There is a version of this we could have live within weeks, funded out of pocket rather than out of a raise, in a market that pays out every single month and never goes out of season. This is that version — and the reason it works is that we are not starting from a blank page.
Becoming a bet365 partner is a marketing business with almost no capital requirement — and it is the half of this industry that quietly prints money. Running an online casino is a licensing and payments business that happens to show games.
Both end with someone placing a bet, so people slide between them. The difference is who holds the licence, who holds the player's money, and who carries the loss when a player wins. As a partner the answer to all three is "not us" — no licence to win, no payment processor to convince, no balance sheet exposed to a lucky punter — and we still get paid every month those players keep playing. That is not the small version of the business. It is the version with the best risk-adjusted return, and it is the one we can start now.
The opportunity
Four things make this different from most ideas that arrive on a Tuesday.
The listed-company and acquisition points above are widely reported rather than personally verified — worth a ten-minute check before either of us repeats them to anyone with money.
The fork
| Affiliate partner | Operator | |
|---|---|---|
| What you own | An audience, and the traffic that reaches it | The licence, the player accounts and the balances |
| Gambling licence | Not needed to be an affiliate in most markets — but you are bound by the operator's terms and by advertising rules, and a few jurisdictions require affiliates to register | Required, per market, before a single real bet |
| Capital to start | Low — content and media | £150k–£1.5m+ |
| Time to first revenue | Weeks | Months to over a year |
| Who carries player risk | Nobody on your side — a winning player costs you nothing directly | You. A big win is paid out of your balance sheet |
| How you get paid | A share of the revenue your players generate, or a fee per player | The margin on everything they lose, minus every cost of running it |
| The main risk | The programme changes terms or closes your account, and traffic sources move under you | Losing payment processing, or failing the licence |
| What it's worth later | Whatever the audience is worth — real, if it isn't rented from an algorithm | A licensed operator, which is a substantial asset |
The practical point: the partner route can be started now, alongside anything else, and it teaches the single most important thing an operator needs to know — what a player costs to acquire and what they return. If those numbers don't work as an affiliate, they will not magically work as an operator, where every other cost is layered on top.
Route A · The deal
Three structures, and the choice between them matters more than the headline percentage.
| Model | How it works | Suits |
|---|---|---|
| Revenue share | A percentage of the net revenue your referred players generate, usually for as long as they keep playing. Industry norms sit broadly in the twenties to mid-thirties, tiered by how many new depositing players you send each month. | Durable traffic — search, a real community, an owned audience |
| CPA | A fixed fee per qualifying new player, paid once. Predictable, capped, no tail. | Paid media, where the cost per acquisition needs to be known in advance |
| Hybrid | A smaller fee up front plus a smaller ongoing share. Usually negotiated rather than offered. | Cash-flow while an audience is still compounding |
Percentages here are industry-typical ranges for sizing the opportunity. The only number that matters is the one in the agreement he is actually offered — read it before modelling anything on it.
Route A · The contract
Affiliate agreements are written by the operator and are rarely negotiated by newcomers, but they should at least be understood. These four do most of the damage.
| Clause | What it does | What to look for |
|---|---|---|
| Negative carryover | If your players win more than they lose in a month, your balance goes negative — and under some agreements that deficit carries into the following months before you earn again. | Whether it resets monthly. With a small player base, one big winner can wipe out a quarter. |
| The tail | Whether you keep earning from a player after you stop sending new ones, and for how long. | Accounts that go inactive for a set period sometimes stop paying out entirely. That clause turns a long-term asset into a treadmill. |
| Permitted territories | You may only send traffic from countries where that operator is licensed. | Traffic from the wrong country can void commissions — including commissions already earned. |
| Traffic restrictions | Bidding on the brand name in paid search is almost universally banned, as is incentivised or misrepresented traffic. | Breaches are the standard reason accounts get closed and balances confiscated. |
Route A · Compliance
An affiliate doesn't hold the gambling licence, but the operator does — and the operator is answerable to its regulator for how its partners promote it. That makes enforcement fast and unsentimental. In practice:
None of this is onerous for someone building a genuine content or product business. It is fatal for anyone planning to arbitrage cheap social traffic.
Route A · The work
This is the whole job, and it is the part that gets waved through in conversation. Four sources that work, in rough order of durability:
The last one is where a technical partner is worth having, and it is the one that survives an algorithm change.
Route A · The build
A partner account and a page of links is not a business — anyone can have one within a day, which is precisely why it is worth nothing. The business is the machine that produces the traffic, and that is the half I build. None of it needs a licence, and all of it is owned rather than rented.
The foundations, first, because nothing else matters without them: the audience product that earns the traffic in the first place; tracking and attribution, so every link is measured and the partner's reporting can be checked rather than trusted; content operations at the volume search results are actually won at; an owned email and messaging list; and a dashboard showing what a player costs against what they return, by source. Those last two numbers are the only ones that decide whether any of this is a business.
On top of that sit the layers that make it defensible.
Route A · The layers
| Layer | Why it earns its place | Reality check |
|---|---|---|
| Telegram bot | Where a betting audience actually lives. Push notifications without owning an app, near-zero running cost, and it converts far better than a website because it reaches people at the moment a match starts. | A bot is advertising like anything else — the same age-gating and content rules apply, and the same partner terms govern what may be said. |
| Odds comparison | The highest-intent traffic in the sector. Someone comparing prices is minutes away from placing a bet, and an entire category of large businesses was built on exactly this. | Needs a licensed odds feed, which is a genuine recurring cost. Scraping bookmakers breaks their terms and their pages change constantly — it is a treadmill, not a moat. |
| Free-to-play prediction game | Builds an audience nobody can switch off, and produces first-party data on who actually bets. It gives people a reason to show up on days they are not betting, which is what makes the traffic durable. | With no stake and no purchase required to enter, this sits outside gambling licensing — but prize competitions have their own rules, so the terms need writing properly. |
| Token or on-chain layer | Works as a loyalty and community mechanic on top of a free-to-play game — points that are actually ownable, and a reason to stay. | The line is whether money is at risk on an outcome. A loyalty token on a free game is a marketing feature. A market where people stake and are paid on results is an operator, and needs the licence the second half of this brief is about. Calling it web3 does not move that line. |
Route A · Terms
Worth settling before anything is written, because it is awkward to raise later. Building this properly is not a job of work handed over at the end — the layers above are the asset, and they take months of continuous development to be worth anything.
Route B · Reality check
The licence and the payment rails. Almost everything else, including the games, is rented — so the engineering is the cheapest line in the budget, not the expensive one.
| Layer | Where it comes from | Buy or build |
|---|---|---|
| Slots & table games | Licensed studios — Pragmatic Play, Play'n GO, Hacksaw, Nolimit City, Relax | Rent |
| Live dealer | Evolution and a few others — physical studios, dealers, camera rigs | Rent |
| Game aggregation | SoftSwiss, Slotegrator, EveryMatrix, BetConstruct, Digitain — one integration, thousands of titles | Rent |
| Wallet, bet ledger, RNG | Platform provider, independently certified by GLI, iTech Labs, eCOGRA or BMM | Rent |
| Brand, site, app, UX | You | Build |
| Retention: CRM, segmentation, bonus logic | The platform gives you a basic version; good operators outgrow it fast | Build |
| Affiliate programme of your own | Income Access, MyAffiliates, or built in-house | Build |
| KYC and AML screening | Sumsub, Veriff, Jumio, ComplyAdvantage — integrate, never reinvent | Rent |
| Safer gambling tooling | Mandated by the licence — limits, time-outs, self-exclusion, national registers | Build |
The certified layers are certified because regulators require independent testing of the random number generator and the bet ledger before granting a licence. Rebuilding them is a year-long project with no commercial upside.
Route B · Structure
| White label | Turnkey, own licence | Self-build + aggregator | |
|---|---|---|---|
| Time to live | 6–12 weeks | 4–9 months | 9–18 months |
| Licence holder | The platform operator — you are a brand on their permission | You | You |
| Who owns the players | Them, in most contracts | You | You |
| Economics | Setup fee plus a large share of net revenue, often 30–50% | Platform fee, commonly 10–20% of net revenue, plus content splits | Content splits only, but every other cost is yours |
| Worth later | Close to nothing — neither licence nor players are yours | A sellable operator | An operator and a platform |
A white label is a rental. Good for testing whether players can be acquired profitably, poor for building something with resale value. If he goes that way, the contract needs a migration clause, and the retention and affiliate layer should be his from day one.
Route B · Licensing
This is where first-time operators lose the first hundred thousand: buy the cheap licence, build the product, then find the players they wanted are in a market it doesn't cover.
| Jurisdiction | Effort | Opens | The catch |
|---|---|---|---|
| Curaçao direct from the Gaming Authority since the 2024 reform | Months | Unregulated and grey markets; the usual starting point | Payment and banking acceptance is the weak spot, and it does not cover the UK, Netherlands, Germany, Denmark, Sweden, Italy, Spain or the US |
| Malta | 4–6+ months | Credibility, EU-facing operations, better banking | Genuine corporate substance required — people and an office, not a shelf company |
| United Kingdom | 6+ months | The only lawful route to a UK player | Personal management licences, affordability and safer-gambling duties, gaming duty. Not a first licence. |
| Isle of Man, Gibraltar | Months | Strong reputation and banking | High bar on capital, substance and track record |
| Anjouan, Tobique | Weeks | Cheapest, fastest paperwork | Weakest acceptance — increasingly refused by payment providers and some aggregators, which defeats the point |
Order matters: pick the market, then the licence. Never the reverse.
Route B · Payments
Gambling sits in a high-risk merchant category. Mainstream processors decline it outright, so acceptance comes from a small field of specialists who underwrite the operator rather than the idea.
If payment providers won't underwrite him, there is no operator business, however good the product. Which is why it belongs early in the sequence.
Route B · Money
Order-of-magnitude planning figures, to size the conversation rather than to quote from.
| Route | Licence & legal | Platform | Compliance | Payments float | Marketing | First year |
|---|---|---|---|---|---|---|
| White label | £15–40k | £20–60k | Bundled | Low | £100–300k | £150–400k |
| Turnkey, own licence | £40–90k | £60–150k | £30–60k | £50–150k | £150–400k | £350–800k |
| Tier-one jurisdiction | £150–300k | £100–250k | £80–150k | £100–250k | £300k+ | £800k–1.5m+ |
The largest line is always acquisition, and it stays that way. Cost per depositing player runs into the hundreds in competitive markets and is recovered over months of play, not days. A plan without twelve months of marketing runway is a hobby with a licence attached.
Route B · Sequence
The first three steps cost legal fees and time, not build time. Nothing built before step three is safe, because any of the three can end the project.
Gate: written legal advice that players there can lawfully be taken.
Gate: a pre-application conversation with the regulator or a licensing consultant.
Gate: at least two providers willing to take him on. If this fails, stop — everything downstream is worthless without rails.
With licence and payment costs known, white label versus own licence becomes an informed choice.
Everyone with influence is assessed for suitability. Surprises here are fatal and cannot be fixed later.
Watch the exit terms and who owns the player data.
Brand, front end, CRM, affiliate tracking, analytics — the part worth owning, and the part I'd build.
Safer gambling tools, AML policy, terms, bonus rules. Bonus terms get tested by professionals within days of launch.
The only test that matters: does a player cost less than they return?
Guardrails
Worth naming, because every one of them is avoidable by design rather than by luck — and knowing them in advance is most of the advantage.
The operator route carries heavier versions of the same question — losing payment processing stops revenue the same afternoon, and a founder failing suitability checks cannot be engineered around — which is exactly why it is the second move rather than the first.
The contact
Genuinely valuable — for supplier introductions, for hiring, and for a sanity check on the acquisition numbers, which are the numbers most likely to be wrong. Worth being explicit about the boundary early, because it becomes a formal question if an operator licence is ever applied for.
His contact's employment contract almost certainly carries confidentiality terms and may restrict competing. Advice, introductions and market judgement are his to give. Documents, player data, pricing models and supplier terms belonging to his employer are not — and that tends to surface during licensing, when regulators assess whether the people behind an application are suitable to hold one. Keeping it to advice protects both of them and costs nothing.
Next
Recommendation
Start as a partner, in one market, with one product that earns its own audience rather than a page of affiliate links. It can begin immediately, it costs time rather than capital, and within a few months it produces the two numbers that decide everything else: what a player costs to acquire, and what they are worth once acquired. Get those right and every door after this one — more operators, more markets, eventually our own licence — opens on evidence rather than on hope.
If those numbers are good, the operator route becomes a financeable proposition backed by evidence, and the audience built as a partner becomes the launch audience. If they are not good, that has been discovered for the price of some months and some content, rather than for the price of a licence, a platform contract and a payments float.
Concretely, the first month looks like this. He confirms the partner terms and the market, and gets the agreement in front of us both. I stand up the tracking, the bot and the first version of the audience product — that part is weeks, not months, because most of it already exists. We agree how we split it and who owns what, in writing, while everyone still likes each other. Then we put the first real traffic through it and read the only two numbers that matter. Inside a quarter we will know whether this is a business worth pouring years into, and it will have cost us a fraction of what most people spend to find that out.
My side of it is settled either way: I build and run the technical half — the bot, the comparison engine, the audience product, the tracking and the reporting — as the technical partner in the venture rather than as a supplier to it. That is the part that turns a partner account into something with a value of its own, and it is the part that carries across if the operator route ever becomes the right move.