Sift as a Provider of Fraud Screening Services
What do you like best?
I've been working with Sift for a couple years now, and one thing that has stood out to me from the beginning is their support. Both the team as a whole, and our account manager, are incredibly responsive, and attentive to our needs. Beyond that, the interface is understandable, the APIs are easy to use, and the overall experience has been positive. As for effectiveness, we have used this tool across multiple brands in the enterprise, and Sift has been a success with regards to each implementation, both in identifying fraud, and allowing us to reduce false positives. This is due in no small part to the machine learning aspect that allows Sift to adapt to any current fraud trends that we are experiencing in a much faster way than we could previously.
What do you dislike?
One area where Sift still has room to improve is the adaptability of the reporting from the user-side. Whereas there are pre-made reports with much of the information that a fraud manager would need to run the day-to-day operations of the department, deeper dives into the data or pulling queries with specific parameters require support/your account manager to do the pull for you and send you the results. This does introduce some delay into the process. Though conversations with our account manager, as well as talks with some of the product team at Sift, we understand that some of these issues are on their roadmap to improve, and allow for more granular insights into the data than we are currently able to achieve ourselves.
Recommendations to others considering the product:
Make sure you have a good understanding of what data is and isn't available, and your ability to get that data, when starting the process with Sift. To get the most benefit from the machine learning model, it is heavily reliant on having complete and accurate data fed into the model from the beginning. Engaging your engineering teams early on what data is available, and what information can be collected, is going to make the entire implementation much smoother.
What problems are you solving with the product? What benefits have you realized?
We are currently using the payment protection offering from Sift in order to screen orders on our platform. Previously, prior to when we first reached out to Sift, we were in a situation where we had multiple brands each with their own disconnected fraud screening tools. Some were in-house built tools, some of which were no longer being maintained, and others were from other solution providers. With all systems, our screening was based on a simple rules-based engine only. With Sift, we have been able to consolidate into one interconnected system in order to gain insight into trends we were blind to before. With the introduction of the machine learning aspects of Sift, we are able to detect fraud that would previously have made it through our rules-based screening. Finally, also have better insight into the exact impact that fraud and fraud screening has on our business with the insights we are able to gain from the implementation.
