Running an affiliate network is all about constant improvement. To survive in a competitive market, you need to implement the best tools and technical features for traffic analysis and sometimes even save the day. Imagine you're an affiliate network, and you just brought in a new team with a great offer. They performed well, with a total payout of $10 000. You checked the traffic for fraud, and everything looked fine. But the advertiser disagrees and refuses to pay. What do you do? It seems like the best option is to pay the team out of your pocket and stop working with that advertiser.
Situations like these are the reality of the affiliate network business. In this part of the article, Ilya, CCO of Trafee, talks in detail about the inner workings of their network: how they scaled their business and entered a new market, how they calculate profitability, and how they combat fraud. At the end of the article, there's a bonus: the latest trends in the iGaming and Dating verticals.
|
In the previous part, we discussed with Ilya whether it's possible to find a compromise between webmasters and advertisers and the kind of competition affiliate networks face. You can read that part here.
|
On Shaving, Fraud, and Resolving Conflicts with Unreliable Advertisers
Webmasters complain about both the networks themselves and their technical capabilities — skewed statistics, tracking issues, unhelpful support. What problems do you see in the affiliate network market, and how do you handle them? What essential features do affiliate networks need today?
The main problem for an affiliate network is finding advertisers. It's not even about finding webmasters — you can get a lot of them quickly. But finding advertisers with quality offers who will pay and not disappear after a month is a challenge. We've been burned by this more than once.
|
Trust is the biggest issue at every stage of the relationship, whether with advertisers or webmasters.
|
There's also a lack of financial literacy among webmasters. Many rush into it for quick money after reading online about how easy it is to make money. They don't fully understand the statistics: hits, clicks, uniques, EPC, EPM. Some say, "We need a higher rate!" But the rate doesn't tell the whole story. You could set it to $100 per conversion, but if you only get one conversion a month, it won't make a difference.
Some networks, for example, don't show EPM. They have good stats, show a lot of data, but no EPM — which is crucial for the revenue side. When we talk to partners who have complaints, we suggest they look at the EPM.
In the previous part of the interview, you shared your thoughts on shaving. What about fraud?
There's a lot of fraud nowadays. A significant portion of our costs goes into our anti-fraud system and technically dissecting such issues. We try to filter out bad traffic ourselves before passing it to the advertiser. Thanks to our well-thought-out anti-fraud measures, we get high ratings from advertisers, which we can then pass on to our webmasters.
The main types of fraud include registration through proxies and VPNs, and the use of emulators. We evaluate traffic based on 20 criteria. We have a massive block written in AI that acts like a smartlink for analysis, which is then double-checked manually to avoid mistakes.
So you've integrated AI to handle anti-fraud?
It's a pseudo-AI, not something like ChatGPT pulled from outside. It's an in-house development.
The algorithm first collects the data, then makes conclusions based on that data, and assigns importance to those conclusions. But the final decision is still made by a human who checks how many fraudulent traffic patterns were triggered and how critical they are. We trust the robots, but we verify everything.
I heard you paid $10 000 to webmasters out of your pocket because a gambling advertiser refused to pay. Can you share more about what happened?
This happened recently, and we're still dealing with it. The advertiser simply refuses to pay, citing fraud. We understand and can differentiate what's fraud and what isn't. We sent them numerous reports, highlighted the patterns we use to assess fraud, and they’re focusing on identical IP addresses from a single virtual mobile network operator. We dug into what network this is, identified the initial IP addresses they use, and even compiled open statistics on the matter and sent it to them, but they wouldn't budge. It’s a clear case of an advertiser trying to dodge payment, so we decided to pay our webmasters out of our own pocket to maintain good relationships.
You've developed a smartlink for dating. Can you share more about how it works?
Sure! Our smartlink integrates 4,000 offers. An AI-based algorithm distributes traffic within this smartlink. The AI makes decisions based on the data available before your traffic is routed, then draws conclusions based on similar user profiles from the past. It then serves the most appropriate offer. This setup allows us to manage 4 000 offers within the smartlink and work with tens of thousands of webmasters, all with just four managers. We don't interfere manually at all.
Our unique algorithm is something we’re really proud of. I always brag about it because our performance metrics are significantly better than our competitors'. And we don’t hide this fact.
How much effort did it take to develop such a smartlink?
We have a large team of developers and mathematicians working on the calculations. There are gigabytes of information involved. It took a lot of time and we’re constantly refining it.
Here's a simple example: say there's an iOS user from a specific geo-location and gender. We already have data showing how users with similar profiles behaved in previous offers and campaigns. We predict their behavior, and the algorithm selects what it thinks would convert best for them. We compared manual and automated analysis, and the "auto" mode was 70-80% more efficient. The numbers are impressive. The key is our extensive database that the algorithm works with.
How to Start, Scale, and Calculate Profitability in an Affiliate Network
You mentioned using your own platform versus using someone else's. What do you think is better: a SaaS solution or a custom-built platform?
Definitely renting. Firstly, it saves you a lot of money. Secondly, you get full support and all technical issues resolved. If you’re considering building your own or renting, renting is the clear choice.
Plus, developers offering ready-made platforms for affiliate networks provide different functionalities based on traffic volumes, making it easy to scale as you grow.
But the main thing is not the tracker, it's the product as an affiliate network. If the product is poor, it won’t grow. If it’s hard to analyze data and other aspects, people will leave.
How do you grow and scale an affiliate network?
I'll share our evolution. Back in 2017, we entered the Asian market (India, Pakistan, Bangladesh). These were our main clients then. It brought rapid exponential growth — increased traffic but significantly dropped quality. We faced negative impacts with advertisers' rates dropping and realized this path was a dead end. Quantity didn't make up for quality.
In 2019, we made a tough decision to completely eliminate fraudulent traffic. This led to huge expenses on anti-fraud measures and AI algorithms.
We had to shift our focus from Asian to European clients, which took years and came with a hefty price tag. We parted ways with thousands of webmasters who brought in money but were unsustainable to work with. After this shift, quality webmasters with good traffic joined us, including many Europeans and even high-quality African segments.
This shift allowed us to get better deals from advertisers and significantly increase rates for our webmasters. Now, for example, USA rates start at $6, Germany at $8, New Zealand at $8, and Switzerland at $12-14.
How do you calculate the profitability of an affiliate network? How do you handle downturns?
Profitability is a complex thing and should be evaluated over one or two years, depending on the data. We have a strong development team, which is a significant expense, but renting a platform involves different costs.
Evaluating Campaign Effectiveness and Choosing an Affiliate Network
How should webmasters evaluate the effectiveness of their campaigns? What metrics should they focus on?
Focus on two metrics: EPC and EPM. EPC is crucial when there's a boom in CPC/CPM networks that sell traffic by impressions and clicks. On the other hand, networks like ours that convert by CPA offer models based on this.
|
It’s important for people to understand: the main metric to focus on is earnings per 1 000 impressions, not just the rate, with a proper cost analysis.
|
Many try to enter the cheapest segment — spending $100 on the cheapest clicks at one cent each, then wonder why it didn't work out. Bots consume the main traffic, and we get sign ups through VPN.
We had a case like this. We even published a news piece and sent a global notice to our affiliates to be careful. Unfortunately, some unethical networks push non-converting traffic through VPN and proxies in lower target categories. Advertisers won’t accept such traffic. It's 100% fraud and gets declined.
How to choose the right affiliate network?
Read reviews and don't hesitate to ask around! If you sign up with our network, ask managers what actions could get you banned, because we aim for long-term cooperation.
And as for shaving — it's now more expensive to shave than not to. There’s no point in setting up a shaving script for a few dollars, as long-term clients bring in more revenue. The numbers speak for themselves.
I'm sure that major networks and advertisers are not interested in shaving. Sales from the end product bring in more money than shaving a few dollars.
"Sales Drop Immediately When the Finals Series Starts in America" — iGaming and Dating: Seasonality, Telegram Casino Development, and New Trends in Verticals
How much does seasonality impact the verticals you work with?
We all know that seasonality affects gambling, but it turns out dating has its own unique patterns as well.
We did a deep dive into the data, and it turns out dating sees a spike in demand on weekends. What really surprised me, though, is how sales plummet when baseball season kicks off in America and during the finals. The same goes for the American football league, both at the beginning and the end of the season. During these times, men are completely focused on football and nothing else.
We also pay attention to national holidays. For Arab countries, it’s the long weekends and holidays during Ramadan.
On these days, we change our strategy, adjust the ads, and alter the GEO targeting. We advise our webmasters to do the same. Stability is never guaranteed when it comes to market fluctuations.
There’s another side to the market — December and January saw fluctuations due to a drop among acquirers in America, which affected everyone significantly, including dating. Several advertisers collapsed during this period, and we had to redirect traffic to others.
How has the Dating market changed? What are the main differences from iGaming?
When running gambling campaigns, you can use schemed traffic. In dating, schemed traffic is very risky for advertisers and is considered somewhat dishonest. Redirecting to Telegram, additional monetizations, and extra pushes — these tactics are fine but don't work well in dating. Different approaches and concepts are required.
I know of a case where iGaming folks redirected from Facebook to a Telegram landing page, then to a very shady site. This wouldn’t work in dating — such traffic would be flagged as heavily motivated and not genuine.
As for iGaming prospects, I think Telegram casinos are a promising topic. In the near future, pouring traffic into them could be very profitable. As long as Telegram maintains user privacy, allowing quick and anonymous betting without much hassle, it’s a winning idea.
It might not explode in wealthy Tier-1 countries, but it depends on the approach. For example, if a German casually and anonymously bets 50 euros just for fun, it’s still an option. The key factors are maximum anonymity and ease of use. These are the two critical elements that will drive this trend, especially with the issues of moving from Google Play to iOS and constant ad bans. I believe these casinos will become a trend.
What do you think about the spread of cryptocurrency globally now? As far as I understand, Telegram casinos involve crypto, mainly USDT. But currently, people are just getting used to crypto — we saw statistics showing that crypto is popular in Tier-3 but not in Tier-1 countries. What are your thoughts on this?
The difference lies in financial literacy between developed countries and Tier-3 countries. Regulators heavily scrutinize crypto use in Europe and America, whereas Tier-3 countries tend to overlook it. I think government efforts to regulate crypto will lead to the closure of exchanges, as we saw last year. However, crypto growth in Tier-3 countries will continue, impacting the entire industry using USDT.
This also affects iGaming, as payment issues remain unresolved. Even if payment issues are sorted, users will still desire anonymity, which won’t change.
Some webmasters have entered the dating sub-vertical by working with OnlyFans models through affiliate networks or personally. They say this is a new type of dating: users are brought from Reddit to OnlyFans, where they are engaged by professional sexters, and the affiliate earns a steady income. What do you think about this market trend and the future of dating in general?
The OnlyFans market is intriguing, but unfortunately, everyone tries to "make quick money," turning a genuinely interesting concept into a money-making machine. One model is surrounded by an office of sexters who send pictures and communicate with fans on the model’s behalf.
Initially, the idea of moving from offline erotica to online was cool. Many public figures in the US have OnlyFans pages and see no issue with it. Not all use erotic content to maintain their fan base; it's not the primary focus. But clever individuals realized they could earn good money within simpler legal boundaries while advertising this.
This will definitely work because the market is fresh and not yet saturated. Currently, related verticals to dating, like casual mainstream, are gaining traction. People genuinely want to meet, date, and start families, without adult content.
Recently, we discovered a new collaboration between two verticals — astrology and dating. It makes sense — if someone visits a dating site to find a partner, it means something isn’t right, and they don’t want to make the same mistake again.
Where’s the guarantee that the algorithm matching them with a partner won’t be wrong? Here’s where dating and astrology come together, offering a well-informed choice for a small fee. And just like that, you double your income!
|
You can find the first part of the interview here.
|
If you want to work with Trafee and test a smartlink with 4 000 offers, join the team here: Trafee.
