Why dropshipping businesses fail — and it is almost never the product

Almost every post-mortem starts with the product: wrong niche, wrong creative, saturated market. In our experience the product is rarely the cause. The cause is that the unit economics were never going to work at the cost of acquisition the store was buying traffic at, and nobody did the subtraction before spending.

Here is a store that looks healthy in Shopify and is insolvent in reality. Average order value $45, product cost $14, shipping to the customer $6.50, card fees at 2.9% + $0.30, a 4% allowance for refunds and chargebacks, and a $28 cost per purchase.

Every order loses $6.91. At 600 orders a month that is $4,143 of revenue-funded loss — and the dashboard says revenue is growing, which is why the store keeps going for another quarter.

Two numbers would have prevented it. At a $45 average order the most this store can pay for a purchase is $21.09. Or, keeping the $28 cost, it needs an average order value of $52.42 — 16.5% higher — to break even. Neither is a marketing problem. Both are pricing and bundling decisions made before the first ad ran.

This is why we ask for average order value, product cost, shipping and card fees before quoting any work. A store with negative contribution does not need better creative. It needs a different offer.

You cannot fix a margin you cannot measure, and most stores are measuring far less than they believe. We opened 47 storefronts that were actively running Meta ads and recorded what their pages actually did on a cold load.

24 of the 47 — 51% — fired no Meta pixel at all within three seconds. 68% had GA4, 26% loaded a Google Tag Manager container, and 6% showed any detectable server-side path.

Some of those pixels are consent-gated and would fire after a click. But a store that cannot tell you which campaign produced which order is optimising on Meta's modelled guess, and Meta will happily spend the budget learning from a partly fictional picture.

The failure mode is quiet. Nothing errors. Orders arrive. The owner concludes the creative is tired and orders more, when the real problem is that purchases are being attributed to the wrong ad set, or not at all.

Advertising is paid for in near real time. Card settlements arrive after a delay measured in days, and platforms hold a reserve on new accounts. Between those two clocks sits every scaling decision a store makes.

The pattern is consistent: a store finds an ad that works at $200 a day, doubles it, and then cannot fund the second week because the first week's revenue has not landed yet. Growth is what breaks the bank account, not failure.

A product page that promises five-day delivery on a shipment that takes three weeks produces three problems at once: refund requests, chargebacks, and a payment processor that starts asking questions. The delivery promise is part of the offer, and it is usually written by someone who has never seen a tracking number.

It is also a policy exposure. Merchant Center and ad platforms both treat shipping and returns accuracy as a trust signal. Contradicting your own checkout — or publishing a delivery window your carrier cannot meet — is the kind of mismatch that gets accounts reviewed, and reviews are a limited resource.

Ad accounts and Merchant Center accounts do not degrade; they switch off. A store can be trading normally at 9am and have no traffic by noon, and the appeal process gives you very few attempts before the decision is treated as final.

The stores that survive this treat compliance as an operating requirement rather than an afterthought: a real business address, policies that match the checkout, product data that matches the landing page, and one honest delivery window. None of that is glamorous, and all of it is cheaper than an eight-month suspension.

The last failure is strategic. If the only difference between two stores is the logo on a product both can buy from the same supplier, then the customer's decision reduces to price and delivery — the two things a new store is worst at.

The stores that last build something the supplier cannot hand to a competitor: a bundle that solves a whole job, a guarantee nobody else offers, a subscription, a niche they understand better than the person copying them. That is not a branding exercise. It is the thing that lets you pay $21 for a purchase and still profit.

None of this is novel. It is simply the work that does not feel like marketing, which is why it is the work most stores skip until the day the account goes dark.