How much should a new DTC brand spend on ads in the first month?

Enough to buy verdicts on five to eight creative concepts, sized at two to three times your breakeven CPA per concept. On a store with a €17.19 breakeven CPA that is roughly €35 to €50 per concept, so a few hundred euros of pure testing spend plus whatever keeps running on early winners. Treat it as an information cost, judged by what it proves, not by its ROAS.

What are kill criteria in a paid media launch?

Kill criteria are the conditions, written before any spend, under which a creative concept, a campaign, or the whole launch stops getting money. A typical concept-level rule: full test budget spent with zero orders, or CPA stuck above breakeven with no downward trend. They exist because the moment of judgment always feels ambiguous from inside; the rule decides so optimism does not.

What ROAS should a new DTC store expect in the first 90 days?

There is no universal number, because breakeven differs per store. Derive your own: compute contribution margin per delivered order, divide AOV by it, and that is the ROAS you must beat. Expect to run below target during weeks one and two while testing, approach it after consolidation, and judge the launch on whether month three holds the target you derived, not a benchmark.

When should a new store start scaling its ad budget?

When four gates clear at once: blended results hold at or above target ROAS for a full week, cash actually returns inside your payback window, you have more proven creative than current spend can feed, and inventory can absorb the extra orders. In a typical launch that is month two at the earliest. Raise in steps of 20 to 30 percent and let delivery settle between steps.

Should email flows be live before launching paid ads?

Ideally the first two, welcome and abandoned checkout, before the first ad runs; they capture value from day-one traffic that paid media has already billed you for. In practice we accept flows going live by the end of month one, built while the creative test runs. What we never accept is scaling spend into a store with no automation: that multiplies a known leak.

What is a creative test matrix?

A grid of angles against hooks. Angles are the distinct reasons someone buys: the problem solved, the mechanism, the proof, the value math, the identity. Hooks are the first three seconds that earn attention on each angle. Crossing three to four angles with two to three hooks gives six to twelve ads, each testable as a concept with its own budget and its own verdict.

What happens if a DTC launch is not profitable after 90 days?

Day 90 forces one of three honest outcomes. If blended CPA holds at or under your allowable number, scale through the gates. If ads execute but CPA sits stubbornly above breakeven, the problem is usually the offer: raise AOV, bundle, reprice, or cut unit costs, then retest. If per-order net stays negative at any ROAS you honestly believe, stop. Day 90 is the cheap place to stop.

DTC Launch Plan: The First 90 Days of Media, Week by Week

What should the first 90 days of a DTC launch look like?

Week 0 is paper: breakeven math, a launch inversion, and written kill criteria before any spend. Weeks 1-2 test angles against hooks with a fixed budget per concept. Weeks 3-4 consolidate winners and switch on email flows. Month 2 scales through gates. Month 3 delivers a verdict: scale, fix the offer, or stop.

TL;DR

The launch is decided in week 0, on paper: contribution margin, breakeven CPA, and kill criteria written before the first euro moves.
Weeks 1-2 buy verdicts, not profit: three to four angles times two to three hooks, each concept funded at 2-3x breakeven CPA.
Weeks 3-4 consolidate winners into one scaling campaign and put the first email flows live.
Month 2 raises budget only through gates: efficiency, payback, creative supply, inventory.
Month 3 ends in one of three verdicts: scale, fix the offer, or stop. All three beat drifting.
Launches rarely die of bad products. They die of unpriced math and unwritten kill criteria.

The 90-day plan at a glance

One table, the whole quarter. Every phase has a job and an exit condition.

Phase	Days	The job	Exit condition
Week 0	Before spend	Paper math: breakeven CPA, launch inversion, written kill criteria	Numbers you actually believe
Weeks 1-2	1-14	Creative test matrix in a testing lane, fixed budget per concept	Every concept has a verdict
Weeks 3-4	15-30	Consolidate winners into one scaler; first email flows live	One scaling campaign holding CPA near allowable
Month 2	31-60	Gated budget steps, second angle wave, retention layering	Target ROAS holding through 20-30% raises
Month 3	61-90	Actuals against the week 0 plan	One decision: scale, fix the offer, or stop

Week 0: the launch happens on paper first

No decision inside the ad account will matter as much as the ones made before it opens. Week 0 produces three numbers and one document, and none of them cost a euro.

First number: contribution margin per delivered order, computed from supplier quotes and courier rate cards instead of hope. Strip VAT off the top, subtract product cost, fulfillment, payment fees, and packaging, then weight for the parcels that come back; the full chain is in the breakeven ROAS article. A €40 AOV store with typical EU costs and a 6% COD rejection rate lands at €17.19 of real margin per placed order. That one number prices the whole launch: breakeven CPA of €17.19, breakeven ROAS of 2.33x, and an allowable CPA of €11.19 if you want 15% net margin.


Breakeven CPA  = contribution margin per delivered order
Allowable CPA  = breakeven CPA - profit kept per order
Concept budget = 2-3 x breakeven CPA

Second: the launch inversion. Fix a monthly profit goal, an expected blended ROAS you honestly believe, and your fixed costs; the required orders per day and spend fall out. The same store aiming for €5,000 monthly profit over €1,500 of fixed costs at 3.0x needs about 55 orders a day and roughly €22,500 in monthly spend. If 55 a day reads like fantasy for your offer, good: you just fixed the offer on paper instead of meeting the same fact in month 2's P&L. Our breakeven calculator runs the margin math and the inversion in the browser.

Third: kill criteria, written down while you are still calm. What makes a concept dead. What makes a month a failure. What day 90 must show for the launch to continue. Boring to write, priceless to hold, because by week 6 every number will feel ambiguous from inside.

Weeks 1-2: buy verdicts with a creative test matrix

The first two weeks of spend have one job, and it is not profit. It is verdicts.

Build the matrix before launch day: three to four distinct angles crossed with two to three hooks each. Angles are the reasons someone buys: the problem it kills, the mechanism, the proof, the value math, the buyer's identity. Hooks are the first three seconds that earn the stop on each angle. Crossed, that is six to twelve ads, and every angle-hook pair is a concept with its own budget and its own verdict coming.

Fund each concept at two to three times breakeven CPA, the testing floor we size every account by. On the €40 store that is roughly €35 to €50 per concept: enough to prove it can find orders near allowable CPA, cheap enough that a loser costs a dinner, not a payroll. Run the test on ad set budgets so every concept is guaranteed its read. A shared optimized budget will quietly execute half your matrix at €4 of spend each, before anything gets judged.

What counts as a signal at this volume? An order or two arriving near allowable CPA is real. Click-through cost and product-page conversion are tiebreakers between concepts, never verdicts on their own. And the blended account number will look rough for these two weeks. It is supposed to. This is the information cost you priced in week 0, judged by what it finds.

Then the criteria you already wrote do the deciding. Full concept budget spent with zero orders: dead. CPA above breakeven, not trending down: dead. At or under allowable: a winner, and week 3 has a job for it.

Weeks 3-4: consolidate winners, switch on retention

By day 15 the matrix has spoken and the account changes shape. Winners graduate into one consolidated scaling campaign, broad, carrying roughly 70-80% of budget; the testing lane keeps running on the rest. It is the same two-lane structure we run on mature accounts: pooled conversion signal exits the learning phase, fragmented signal never does. The full doctrine is in our Meta account structure guide, and at launch volume it matters more, not less: a store doing ten orders a day cannot afford to split them six ways.

Feed the winner. The week 3 temptation is to keep resuscitating the losers you privately liked. The discipline is to put money behind what the auction endorsed and let the matrix keep producing challengers.

These two weeks also close the retention gap. The first email flows go live now if they were not live at launch: welcome and abandoned checkout first, the two closest to the money, with abandoned cart and post-purchase behind them. Email for an established DTC store is commonly benchmarked at 25-35% of revenue. At day 20 yours will earn a fraction of that, but every week of paid traffic into a store without flows is margin you billed for and never collected. Build order and per-flow setup live in the Klaviyo flows checklist.

Day 30, if the plan holds: one scaler holding CPA near allowable, one testing lane, the first flows earning, and a written record of what died and why.

Month 2: scale through gates, not by calendar

Month 2 is where launches get greedy, so we run it on gates. Before any budget step, four things must be true at once:

Efficiency. Blended results hold at or above target ROAS for a full week, not one good afternoon.
Payback. Cash has actually returned inside your payback window. Ad platforms bill within days; COD remittances and payment payouts land weeks later. ROAS does not pay invoices. Remitted cash does.
Creative supply. More proven concepts than current spend can feed. Scaling one winner just fatigues it faster.
Operations. Inventory and fulfillment absorb the extra orders without delivery times slipping, because slow delivery raises COD rejection rates and quietly rewrites the margin math.

Gates clear? Raise budget 20-30%, let delivery settle for a few days, check again. One gate fails? Fix that constraint and leave the budget alone. More money into a failed gate converts a bottleneck into a loss.

Month 2 also launches the second angle wave. The week 2 winners will fatigue, and a replacement cannot be started the day the winner dies. Same mechanism, fresh angles and hooks, through the same testing lane and the same kill criteria.

Retention keeps layering: winback, back-in-stock, sunset, review request. The first review requests matter beyond their revenue: by month 3 they are feeding the product page proof every cold click lands on.

Month 3: the honest review

Day 90 exists to force a decision, and the judge is the week 0 document. Put actuals against paper: real courier invoices against assumed fulfillment, real rejection rate against the modeled 6%, blended CPA against allowable, payback against the window, creative hit rate against the matrix.

Three outcomes. All of them beat drifting.

Scale. Unit economics at or above plan, gates clearing, pipeline producing. Raise budgets through the same gates, open the next angle wave or market, keep the weekly ritual.
Fix the offer. The ads execute, the clicks arrive, the page converts, and blended CPA still sits above breakeven. That is rarely a media problem. It is an offer problem: AOV too low, margin structure too thin, price wrong. Raise AOV with bundles, cut landed cost, reprice, then go back to paper and retest. Media buying cannot outrun a €9 contribution margin.
Stop. Per-order net stays negative at any ROAS you honestly believe, even after offer surgery. Stopping at day 90 costs one quarter of controlled testing. The same lesson in month 9 costs the year, plus the inventory.

Notice the option that is missing: keep going and see. A launch without a verdict date does not end. It just gets quieter.

The weekly review ritual

Every one of the twelve weeks gets the same 30-minute review. Same day, same order, because the early numbers veto the late ones.

A calm desk with coffee, a notebook with a hand-sketched rising chart and a blank clock, representing the weekly launch review ritual

Spend against plan. Did the money move as designed, or did a paused ad quietly underspend the test?
Blended CPA, or MER, against allowable. The single health number; above allowable and trending wrong, it vetoes everything below.
Verdicts due. Which concepts finished their funded budget this week? Kill or promote per the written criteria. No extensions granted out of affection.
Cash and payback. What couriers and processors have actually remitted, held against what the platforms have billed.
Pipeline. How many concepts are queued for the testing lane? Fewer than two is an emergency that arrives in three weeks wearing a fatigue curve.
One written decision. Every review ends with a dated line: what changes this week, and why.

Thirty minutes is enough because the thinking already happened in week 0. The review compares; it does not deliberate. The moment it turns into a debate about what breakeven "really" is, the launch has lost its instrument panel.

The common launch deaths

We have watched more launches die of process than of product. Five causes keep returning:

Spending to "learn" without kill criteria. A learning budget with no verdict date is not a test, it is a subscription to hope. The fix costs one page of writing in week 0.
Scaling on ROAS before payback proof. The dashboard says 3.2x, the bank account says empty, because ads billed on Tuesday and the COD remittance lands in three weeks. Working capital sets the speed limit, not ROAS.
Verdicts at €10 of spend. Killing a concept at half a breakeven CPA is not filtering losers, it is executing winners early. Fund the verdict or skip the test.
Re-fragmenting in week 5. One flat week and the account sprouts six new campaigns. Signal splits, learning resets, and the structure that worked gets blamed for the panic that broke it.
Retention postponed until "after we scale". Scale multiplies traffic into whatever store exists. If the store has no flows, scale multiplies the leak.

Every one is a decision, which is the good news. Decisions can be made once, calmly, in week 0, instead of badly in week 7.

Run the plan, or have it run for you

Nothing in this plan is clever, and that is deliberate. Paper math, funded verdicts, one consolidation, gated scale, a forced decision at day 90: boring on purpose, because boring survives contact with a real P&L. We run this sequence on every launch we touch, our own brands included, and the failed launches we get called into are almost never short of enthusiasm. They are short of a document written before the first euro moved.

If you would rather have the whole loop operated for you, this plan is the first quarter of our full-funnel growth retainer: media, creative, retention, and the weekly P&L review run as one system, with one team accountable for the number. Either way, write the week 0 document tonight. The brands still standing at day 91 are the ones that decided in advance what would make them stop.