We turned an affiliate programme into a product line.
Bloggers sent CryptoRobotics traffic for a one-time commission. We designed the layer that lets any of them launch their own branded platform in an evening — and keep the customer.
Research, business model, the partner-facing product.
Taken apart before a screen was drawn.
The market benchmark is 4–12 weeks.
Through partners, at no acquisition cost.
The audience was already there. It just had nothing to sell.
Partners send a client, collect once, and the client belongs to the platform from then on. Everyone else is bought in the same ad auction as the whole category.
- Engines, exchange integrations, billing and demo mode — built, paid for, serving one brand.
- A partner network with real audiences and real trust.
- Non-custodial architecture: users connect their own exchange keys.
- A partner is paid once. After that the client is the platform's.
- Everyone else is bought, keyword by keyword.
- Infrastructure that serves one brand earns from one brand.
A partner with forty thousand followers could send them somewhere. They could not sell them anything.
Fourteen platforms, two questions.
What is automated trading worth per month, and what does "white label" already mean to a buyer? Both answers changed the product.
- Sell
- One brand, $15–150 a month
- Cost
- Free to promote, nothing to own
- Blocker
- You can advertise it, never sign it
- Sell
- An exchange — engine, order book, liquidity, custody
- Cost
- $15–60K and 4–12 weeks
- Blocker
- The operator has to hold customer money
Nobody was selling the algorithmic layer on its own.
to go live, against $15,000 for the cheapest white-label exchange. They hold customer money. We never do — every user brings their own key. ChainUP, B2Broker, AlphaPoint — published 2026 pricing
A product that cannot afford a salesperson.
At $249 the first support conversation eats the margin. So the launch had to be something one person finishes alone, at night. That constraint is the design.
Each stage carries its own status, so a partner fills in what they know and comes back to where they stopped. Step one is also the first pricing decision — $249 on a subdomain, $399 on their own — stated before the first field.
Which venues appear, whether demo trading is offered. Every user brings their own API key, so the partner never holds a coin.
A signal-channel owner ships a platform that is only signals and adds bots when ready. A product on day one beats a catalogue nobody chose.
The business model is the interface.
Two radio buttons decide whether the partner's platform is a funnel or a paywall. Most of them have never priced a subscription before.
Freemium charges for the good parts; Membership charges at the door — the consequence of each written next to the choice. Package defaults land inside the $15–150 band the research found.
Crypto settles every Monday, fiat between the 3rd and the 6th — on the screen where the method is chosen, not in a contract.
recurring revenue per converted affiliate, against the one-time commission they used to collect from the same audience. Averaged across partners active six months or longer
The screens that decide whether a partner stays.
White-label businesses churn in month two, when the operator can no longer explain their own numbers. So the operator side was built as a product, not a settings page.
A discount, the packages it applies to, and a live count of activations against completed purchases — what was given away sits next to what it bought.
What the operator watches
Exchanges by user and by key, subaccounts, tools launched split real against demo, active packages against the discount that bought them. It exports to a spreadsheet — reporting to your own investors needs a file, not a screenshot.
Why that is the retention product
Every surface answers a question a partner will be asked: what sold, who came back, which code paid for itself. Software ships once; a business has to keep making sense on a Tuesday morning.
Everything a partner changes without calling us.
A dozen surfaces, every one self-explanatory — the whole economic case is that nobody is on the other end.
Subdomain or their own, with the fee stated up front.
Logo, palette and naming, so the product reads as theirs.
Bots, signal bots, signals and smart trading, one at a time.
Which venues appear, and whether demo trading is offered.
Several accounts inside the platform, none on the exchange.
Freemium or Membership, and the PRO packages under it.
Crypto or card, each with its settlement schedule.
The partner's own affiliate layer, one level down.
Discounts by package, in every language the platform runs.
Who signed up, who authorised, who came back.
What sold, at what price, on which package.
Keys, subaccounts and activity, venue by venue.
What it changed for the business.
One product with one growth channel became infrastructure with a sales force that pays for itself.
A second revenue line on infrastructure already paid for, and users arriving at no acquisition cost.
An audience they monetised once now pays them every month, on a product with their name on it.
The same engine, reached through someone they already trust.
What we actually did.
Market & competitor research
Fourteen platforms — subscription bots for the price band, white-label vendors for what the words already meant. The pricing and the non-custodial positioning both came out of it.
Business-model design
Freemium against Membership, packages, promo and referral mechanics, settlement schedules — designed as screens, because that is the only place a partner meets the model.
Self-serve onboarding
A six-stage wizard with resumable state, so a platform launches without a call and the unit economics hold at the fortieth customer.
Operator product
The dashboard, statistics and growth tools a partner needs in month two — exactly when white-label businesses churn.