The most common leaks in a DTC marketing funnel (and how to find yours)

A funnel rarely fails the way founders expect. There is no single broken page, no one bad ad. The brand is doing seven figures, the dashboard is green, and yet the profit is thinner than the revenue suggests. That gap is the leak, and the pressure underneath it keeps building. SimplicityDX (2022) reports that customer acquisition costs have risen 222% over the last decade, and that brands now lose on average $29 for every new customer acquired on a fully loaded basis, up from $19 ten years earlier.

At The Social Target, we have run this engine for established brands for nine years, across 600+ clients with 50+ active today. The leaks repeat: a skincare brand and a furniture brand lose money in almost identical places. Here are the six we find most often, and how an operator pins yours down.

What is a DTC funnel leak?

A DTC funnel leak is any point where money or customer lifetime drains out of the system faster than the funnel replaces it. It is not a crash but a slow loss that hides inside a profitable-looking account, because the top-line revenue keeps the dashboard green while the leak quietly raises what every sale costs.

Leaks are dangerous because they are invisible at the level most founders watch. Revenue and ROAS can both look fine while the brand quietly loses efficiency. A leak shows up as a trend, not an event, and by the time it is obvious in the bank balance it has usually run for a quarter. The job is to stop watching the green number and start watching where the system over-relies on one part of itself. That over-reliance is the leak.

What are the most common leaks in a DTC funnel?

The six most common leaks are: one creative carrying the account, no retention engine, one saturated audience, an untested offer, measuring ROAS instead of contribution margin, and a founder who is the bottleneck. Most brands run two or three at once, and they compound.

The six leaks below are the ones we see in almost every audit. Because they compound, the funnel feels heavy long before any single one looks like a problem.

Leak one: one creative is carrying the whole account

The most common leak in a profitable DTC funnel is a single winning ad: one creative quietly accounts for the majority of spend and return, and the brand builds its forecast on it. This is the riskiest possible bet to concentrate: according to NCSolutions (2023), in its update to the Five Keys to Advertising Effectiveness study, advertising creative is the single biggest driver of incremental sales, responsible for 49% of them, ahead of brand at 21%, reach at 14%, targeting at 11%, and recency at 5%. Creative is your largest performance lever, and no creative lasts. The day that ad fatigues, your whole account drops with it.

What makes this leak insidious is that most operators underweight the very lever they are over-concentrating. An Advertiser Perceptions survey of 606 marketers and agencies, reported by Cumulus Media | Westwood One (2024), found brands estimate creative drives only about 19% of total sales effect, roughly the same figure they have assumed for five years, while the NCSolutions science puts the real number near 49%.

You find this leak by ranking spend and return by individual creative. If your top one or two ads carry the bulk of both, you have a fragile account, not a strong one. The fix is a creative pipeline, not a single hero: enough new angles in test that you are never one fatigued ad away from a bad month.

Leak two: no retention, so you keep re-buying the same customer

The second leak is the most expensive and the easiest to miss. With no retention engine, every quarter you pay acquisition cost to win back customers you already had: a bucket with a hole in it. The hole is bigger than most founders assume: BS&Co (2026), analysing 156,110 DTC customers across consumables, fashion, and durables, found the average e-commerce repeat purchase rate is just 18.8%, meaning 81.2% of customers never place a second order within a 365-day window.

How expensive a new customer is depends on your vertical: Shopify's customer acquisition cost data puts the average at $377 in electronics, $129 in both home, furniture and garden and fashion and accessories, and $127 in health and beauty. Pay that figure twice for the same buyer and the economics of the funnel quietly invert. A brand that builds email, post-purchase flows, and a reason to come back changes its entire economics, because the second and third purchase carry almost no acquisition cost. Timing matters too: of the customers who do return, BS&Co found 50.3% buy again within 30 days, so the window is short. You find this leak in repeat-purchase rate and lifetime value: if both are flat while spend climbs, the hole is in the bucket, not the tap.

Leak three: one saturated audience

The third leak is reach. Many brands scale by pouring more budget into the same core audience until the cost to reach the next buyer climbs past what a sale is worth. The audience is not bigger; the budget is. You pay more per customer and call it rising ad costs when it is really saturation.

You find this leak in frequency and the cost trend within a single audience over time. When the same people see the ads more often and each new sale costs more, the audience is tapped. The fix is genuinely new audiences and angles that open fresh demand, not more spend aimed at people you have already won.

Leak four: an untested offer

The fourth leak sits underneath all the others. Brands test creative, audiences, and landing pages obsessively while leaving the offer untouched for years, yet the offer changes the maths on every click that follows it: a weak one makes even great traffic convert poorly, and no amount of creative testing fixes that.

You find this leak by asking when you last changed the offer, not the ad. If the answer is "we have not", that is the leak. Bundles, guarantees, the first-purchase incentive, the free-shipping threshold: these move conversion harder than another creative variant ever will.

Leak five: measuring ROAS instead of contribution

The fifth leak is not in the funnel. It is in how you read it. ROAS tells you revenue over ad spend and nothing about margin, cost of goods, shipping, or discounts, so a brand can hit a healthy ROAS and still lose money on the order. This is exactly the trap SimplicityDX (2022) describes: because their $29-per-customer loss assumes the first sale carries the full cost of acquisition, profitability now depends on the second and third sale, which ROAS on the first order cannot see.

Contribution margin, what is left after the variable cost of fulfilling that sale, is the number that does pay the bills. We put it this way to clients: we make your cents work like dollars, and you cannot do that while measuring the wrong cents. Rebuild your reporting around profit per order rather than return on spend, and where every next pound should go often changes.

Leak six: the founder is the bottleneck

The sixth leak is the one founders least want to hear, because it is them. When every creative decision, offer change, and approval routes through one person, the funnel moves only as fast as that person has hours. Growth stalls not because the strategy is wrong but because the system cannot decide without them.

You find this leak by looking at what is waiting on you right now. If tests are queued behind your approval and nothing ships while you are away, the constraint is not the market; it is the org chart. The fix is to move from doing the work to owning the outcome: clear decision rules, a team or partner who can run the pipeline, and a founder freed to work on the brand rather than inside it.

How do you find which leak is draining your funnel?

You find the leak by isolating each part of the funnel and testing which one carries more weight, or loses more value, than it should. Rank creatives by spend and return, check repeat rate and lifetime value, check audience frequency, audit the offer, rebuild ROAS into contribution margin, then look at what is queued behind one person.

You measure before you change anything. The teardown works through the six in order, and the one number worth holding against your own is retention: benchmark your repeat-purchase rate against the 18.8% DTC average BS&Co (2026) reported, because re-buying your own customers is usually the costliest leak hiding in a profitable account.

The fatiguing ad is loud; the missing retention engine is silent and drains far more. Once the leaks are patched and you are trying to scale, see the companion piece on why your ads stop scaling.

Why do these leaks survive in profitable brands?

These leaks survive because profit hides them. When the top line is healthy, nobody hunts for the slow drain underneath, so the leak gets a year or two of runway. And because each leak lives in a different team's remit, no single person owns the whole-funnel view that would expose it.

The brands most exposed are not the struggling ones. They are the comfortable ones.

There is also a structural reason. Each leak lives in a different place, so no single person owns the whole picture. The media buyer watches ROAS, the email team watches retention, the founder watches revenue, and the leak lives in the seams between them. Catching it takes a whole-system view, and most in-house setups are missing it.

This is the difference between running campaigns and running a funnel. The campaigns can all look fine while the funnel they sit inside loses money, because the leak is never in the campaign; it is in how the campaigns fit together.

What does fixing a leak actually look like?

Fixing a leak means patching one hole at a time, in order of how much it costs you, and measuring the profit line rather than the vanity line as you go. It is not a rebuild. Most funnels need only the two or three biggest leaks closed to change the economics of the whole thing.

Most funnels do not need tearing down; they need the two or three biggest leaks closed, which is usually enough to change the whole funnel's economics. We describe our model as run by an artist, operated like an engineer, and the engineer half is what closes leaks: isolate the variable, change one thing, measure against contribution, keep what works. Each fix compounds, because the leaks were compounding too.

The brands we have moved furthest are the ones that let an operator find the leak before spending another pound trying to outrun it. Old Dirty Brasstards reached 15x ROAS not by spending more but by running a tighter system. FoundPop has held a 10.3x blended return across a multi-year retainer because the funnel underneath it does not leak. The growth came from fewer holes, not louder spend.