Kelkoo Optimizes Ad Spend & Saves a Full-Time Engineers’ Work With DataDome

Kelkoo specializes in e-commerce advertising, helping publishers monetize their audience through ads and ensuring that merchants receive quality traffic. To avoid paying publishers for clicks generated by bots, the company developed an in-house fraud detection system, but publishers often challenged the statistics. Using DataDome as an impartial third party to classify traffic, Kelkoo can more easily pay publishers fairly, and their engineers can focus on the company’s core business.

The Problem: Bot Detection Challenged by Publishers, Distracting From Core Business

“When we pay publishers for sending us traffic, we obviously only want to pay for real, human visitors,” says Michael Trommsdorff, CTO at Kelkoo Group. “The problem is that when it comes to defining what’s ‘good’ and what’s ‘bad’ traffic, the conversations can become a little heated.”

As an e-commerce pioneer with more than 20 years in existence, Kelkoo’s engineers have always been conscious that some of the activity on their website is robotic. To avoid paying publishers more than what’s fair, the company created and implemented in-house systems for detecting fraudulent traffic.

“We were actually quite successful in detecting spam,” says Michael. “However, because we have a vested interest in what is defined as ‘good’ traffic or ‘bad’ traffic, publishers would often challenge our numbers. With 20-30% automated traffic, there’s a lot of money at stake.”

“Secondly, all that automated activity is increasingly complex, which means that detecting it properly takes more and more time and resources. It’s not our core business, and it took our attention away from the things that we want to focus on.”

The Solution: Real-Time, Efficient Traffic Classification

Michael and his team finally decided to look for a third-party bot detection solution, which would provide Kelkoo with impartial traffic classification and take the responsibility of keeping up with new automation technologies off their plate. Among their top requirements were reliability and execution speed.

“We have a high volume of traffic, and clicks are transactional, so the detection needs to be very quick,” Michael comments. “We also have a fairly advanced setup and an untraditional use case, so it was important for us to find a reliable, long-term technology provider.”

During the evaluation process, another quality that stood out to the Kelkoo team was the ease of working with DataDome.

“One of the very positive things we saw is that the DataDome team is very responsive. Whenever we had feedback, it was taken into account in a swift manner, which raised our confidence level,” says Michael. “So although the technical solution was our number one selection criteria, the collaboration itself is a second important pillar.”

The Results: Impartial Classification, Important Time Savings

Since they implemented the DataDome solution, Kelkoo’s conversations with publishers have become easier. If any discussions about traffic classification come up, they can argue that the data comes from a neutral third party with specialist bot detection expertise.

Perhaps even more importantly, Kelkoo’s engineering team no longer needs to spend time keeping up with new traffic automation technologies, and can focus on the company’s core business. 

“In fact, our own detection systems showed quite similar results to DataDome,” Michael observes. “But to maintain that performance going forward, we would have needed to spend a significant amount of time. Staying competent on the topic and maintaining an efficient system would probably have required at least one developer’s full-time work.”

To conclude, what would Michael say to peers who are considering implementing DataDome?

“Given our experience, we would definitely recommend it. And again, I would like to highlight the very positive collaboration, with regular meetings and fast responses to our questions and feedback. Not all providers are so adaptive to the customers,” he attests.