You sell digital products — courses, templates, memberships, ebooks, software. No inventory. No shipping. Margins above 80%. On paper, the best business in online commerce. But the month-end story is always the same.
30 min. We read your numbers, not a deck. No obligation.
The ad account shows a healthy ROAS — the revenue Meta took credit for, divided by what you spent. The payment processor shows money coming in. Yet your bank balance is smaller than both numbers promise. The gap is almost always in the data the algorithm learns from. It is optimizing toward the wrong buyers, because it never saw the full picture of who paid and who refunded.
Your checkout platform rarely sends Meta a clean purchase event. ThriveCart, Kajabi, Teachable, Gumroad, and Hotmart all handle the sale on their own servers. The pixel fires on a thank-you page redirect — which breaks the moment a buyer closes the tab too early. Or it fires on the checkout platform’s domain, not yours, which Meta counts differently. Across the digital product accounts we audit, a quarter to two-fifths of confirmed sales never reach the pixel at all.
Your upsells disappear. A customer buys your core course, then takes a one-click upsell for the advanced module. The first purchase event fires. The upsell fires seconds later, often on the same page — and the platform strips it as a duplicate. The revenue that makes your numbers work is the revenue the algorithm cannot see.
Refunds are never sent back. Online courses refund at 5–15%. That is normal. But those refunds never reach Meta. So the campaign keeps chasing more people who look exactly like the ones who ask for their money back inside 30 days. The honest number is contribution profit — what’s left after returns, payment fees, and ad spend come out. ROAS before refunds is the flattering one.
A cold buyer trusts a digital product slower than a physical one. They cannot touch it, hold it, or taste it. So the creative has to carry more weight — proving authority, showing results, removing the risk. Most digital sellers run three to five ads a month, then wonder why the same faces stop converting. The algorithm needs far more genuine variety than that to keep learning.
Take Chocolissimo, a multi-market e-commerce brand. We rebuilt how Meta saw their sales and cut the cost to win a customer by 22% — same ads, same budget, just cleaner data feeding the algorithm. For a digital product the move is the same one: close the gap between the event the algorithm learns from and the real payment your processor confirms.
The team has also shipped tracking for Volkswagen, Audi, KFC, and WizzAir — places where measurement, not creative, was the thing holding growth back. That is the discipline a digital product needs most, because almost everything that breaks for you breaks in the data.
On the creative side, we produce 100+ distinct ads a month across 10–15 genuinely different angle families. For digital products that means authority proof, student results, objection handling, plain explainers of how it works, and risk-reversal formats — the angles that earn trust for something the buyer can never hold in their hands.
Step 1 — The free diagnosis. Step 1 is free. We put your Meta purchase events next to your confirmed payment processor data — ThriveCart, Kajabi, Stripe, or whatever you use to take money. Then we count how many real sales the algorithm never saw. We turn that into a monthly figure of misdirected spend. And we show you which campaigns look profitable on the dashboard but lose money once the refunds land.
Step 2 — The Profit Foundation Audit (€5,000, three weeks). We match your ad data against your payment processor across the last 12 months. For each product we calculate contribution profit — what’s left after you subtract refunds, payment fees, and ad spend from the revenue. That is the number your bank account actually keeps. We also map which angles bring in buyers who finish the course, and which bring in buyers who ask for a refund.
Step 3 — The Profit Proof System (ongoing). We rebuild the connection so every confirmed sale travels from your payment processor straight to Meta’s servers — no more relying on a thank-you page pixel that breaks. We run your campaigns against contribution profit, not ROAS. We produce at least 50 fresh ads a week built to move a cold buyer to a warm one.
Fresh ads are what scaling needs. So we ship at least 50 a week — and the production fee simply covers what they cost to make. Gemini will render you an image for cents. So will we. You are not paying for pixels. You are paying to know which fifty to make.
The second part is how we earn. We take a share of the new profit we create. Profit here means the cash left after VAT, returns, product costs and ad spend — we call it contribution profit. If that number does not grow, we earn nothing on that side.
30 minutes. We read your numbers, not a deck. No obligation.
Yes. We build the event pipeline from each payment backend to Meta independently and deduplicate at the Meta end. Each product gets a clean purchase event with the correct value and the correct product identifier.
Usually both. The tracking problem is that refunded purchases are counted as successful conversions and the algorithm scales toward buyers who look like refunders. The creative problem is that ads are generating low-intent clicks. We fix the signal first, because you cannot diagnose the creative problem accurately until the data is clean.
The profit share model works for any revenue pattern. In between launches, we build the creative library and maintain the measurement infrastructure so the next launch starts with a warm audience and a trained algorithm.
Roughly €5M a year or more, spending at least €12,500 a month on Meta. Below those thresholds the profit share rarely generates enough for either side to justify the engagement. We will tell you honestly on the free call.
On the call we put your ad platform purchase events next to your confirmed payment processor data. Then we show you the gap in euros per month — and what your campaigns really earn once the refunds come out. You keep the
30 minutes. We read your numbers, not a deck. No obligation.