Predictable influencer marketing.
Influencer marketing is the highest-leverage channel most brands have available — and the one most often run on instinct rather than data. We exist to close that gap.
Since 2022 we've delivered 1,400+ placements, 180M+ views and managed $8M+ in campaign spend across mobile games, apps, beauty, fashion, tech and adjacent verticals. Every campaign feeds a structured dataset: which creators move which audiences toward which conversion events, in which markets, at what cost.
That dataset — combined with a 287M+ creator index — powers InfluBase.io, our partner platform. It is the prediction layer behind every campaign we ship today, and what lets us tell you the expected CPA before you commit budget.
From a $32K disaster to a system.
In early 2021, I ran my first influencer campaign. Budget: $32,000. It delivered 2 conversions. Cost per conversion: $16,000. The product's average paid-search CPC at the time was around $4 — meaning the influencer channel was performing roughly four thousand times worse than the channel everyone considered “expensive.”
The natural reaction is to blame the creators. The actual problem was upstream. Nobody had been able to tell me, before signing the deals, what return to expect — because nobody was tracking influencer outcomes the way performance teams track every other channel. We had run a real budget through an industry that talked entirely about reach and ignored conversion.
That campaign became the brief for everything we built next. In 2022 we launched GrowThunders as a deliberately narrow agency: every campaign instrumented for full-funnel attribution, every creator scored against historical performance, every result fed back into a structured dataset. Over the next three years we shipped 30+ performance-driven campaigns with brands including BoomBit, Modivo, Wargaming, Gamesture and Ten Square Games — then expanded the same playbook into beauty, fashion, fintech, tech and automotive.
That dataset is what powers InfluBase.io today — our partner prediction platform. Every brief we run gets matched against a 287M+ creator index and the closed-loop history of every placement we've shipped. The forecast you see before signing is a model output, not a salesperson's guess.
The $32K disaster became a discipline. That's the whole company.
Meet the team
Two founders. $8M+ tracked spend across 1,400+ delivered campaigns — we ship the work ourselves rather than scale headcount before the playbook is proven.
Oskar Porębski
Founder & CEO
- Closed 1,400+ creator deals across 67 markets — the dataset every InfluBase.io forecast leans on.
- Built the playbooks the agency runs from end-to-end: brief, match, ship, measure.
- Background in performance marketing — brought paid-acquisition rigour into a channel still sold on impressions.
Konrad Giżycki
CTO & Co-founder
- Former CEO/CTO of GO.PL — built and exited a high-traffic platform to Wirtualna Polska, one of Poland's largest media groups.
- Architect of the 287M-creator index and the matching engine that scores every brief.
- Owns the prediction model: takes country, vertical and audience overlap as inputs, returns the CPA band we sign.
We don't sell creators. We match them — by predicted outcome, not promised reach.
Problems we solve
Which creators should I choose, and what should I pay them?
InfluBase.io scores every creator on audience-product fit, expected reach, conversion likelihood and a fair-rate band. We send you a shortlist with predicted CPA per placement — not a list of suggestions you have to vet yourself.
How do I structure deals so the campaign is ROI-positive, not a guessing game?
Performance-style deal structures (rate cards calibrated to predicted impressions / installs / signups), milestone-based payouts and live tracking. If a placement underdelivers we know within 72 hours and adjust the next ones.
What assurance is there that a $50k–$300k campaign delivers expected results?
A pre-spend forecast you can hold us to. Our prediction model has 80–90% confidence on CPA bands, calibrated against $8M+ of delivered campaigns. If the forecast is wrong we eat the variance, not you.
How do I scale influencer marketing like I scale paid ads?
Programmatic creator selection, automated brief generation, standardised tracking pixels and dashboards that mirror what your performance team uses for Meta and Google. The operational overhead per placement drops to roughly the same as a paid-ad creative iteration.
How we operate
Four operational disciplines built into every campaign. They are the reason our delivery numbers stay consistent across verticals and geographies — not optional add-ons, not premium tier features.
The thing that ties them together: we don't sign exclusive creators and resell them. We index the entire creator market — 287M+ verified records — and let InfluBase.io score the right match for your specific product. The agency's economics don't bias the recommendation.
Brief → forecast in 2 days
Send product, market and budget. We come back with a 1-page brief and a calibrated predicted-CPA band. InfluBase.io scores against the entire 287M-creator pool before we recommend any roster.
Market-neutral matching
Zero exclusive-creator deals. AI scores the full creator market against your product on overlap, engagement quality and conversion probability — the right match wins, not the creator we'd locked a discount on.
Authenticity filtering
Engagement-pattern audit on every candidate — bot networks, comment pods and engagement farms get filtered before the brief is written. Filter 2 of 7 in our vetting sequence, not a premium add-on.
Lifetime attribution + D60 report
Every placement gets a unique link or promo code, tracked for the lifetime of the cohort. At D60 you get a deep-dive report — variance vs. forecast, per-creator economics, retention by source, raw CSV included. Not curated slides.