creative Testing System is built on four pillars. Each one closes a specific gap that produces the failures above. You can adopt one this week and see a result. Running all four is how you compound.

Pillar 1: Briefing

A six-field brief that turns ideas into design-ready inputs. Concept, hypothesis, angle, copy direction, designer notes, format. Each field has a specific job. The hypothesis is the field most briefs are missing entirely, and the one that produces the cleanest reads when it is in place. "We think [audience] will respond better to [angle] because [reason]" is the structure. Without it, you are testing variations rather than ideas.

Pillar 2: Scoring

A scorecard that declares winners in a defined order. CTR, then CVR, then CPA, with a minimum spend threshold that has to be hit before any ad is allowed to be called a winner. The order matters. CTR tells you whether the hook works. CVR tells you whether the promise lands. CPA tells you whether the unit economics work. Reading them out of order is one of the most common mistakes in DTC creative testing, and it produces a lot of false positives.

Pillar 3: Kill criteria

Five rules that close losing ads without ego or guesswork. Spend threshold, CTR floor, CVR floor, frequency cap, days-running ceiling. Each kill triggers a learning log entry, which feeds Pillar 4. The discipline is to kill on the rules rather than on instinct. Teams that kill on instinct keep favourites alive too long and shut new ideas down too early.

Pillar 4: Feedback loop

The learning log turns every kill and every winner into a structured input for the next brief. Brands without this end up testing the same idea repeatedly because nobody remembers the test happened. Brands with a working feedback loop compound their learnings over quarters rather than starting from scratch every month.

What this produces

When the four pillars run together, three things change in the account.

Output goes up materially. Brands that adopt the system typically move from 4 to 8 ads a month (the common baseline) to 30 to 40 ads a month inside a single quarter, without adding headcount. The bottleneck was never production capacity. It was workflow clarity.

Performance moves the right way. Because losers are killed earlier and winners are identified faster, more of the budget sits on winning creative for longer. CPA improves even as volume climbs. We have seen this pattern repeat across health, fashion, and clothing accounts, with the strongest cases showing 67 percent lifts in new customer acquisition alongside an output increase from 5 to 40 ads a month.

The team gets faster at the part that matters. Briefs become sharper because the hypothesis field forces clarity. Designers ship better work because the brief is unambiguous. Media buyers make cleaner calls because the scorecard is shared. The founder gets to ask different questions because the reports stop being a debate about whose metric is real.

What you can apply this week

If you want to take one pillar and run it without downloading the playbook, start with Pillar 2: scoring. Most teams are operating without a shared definition of a winner, and that single shift produces faster value than any other change in the system.

Define the order: CTR, then CVR, then CPA. Set a minimum spend threshold (most accounts at this stage need somewhere between $1,500 and $5,000 spend on an ad before any call gets made). Agree the floors. Get the agency, the creative team, and the founder to use the same scorecard on the same call. Hold it for two weeks. The conversations start changing inside the first week, and the way budget allocates inside the second.

When you are ready for the rest, the playbook has the brief framework, the kill criteria detail, and the feedback loop architecture in full.

The bigger picture

A lot of the public conversation in DTC creative testing focuses on the wrong layer. Format mix matters. Hook variety matters. Entity ID diversity matters (we wrote about this in the Andromeda Q2 update). But all of those operate at the level of the asset. The system that decides which assets to brief, which to ship, which to scale, and which to kill operates one layer above. That layer is where most accounts are quietly broken.

The brands compounding in 2026 are not the ones with the most ad spend or the biggest creative teams. They are the ones with the cleanest system underneath the spend, where every brief produces a clear test, every test produces a clear read, and every read produces a clear learning. That is a structural advantage, and it does not require new tooling or new hires.

Where to go next

Download the Creative Testing System playbook for the full version of the four pillars, including the brief framework template, the scorecard, the kill criteria, and the real-account results behind the system. It is free, written for $5M to $30M DTC founders, and applicable from this week.

If you want a view on what your current creative testing system is missing, book a call and we will walk through your last 90 days of creative output with you. For the broader Full Picture context on how creative output connects to acquisition cost and contribution margin, the First $100k of Meta Spend playbook is the companion piece.

Frequently asked questions

How many ads should a DTC brand test per month?

For DTC brands at $5M to $30M, 30 to 40 ad variants per month is the volume the system typically produces when the workflow is clean. Brands shipping 4 to 8 ads a month are usually constrained by briefing and feedback-loop friction, not by production capacity. The fix is structural, not headcount. Most accounts can hit 30+ variants a month without adding designers once the briefing framework and scoring discipline are in place.

What metrics should I use to score creative tests?

CTR, then CVR, then CPA, in that order, with a minimum spend threshold before any call is made. CTR tells you whether the hook works. CVR tells you whether the promise lands. CPA tells you whether the unit economics work. Reading these out of order produces a lot of false positives. Spend threshold matters because reading any of the three metrics under-volume produces noisy data. Most accounts at this stage need $1,500 to $5,000 of spend on an ad before a winner or loser call is statistically meaningful.

When should I kill a losing Meta ad?

Use five rules: a spend threshold, a CTR floor, a CVR floor, a frequency cap, and a days-running ceiling. Kill on the rules, not on instinct. Teams that kill on instinct tend to keep favourites alive too long and shut new ideas down too early. Each kill should produce a learning log entry that informs the next briefing cycle. The Webtopia playbook covers the specific thresholds in full.

What is the difference between a creative test and creative volume?

A creative test is a structured comparison of a specific hypothesis (audience response to angle, format, hook). Creative volume is the total number of ads shipped. Brands that confuse the two end up producing high volume with no learning, because the variants are not structured to read against a hypothesis. The fix is the briefing framework, which forces every brief to declare a hypothesis before any design work begins.