Team track record · Risk
Deduplication saves a UK bookmaker about £80,000 a month in affiliate fraud
One of the world's largest and oldest multi-channel bookmaking and gaming companies, headquartered in the UK and operating in more than 25 countries, with annual revenue above €3 billion (2020).
~£80Ksaved per month in suppressed affiliate payments
93.7%accuracy identifying duplicate accounts
25+countries covered
The challenge
- Affiliate fraud through multi-accounting inflated acquisition costs (CPA).
- Thousands of wagers a day across phone and internet channels.
- Duplicate accounts had to be found across the whole brand portfolio.
The solution
- Weighted biographic and demographic attributes to score how likely two accounts belong to one person.
- Cleaned and standardised account data, then used text mining to match names and addresses.
- Tested multiple machine-learning algorithms and fuzzy-logic matching to pick the most accurate approach.
The outcome
The model identifies duplicate accounts with 93.7% accuracy. Suppressing affiliate payments for duplicates saves the operator approximately £80,000 every month.
A project by the BizAcuity team that built GAMWIT. Originally published at bizacuity.com.
More from our track record
Revenue · Gibraltar · UK market95%LTV prediction accuracy
Affiliate optimisation with player LTV
Read case study →Revenue · UK · Europe · Asia10%reduction in player churnPredicting churn before players leave
Read case study →Revenue · Asia5.33%increase in revenueSpending bonuses on the right players
Read case study →Talk to us
See what GAMWIT would find in your data.
Book a 30-minute demo with our gaming analysts, or start with a health check to see where your business is leaking value.