Transforming a
credit bureau in real time
How Regis24, a German consumer credit bureau, matched hundreds of millions of identity records in near real time with Tilores — and turned its data infrastructure into a competitive edge.
Consumer credit bureau — credit risk & fraud data
Real-time identity resolution across bulk & streaming supplier data
Hundreds of millions of datasets on tens of millions of people
In production
A credit bureau built on data
Regis24 is a German consumer credit bureau providing data and AI-based risk scoring that help companies make credit-risk and fraud decisions on the richest possible data. Its hundreds of customers range from law firms and banks to e-commerce companies and online payment providers.
Founded in 2003 as a customer-address research agency for law firms — an alternative to querying local government residents' registration offices directly — Regis24 grew with the rise of e-commerce, spotting an opportunity to put its fraud-detection knowledge to work for online businesses. But winning that segment meant pairing that knowledge with superior, fast technology.
“The challenge is that we need to very quickly match data from many suppliers to common identities so that customers have the full data picture of an individual. Then we need to be able to search that complex, matched data. Technically, it is very challenging.”
Big data, in real time, without compromise
As a consumer credit bureau, Regis24 handles a vast volume of datasets whose complexity has always challenged the industry. The most critical part: maintaining absolute accuracy of the data it supplies to customers while delivering it at maximum speed.
Some supplier data arrives in bulk; some arrives as a near-real-time stream. All of it has to be matched to common identities fast, then made searchable — and customers in the new e-commerce and online-payment segment cannot wait more than a few hundred milliseconds for an answer. As Regis24 shifted its focus toward these clients, the big-data, real-time challenge only intensified.
“Relational databases just don't work for this sort of data — it's too complex and the databases get locked very easily. Graph databases are great if you want to know everything about a specific individual, but they don't work at this scale when you need to search that linked data very quickly. Our customers can't wait more than a few hundred milliseconds.”
When nothing else met the speed, scale, and cost bar
Regis24 experimented with a range of technologies — graph databases, Elasticsearch, Apache Spark — to crack the problem. None could deliver the maximum speed, swift scale, and effective cost control its new customer segment demanded.
In Tilores, Regis24 found a data-infrastructure technology that could assemble all its complex identities — built on hundreds of millions of datasets — master data usability, reach effectively unlimited scale, and allow near-real-time searching under 150 milliseconds, all while cutting unnecessary operating costs.
A credit bureau doesn't only need to search records. It must match data from several suppliers to common identities, keep those identities current as new data streams in, and return the resolved context fast enough for live credit-risk and fraud decisions — all at a scale where relational and graph databases lock up or slow down.
“The unlimited scaling is impressive. Sometimes we have imported a few tens of millions of datasets in one go, using Tilores' serverless ETL tool, and Tilores just handles it with no problems. The data is ingested, deduplicated, and assembled into entities and available for searching nearly instantly.”
Their matching rules, running serverless
Regis24's data science team had spent years refining its matching rules, so confidence in them was high. Recreating those rules in Tilores proved straightforward — and while testing the software before importing all their data, the team was even able to fine-tune the rules to improve the matching rate.
Because Tilores runs on serverless technology, Regis24 scales operations without limit. When an e-commerce or payment customer sees a sudden surge — a Black Friday spike, for instance — the Tilores infrastructure scales with them exponentially, with no hitch.
Deployed in Regis24's own cloud
For core, critical data infrastructure involving sensitive data, Tilores runs inside Regis24's own cloud account. Regis24 continues to handle all its customers' data directly, which simplifies GDPR compliance — there is no external data controller or processor to consider.
“Our data scientists have been working on data matching for years, so we were confident our rules were working well. Fortunately, it was easy to recreate our previous matching rules in Tilores. While testing the software before importing all our data, we were even able to fine-tune our rules to improve the matching rate.”
A future-proof data infrastructure
With Tilores, Regis24 found a technology that supercharged its data infrastructure. Scaling and capacity are now handled automatically, and the ability to provide complex data in real time lets Regis24 sell superior products into the high demands of e-commerce clients.
- ▹ Reduced response time by 95% — from an average of 3 seconds to 150 milliseconds
- ▹ Cut prototyping time by 86% — from an average of 23 days to 2 days — freeing data-scientist time
- ▹ Reduced monthly infrastructure IT costs by 60%, thanks to serverless technology
- ▹ Lowered server administration and maintenance FTEs by 90%
- ▹ Connected different data sources into one single source of truth, replacing three separate databases
- ▹ Improved data redundancy — data now stored across three data centres instead of one
Thanks to Tilores, Regis24 transformed its data infrastructure, acquired a new customer segment, and made a transformational change from a consumer credit bureau into a data-tech company and e-commerce fraud-detection partner.
“We love hosted services, but for core critical data infrastructure concerning sensitive data, the software runs in our cloud account. Tilores provided results that went beyond the pure entity-resolution technical solution.”
Resolve identities in real time —
at credit-bureau scale
See how Tilores ingests, deduplicates, and assembles hundreds of millions of records into searchable entities. Available on AWS Marketplace.
In financial services? See the Financial Services and Fraud Detection solutions.