E-commerce marketing that doesn't rely on organic luck

A store launches. The advice is to post consistently, engage with the community, and let the algorithm find its audience. Sometimes it works. When it works it is called a strategy, and when it does not it is called bad luck, which should tell you something about the category.

The problem with organic-only is not that it is ineffective. It is that it has no control surface. If it is not working, there is no lever. You post more. That is the whole intervention.

Paid acquisition has a control surface. It also has a way of losing money quickly, and the difference between those two outcomes is almost entirely decided before any money is spent.

Three numbers, and the only equation that matters

Nearly every failed campaign is a violation of one line:

Customer acquisition cost must be less than gross margin per customer, within a period you can survive.

Three quantities. Almost nobody knows all three before they start.

Gross margin, not revenue

A £60 order is not £60. Subtract cost of goods, payment fees, shipping and packaging, and expected returns. A £60 order at a healthy retail margin is perhaps £25 of gross profit.

That £25 is your entire budget for acquiring the customer, not £60. This single confusion is responsible for more failed ad accounts than any targeting mistake.

Repeat rate — the number that decides everything

How many customers buy again, and how many times?

This is the difference between a business that can outbid competitors and one that cannot. If a customer buys once, you can spend £25. If the average customer buys three times over two years, you can spend £75 — and now you can win auctions your competitors cannot afford.

The catch is the "within a period you can survive" clause. Spending £75 to earn £25 today and £50 across two years is correct on a spreadsheet and fatal if you do not have the cash to bridge it. Many businesses that "scaled profitably" ran out of money doing exactly this.

If your product is genuinely one-purchase, that is not a failure — it just means acquisition economics are tight and referral, bundling or a related second product matters more than ad optimisation ever will.

Cost per acquisition, measured honestly

Total spend divided by new customers. Not cost per click, not cost per add-to-cart, and not the platform's attributed figure taken at face value.

Every ad platform reports the conversions it can claim. Run Meta and Google simultaneously and their combined reported conversions will exceed your actual orders, sometimes substantially. Both are telling the truth from where they sit.

The only number that is not arguable is: total spend across all channels, divided by total new customers, over the same period. Blended CAC. Less flattering, considerably more useful.

What to fix before spending anything

Paid traffic is a multiplier on your conversion rate. Multiplying a broken funnel gets you an expensive lesson delivered faster.

Know what a conversion is, in your own system

Not the pixel's opinion. Your database should be able to answer: how many new customers, from which source, at what margin, in the last 30 days. If it cannot, you are optimising against a number you do not control and cannot audit.

This is the argument for tracking being designed into the product rather than bolted on afterwards. Retrofitted analytics answer the questions the tag manager can reach. Built-in analytics answer the question you actually have.

Fix the landing experience

Sending paid traffic to a homepage is the most common and most expensive mistake in the category. Someone who clicked an ad for a specific product should land on that product, with the ad's promise repeated in the first line.

Speed is a conversion factor, not a technical vanity metric — and mobile is where ad traffic lives, often on a connection nothing like the one you tested on. A page that takes six seconds on a congested mobile connection has already lost a meaningful share of the traffic you paid for, before anyone has read a word.

Have something to say beyond the discount

If the only reason to buy is 20% off, you have trained the customer to wait for 20% off. Discounting is a pricing decision disguised as a marketing one, and it is very hard to walk back.

Sequencing, roughly

An order that avoids the common failure of scaling something that was never working:

1. Measurement first. Know your margin, your repeat rate, and how to count a new customer. Weeks, not months.
2. One channel, small budget, learning only. Whichever your customers actually use. The goal is a real CAC number, not profit.
3. Fix the funnel with what you learn. Traffic that does not convert is usually a landing problem, not a targeting problem.
4. Scale the thing that works. Only after CAC is comfortably under gross margin, and only in increments — ad platforms respond badly to sudden budget changes.
5. Add the second channel. Now blended CAC starts mattering more than either platform's reporting.
6. Then retention. Email, post-purchase, the second product. This is where the economics get good, and it is the part everyone skips to.

Why we do this in-house

The usual arrangement is one company that built the store and another that markets it. Each has half the picture, and each has a reason to believe the other half is the problem.

The specific failure is not personal, it is structural: when conversions drop, the agency sees a traffic quality question and the developer sees a site that is working fine. Nobody owns the middle, and the middle is where the money is.

Building and marketing in one team removes the seam. Tracking is designed with the application instead of retrofitted through a tag manager. Landing pages are built by the people who built the product. And when a campaign underperforms there is no argument about whose fault it is, because there is only one team to ask.

That is not a claim that the campaigns will work. It is a claim about where the accountability sits, which is the part you can verify before you sign anything.