"Trust in iGaming Is a Matter of Data": Interview with aff.studio CEO

aff.studio builds tools for iGaming affiliates — for those who work with traffic. An affiliate business is a business like any other: margins, receivables, partner-side risk, asset value. But almost all the software in this industry is written for operators. We're building what we were missing ourselves.

Right now that's three products. CasinoMass — a database of casinos, slots, providers, and payment methods: SEO and content teams can publish brand and slot reviews without spending weeks on manual research. AffStata — a tracker that consolidates all affiliate dashboards into a single interface, where commissions, trends, and performance dynamics for every program are visible at a glance. And a catalog of affiliate programs with ratings based on an open methodology — so you can choose a new partner based on data and reviews, not rumors in chats.

The brand is young — we launched in October 2025. The products are not: CasinoMass started as an internal project in 2019, AffStata in 2020, both built inside the active affiliate business of our founder Roman Doroshenko, who has 10+ years in affiliate marketing. He assembled a team to bring the products to market and stepped back from day-to-day operations into an investor role.

Roman asked everyone he knew — no one had a proper solution, yet almost everyone had the same problem. Every affiliate has hundreds of affiliate programs — meaning hundreds of dashboards: log into each one, export a report, consolidate it in your own spreadsheet. That doesn't create a standard; it creates thousands of spreadsheets.

Casino data was entirely dependent on manual work. To gather licenses, withdrawal limits, payment methods, and a provider list, you had to crawl the casino's own site and the game developers' sites — often through a VPN, since many of them simply don't load from certain countries. One casino takes a researcher at least half an hour, and a few months later the data is already outdated and you start from scratch.

Tools in this category did exist — trackers, for instance — so I won't say there was nothing. What was missing was the key thing: treating affiliate data as infrastructure, not as a routine to wrap up by the 10th of the month and forget about until next time.

We didn't start with a product — we started with the market: discovery calls with CEOs of affiliate companies to validate whether the pain we were solving internally was real for others, and real enough that they'd pay for a solution. AffStata looked ready for external users, so we launched the tracker first — in October 2025. CasinoMass needed more work; we opened it to the broader market in June this year.

The first few months gave us two insights. The first — the product holds up in direct comparison with the market: our first two paying clients were actively choosing a tracker and, between calls with me, were going to demos with other products. We won that comparison — not for a free pilot, but for a real monthly subscription. Both are still with us.

The second — the market is far more closed off than I expected. People come to demos with cameras off, don't show their sites, and almost every conversation hits the same wall: access to affiliate dashboards and financial data. "I like the product, but I'm not ready to hand my commissions over to someone else's software" — you hear that directly on the call. So security for us isn't a technical question — it's a question of trust, and trust is what determines the order in which we build everything else.

Partly, but the sharper problem isn't low rates — it's that you can't see it. Most affiliates evaluate performance by the front end: clicks, registrations, first deposits. Those numbers appear fastest and get paid out earliest — which is exactly why they have the biggest influence on decisions.

But FTD only confirms one thing: the brand is converting your traffic. Whether it retains the player beyond that, and whether the amount credited in the dashboard actually reaches you — none of that is visible from FTD. Two programs with the same number of first deposits can deliver completely different revenue six months down the line.

There's another layer — geography. The same brand in different countries is effectively a different deal: in some markets it both converts and retains players, in others it simply collects registrations. In a consolidated program report, all of that gets averaged into a single number — and that's the number decisions about where to send traffic next are based on.

So "put traffic where it actually pays" means measuring results all the way through, in numbers you can put side by side. Not the stated rate, but the effective one — the one that accounts for deductions, real player LTV, and what the terms actually say.

The reason is structural, not personal. In these relationships, only one party keeps the books — the operator. They measure the traffic, define who counts as an active player, calculate NGR, and decide what goes into deductions. The affiliate sees their own results only in the form the operator chooses to present them. Without your own numbers to compare against, you're not verifying the report — you're just reading it.

The second reason — the terms of a partner agreement can change after you've already joined the program. We pulled archived versions of terms across 86 programs: 1,009 documented changes since 2015. Across the market as a whole, terms changed against the affiliate's interests 6.7 times more often than in their favor. 23 changes were applied retroactively — backdated to traffic the affiliate had already delivered.

So trust here isn't a question of the character of the people you work with — it's a question of whether you can verify what you're being shown.

The fear is rational, because information is asymmetric. Before you start, you see the program's marketing numbers — but how it actually behaves on payouts three months in is something only those who've already worked with it know.

The best thing that works today is people. Keeping those you trust close and staying in a live network: we have one within NetSolid. Members share lists of reliable partners, manager contacts, help connect you with the right person — that's pre-validation of a partner before you send them your first traffic.

But this mechanism doesn't give a 100% answer: you can get excellent results from an operator that gets bad reviews in chats — simply because you worked with a different manager. Often it's not about the brand, it's about a specific person — which is why affiliates stay in touch with a manager even after they've moved to another company.

We approached this problem through a directory: we provide data on an operator before you start working with them, we collect and verify real affiliate reviews, and in our scoring, that feedback carries no less weight than the numbers a program reports about itself. Fear doesn't go away because of promises — it goes away when information stops being one-sided.

CasinoMass is for SEO and content teams: a database of casinos, slots, providers, and payment methods with filters by license, GEO, and payment method, with export functionality. Instead of manually verifying every brand, the team takes already-verified data and goes straight to writing content.

AffStata is for affiliate teams working with dozens or hundreds of programs. Data from all partner dashboards is normalized — one currency, consistent metric definitions — and consolidated into a single dashboard with breakdowns by brand, GEO, and cohorts. The point isn't to replace dozens of tabs with one report, but to compare programs against each other on equal numbers.

Affiliate Programs Directory is for the moment of choosing a partner: programs are rated using a transparent methodology, with filters by vertical, GEO, license, and deal terms. The second layer is reputational: affiliates leave reviews, we verify them before publishing, and a program can claim its page and respond publicly. In the first month, 12 programs claimed their pages.

The affiliate manager database answers the question "who do I talk to": a classic situation — your manager left a brand that still owes you a payout and stopped responding. You open the database and find another person at the same company, Telegram handle included. The database is crowdsourced: a user can flag a contact as outdated — which affects the contact's reliability rating — and add a new one.

Together, this is one working cycle: find a program in the directory, find the person to negotiate with, pull data for content, see how much traffic actually delivered. Each product makes sense on its own, but together they form a single workflow.

Right now it's one hundred percent manual work. A researcher opens the primary source, collects the data, and enters it into the database following a defined process and strict guidelines. We update a brand as frequently as the client requires: if a client needs a fresh snapshot of a specific casino, the request is turned into a task for the research team.

We often talk to teams that tried to build this kind of database in-house, and almost all of them hit the same two walls — cost and data quality. Removing the human from this process entirely is simply not possible today: the moment you try to collect this data automatically, you start losing the very thing people are actually paying for.

In parallel, we're building a semi-automated solution: part of the process is already automated, and we're now training the team to work with the new workflow. The goal isn't to replace the researcher — it's to have them spend their time on verification rather than routine tasks.

The last piece is control. Every change goes into a daily report, so the client can see exactly what changed and when. For a data product, being verifiable matters more than being large.

All the time. But it's worth separating two things here: most discrepancies aren't abuse — they're differences in terminology and calculation methodology. "Active player" and even "registration" mean different things depending on whose dashboard you're looking at. Put two such reports side by side and you get a number that looks precise but isn't. Deliberate shaving gets discussed far more often than it gets proven — precisely because there's usually nothing to prove it with.

Tracking gives you something different — grounds to ask a question. Campaign reports in dashboards update several times a day, while a partner's financial report is generated once a month. There are situations where your daily data showed, say, one hundred FTDs, but the final financial report has thirty percent fewer. Sometimes there's a perfectly normal explanation — duplicates, fraud filters, players who didn't meet the conditions for attribution. But you can only ask where that thirty percent went if you have your own record.

Does it close the gap completely? No. A tracker doesn't audit the operator and doesn't make your number the truth in a dispute. But your own record, preserved over time, turns a gut feeling into an argument — and a decline that's been dragging on for two months becomes visible at a point when you can still do something about it.

No one has a reliable figure here, so I won't quote one. The issue isn't that promises go unfulfilled — it's that the promise and the result describe two different things. The numbers an operator gives at the start of negotiations are an average across the brand's entire traffic: all GEOs, all channels, all affiliates. An individual affiliate's traffic is just one slice of that average — and almost always an atypical one.

That's why an affiliate can't know their retention in advance — it has to be measured after launch. A cohort shows how players acquired in a given month behave: whether they stayed after two, three, or six months, how many deposits they made each subsequent month — in volume and in value, broken down by brand.

In my experience, this is one of the least-tracked metrics in the industry — and one of the most costly to ignore. An affiliate sending traffic to a brand that doesn't retain players will find out about it roughly a year too late. We've seen brands that hold a cohort even after a year, and brands where almost nothing is left of the first-month cohort.

This is exactly what changes traffic allocation decisions on RevShare: the affiliate isn't comparing promises made at the start — they're comparing how two brands actually performed with their own players.

There's been no pressure. There have been requests: a few programs asked us to update information in their listing. We did update it — but not because of the request itself, rather because in each case the program provided proof that supported their version of events. That's a data quality control process, and it works the same way for everyone, regardless of who reaches out or what tone they use.

It works because there's nothing in a listing that can't be verified: license status comes from the registry, terms come from the program's own official resources, complaint history comes from public sources. So any dispute comes down to evidence, not opinions.

We're prepared for pressure, and our position is defined in advance — otherwise it falls apart the moment the first conversation gets difficult. We won't remove an accurate fact just because it's inconvenient for a program. Our evaluation methodology is published, so anyone can see which factors moved a program up or down and check our work. A directory that quietly rewrites itself to suit the companies it covers isn't worth reading — and affiliates pick up on that fast.

In most niche rankings, rating and visibility are one and the same: pay and move up. We separated the two: visibility can be bought, and it's labeled as Sponsored; rating cannot.

Organic ranking works like this: by default, programs are sorted by number of reviews, with a separate sorting option by House Grade — our proprietary metric, conceptually similar to Clutch Rank. It's calculated based on a published rubric, so anyone can see exactly what makes up a given program's score. Commercial arrangements have no effect on House Grade or a program's position in organic sorting. Sponsored is a separate format: a way to appear at the top quickly, which adds neither reviews nor points to a program's House Grade.

A separate note on monetization: we're starting to add sub-affiliate links to some programs — in certain cases we earn when a visitor signs up through our link. That's precisely why the terms of those programs have no bearing on listing placement or rating.

For us, this isn't a question of ethics — it's a long-game calculation. A catalog that sells ratings makes money fast, right up until the moment affiliates stop trusting it. We want to be the place affiliates go first when choosing a new partner, and that only holds up on the strength of reputation.

I'd highlight three. The first — part of the teams in iGaming are starting to play the long game in SEO: many have grown tired of constant traffic drops, and the long game is about precise, up-to-date, and regularly updated content. For us, this means data quality is no longer a content planning question — it becomes a daily operational task, and that's where CasinoMass comes in.

The second — the market is maturing and the cost of player acquisition is rising. As a result, a mistake in choosing a partner costs more than it did two years ago. Teams that are scaling are shifting focus from front-end metrics to retention, LTV, and net margin, and program selection is evolving from a quick decision into a structured process. A catalog with transparent ratings and verified reviews is a direct answer to this trend.

The third — regulation: it constantly reshapes which GEOs are worth working with and on what terms. Because of this, license status and the history of terms changes are moving from the legal side into the commercial side — they're now part of the decision about where to send traffic. Together, these trends raise the cost of decisions made blindly. We're building something that lowers that cost.

Most of the work right now is going into AffStata. The dashboard and cohort analysis already answer the question of which program to scale and which one to exit. The next step — making sure that data isn't locked inside our interface: we're developing an API so clients can pull their numbers into their own BI system — Tableau, Power BI, or a custom solution. Alongside that, we're building notifications: when a metric hits its target value or the numbers behave differently than usual, the client finds out on their own.

The second priority is security: we're moving toward full isolation between clients and working on individual domains for each one. The logic is simple: an issue on one client's side shouldn't create risk for the rest.

The catalog currently has five verticals open; we're planning to add up to ten more. In parallel, we're deepening the data for each program and surfacing the history of terms changes on the program page rather than burying it in an archive. In the affiliate manager database, we're adding GEO to each contact — so you don't reach out to someone who doesn't work with your GEO.

The most notable change we can back with data — programs have become more aggressive about rewriting their terms. And there's a counterintuitive detail here: among strictly licensed operators (Malta, UK, Gibraltar), for every one change made in the affiliate's favor, there were 7.8 changes made against them. For offshore programs — 4.5. In other words, stricter regulation of the player relationship does not mean a softer contract for the affiliate — if anything, it's the opposite.

Second — margins are being squeezed from both sides: traffic is getting more expensive, and organic is delivering less than it was a year or two ago, so revenue is shifting toward retention. For day-to-day operations, this means one thing: the risk has moved from "will there be traffic" to "what will the terms say at the time of payout." Reading terms before launch has become as much a part of the job as checking the rate.

Fewer companies will remain in the market, but they'll be larger and operationally stronger. Smaller teams will gradually be pushed out by content costs, search changes, and compliance requirements. Part of this process is already underway.

Second — what operators pay for is changing. Operators are getting better at measuring player value, so rewards are increasingly tied not to the deposit itself, but to what happens after it: retention and LTV. Those who can measure the same thing win; those who can't, lose.

Third, and this is where we're placing our bet: affiliate companies will start being bought and sold like proper businesses. And that means due diligence: buyers won't be looking at traffic volume — they'll be looking at revenue predictability, partner concentration, and how well contracts protect against terms changes. Software in this market either serves exactly that kind of business, or becomes irrelevant.