Interview with Klaus (Gipsy MBA): Working Solo, Traffic, and Getting into Gambling
The gambling vertical is a constantly shifting landscape, where quick adaptation and data‑driven decisions separate strong media buyers from the rest.
In this interview, Klaus, team lead at Gipsy MBA, shares his journey from being a solo player to managing a high‑performing team, discussing growth, experimentation, and the principles that keep his results consistent.
Working with Gipsy MBA
How did you join Gipsy MBA?
I used to have my own team for nutra, and I have made good friends in this field while travelling to conferences. One of my best friends is the CEO of a well-known nutra platform, and last year, I was really looking forward to getting into gambling, as it seemed easier and more scalable.
When my volumes started to drop a bit, I asked him to help me join a CIS-based team, as I wanted to acquire knowledge more quickly and needed a wallet to work. He already knew one of the owners of Gipsy MBA, and he arranged my first meeting with him. I also had some friends that I met in Dubai a few years ago who were actually working in Gipsy MBA, so I had already heard about this team.
As a foreigner, it is a bit challenging to get into the CIS market due to the language barrier, so many teams rejected my proposal.
How quickly did you grow from buyer to team lead?
It took me around 5 months. I didn’t start as a media buyer — I was running solo, and I achieved my first results on my own.
In the CIS market, the terms "solo" and "media buyer" are often considered equivalent. However, in our experience, it’s not the same: a media buyer typically works as an executor, following an existing model, while a solo buyer actually develops the strategy, tests, and scales solutions — basically acting as a team lead.
How many people are on your team now, and how many were there at the start?
Throughout this year, I’ve worked with around 10 buyers. As a part of the team-building process, the structure was refined to a core group of four — a necessary step to ensure that the team consists of buyers who can adapt to evolving market conditions. With this foundation in place, I’m now gradually expanding the team and planning to onboard three additional buyers while continuing to scale my own traffic.
What helped you grow into a team lead?
I realized I was becoming a team lead when I started testing more approaches and GEOs and understood that staying in one operational role was no longer possible. To continue growing, I needed to build a structure.
At that point, I was working with multiple GEOs simultaneously and achieving good results, but I couldn’t give full attention to each flow. I was making money, but I knew I could scale much further by involving experienced people who could focus on specific directions.
Together with the owner, we realized it was the right moment to expand the team. I still work with helpers who simply repeat my traffic, and that’s important. But for me, a real buyer must be able to readapt, build funnels, and think independently. Copycats can help, but they’re not the core.
Current Work
Which GEOs do you work with?
For some time, I used to work a lot in LATAM, like Argentina or Mexico, but now I’m mostly running Europe, focusing on key markets.
Which GEOs bring the best results?
The top-performing markets are usually some of the main European ones, like Italy. For these GEOs, traffic tends to be more stable, scalable, and easier to optimize.
What key changes have occurred in your team’s structure/strategy over the past year?
I think we’ve been evolving with the market in terms of team structure. Since we started redeveloping the Facebook area, it’s been a real challenge. Finding good buyers right now is complicated — there are a lot of people who want to learn, but identifying early-stage skills to avoid an extremely long learning curve is hard.
We ended up implementing a sandbox system, which we’re still polishing, but it has already delivered its first results. Basically, we work with different phases and objectives, focusing on the logic of process development.
For example, first we try to understand how a buyer can develop in an environment where they need to create with limited resources. This is extremely important right now, as approaches are migrating and there’s a constant hunt for the golden egg. A buyer who can create has a much longer LTV than one who is stuck only using spy tools.
This also changed the team structure. We’re trying to hold internal calls with buyers and do one-on-ones, but at the same time, I’m also promoting partnerships, because I believe networking is really important at this stage.
We also try to avoid overlapping the same GEOs unless it’s really necessary. Another challenge is that everyone wants to pick the same markets — and this happens even inside the same team. So we’re trying to open their eyes to the fact that there is food in every country; you just need to understand how to cook it.
At the moment, my team consists of 5 buyers, including me, and I’m happy that all of them are in profit. There are no red numbers right now, and some of them have already started scaling to thousands. After a lot of work from the whole team — designers, helpers, BDMs, and everyone involved — we’re finally seeing solid results.
How do you currently handle the storms?
When it comes to handling the storms, we have a lot of resources, and there are open opportunities for me to test new initiatives and adapt to restrictions.
That’s why working with agencies is especially important for me right now, mainly because they have access to stronger ad accounts and, in some cases, older ones that can be difficult to build or acquire independently.
Having solid assets with spending history — and, more importantly, high trust — is a must to work smoothly. At the same time, even within the same agency, not all assets perform equally well.
What I usually do is request a larger pool of ad accounts and then identify the strongest ones. In practice, out of ten accounts, only one or two tend to show stable approvals, good metrics, and long-term potential — that’s how I find reliable and scalable traffic.
Latest Cases
Tell us about your latest successful case
One of my biggest cases was Chicken Road in Tier 1 during May and June, where I made $1MM+ in Tier 1 and $300K+ in LATAM in revenue. I managed to achieve strong results across multiple GEOs, focusing on scalability and traffic quality.
A detailed breakdown of the structure, scaling logic, and creative approach will be shared in a separate post about working with Chicken Road traffic in 2025 in my Telegram channel.
What lessons have you learned from unsuccessful campaigns (if there were any), and how did they influence your future strategy?
Failure is a daily part of media buying — especially when you manage several GEOs at once. Right now, from my own traffic, I have 6+ GEOs running.
Some of my recent failures have actually been very insightful. For example, I once tested a funnel using only a single ad account. I was confident in the setup, but the test didn’t deliver the expected results.
After that, I tried the same funnel on several other ad accounts. I recovered the money in one day and validated the approach. The next day, it made $7K profit in a single day. The only things I changed were the assets, fan pages, and ad accounts.
Because of this, what I do now is launch multiple ad accounts even during testing. Always with small budgets, but I think it’s a must. I was actually talking about this today with one of my buyers.
For example, in one of my GEOs, I currently have 7 ad accounts running:
- 2 of them are heavily scaled
- 3 are stable, but can’t scale too much because of higher costs
- 2 of them suck and are losing money (obviously, those will be killed soon)
The key point is that all of them run the same setup — same creatives, same campaigns, and even the same ad account configuration.
Sometimes I also launch different campaigns inside the same ad account, and you can still find similar results. It’s not rocket science: we launch broadly. But then the question is, why do I still launch different ad accounts?
That’s probably the second big lesson I learned recently.
Anyway, I think we learn new things every day — and not only about traffic. In some cases, we just need to improve the design process to scale faster. In others, we need to work harder on the offers. And occasionally, the problem isn’t the funnel at all; it’s the offer.
Future and Trends
Which trends do you expect in 2026, both globally and in your GEOs?
I think we are exactly at the point where new trends will start to emerge. You know that most of the traffic was going to Chicken Road; so far, it has been going down in the last few months, like the trend is slowly dying.
I see a really high trend in using AI UGC. I think the possibilities with the actual AI platforms to create videos are really broad. There are so many things that can be done. So I think most of the ads that we are going to see in 2026 Q1 will be AI-generated, and a lot of UGC will be based on this.
As for the games, I think there is no clear winner yet. I hope there will be a higher trend for the next year, but I see that most of the game providers are in the process of testing new approaches. I think we’ll see a lot of rotation — fast‑rise, fast‑fall games — until the next viral formula appears, as well as new approaches.
Do you think Meta’s requirements will become stricter?
I think we are still far from fully restrictive policies from Meta. However, every time Meta becomes more demanding, it happens gradually rather than all at once. From a numbers perspective, gray traffic is still part of the platform’s revenue, so it would not make sense to remove it completely. So far, I have mainly noticed that Meta’s AI analysis has become more advanced.
For example, you can already see that many words are being detected by AI in the videos. You can search for gambling-related keywords and find ads with no text or subtitles at all, but the words are still spoken or appear somewhere in the video. This shows that the system now listens and transcribes video content quite accurately.
I don’t think there is a strong motivation to limit everything completely — it seems Meta already has enough tools to restrict ads when needed.
At the same time, I’ve noticed that rejections have become tougher lately. With images, it can already be difficult to pass moderation, and depending on the ad account, using more aggressive video approaches, such as news-style creatives, can be nearly impossible.
What advice would you give to those who work with Facebook?
A lot of people think that using Tyver or any spy tool is the solution to all their problems with traffic. And yes — that’s true at certain points of the year, especially when a new trend starts and traffic is flowing heavily, with many people running profitably.
But right now, we’re at the end of one of the biggest trends, and this is where the real workers appear. At this stage, you either create something new, build a fresh funnel, or get lost. Everyone is on the same road, copying and testing the same funnels, without anything truly scalable or solid.
My main advice is to go further: understand the audience and create from scratch.
One more key recommendation is to use team data. This is critical during the testing phase. If you have your own data, everything becomes much easier compared to working blind. Metrics like CPM, CPI, CPR, CPA, and especially EPC allow you to compare performance with teammates when needed.
We already have everything to understand where we are before waiting for an FTD. This is a real game-changer, especially when moving into Tier-1, where CPAs are high, and testing can be wild. That is why I regularly share insights and updates from my current live campaigns in my Telegram channel, as well as in the Gipsy Channel — feel free to follow and stay up-to-date.