Layer 1 copy

Sift Science is an obvious solution. Once you start using it, it's immediately useful.

Andrew JervisCo-Founder

Overview

  • On-demand marketplace connecting mechanics with people in need of car repair

  • Available only in the UK, growing rapidly

Challenge

  • Real-time booking and payments fraud

  • Rules didn't scale as business grew

Solution

  • Full integration with existing CRM in just days

  • Device fingerprinting and IP address data to quickly spot bad users

Results

  • 80% reduction in manual review time

  • 100s of suspicious bookings caught in the first 6 months

Overview

Paving the way for easy car repair

ClickMechanic is an online marketplace for car repair, making it easy for any car owner in the UK to get their car fixed. ClickMechanic matches customers with trusted local mechanics, offering users an easy way to find, book, and pay for services – all through the app. Users can access mechanic-specific quotes online and get matched with a mechanic in real time, while mechanics have a one-stop shop for invoicing and charging. To top it all off, the company offer an interface optimized for desktop navigation and responsive design.

As a thriving young company, ClickMechanic’s site and services enticed an increasing number of visitors. Customers were drawn to ClickMechanic’s real-time quoting engine and breadth of vehicle services, and the company's user base quickly grew 10% month over month growth. Now, ClickMechanic facilitates thousands of bookings per month across the whole of the UK.

Challenge

Roadblocks to happy customers

ClickMechanic’s business relies on creating positive, trust-based connections between mechanics and their local customers. The company facilitates payment and communications, but is often the last to know when chargebacks and friendly fraud occur. Although fraudsters have been fairly rare on the ClickMechanic site, credit card fraud is the main type of bad behavior exhibited. Whether being tested by fraudsters with stolen credit card numbers or potentially good customers attempting to dispute a service because they simply don't want to pay (known as friendly fraud), ClickMechanic’s reputation in the marketplace and the real threat of losing mechanics’ business were on the line.

As an on-demand business, ClickMechanic operates in real time and outside of the static fraud prevention rules that govern many other online businesses. For example, a good user may break down while traveling and book a service with a different address than they usually use, which you would expect from someone in need of car repair! While unmatched billing and service addresses may raise flags for traditional e-commerce or on-demand companies, ClickMechanic’s users are often operating from locations outside of their homes. However, a bad user may use a stolen credit card number to book a similar service or order repairs with expensive new parts, capitalizing on the expectation that ClickMechanic’s appointments are often next-day. As part of their process, ClickMechanic places a pre-authorization charge on the credit card to hold the appointment with the mechanic, and then when the service is actually completed, the mechanic processes payment through ClickMechanic’s platform. However, weeks or months later, that mechanic might get a chargeback when the real cardholder disputed the purchase with their bank. Only occasionally did the mechanics report the chargebacks to ClickMechanic, which made for a bad experience all around. Mad mechanics, mad credit card fraud victims, and ClickMechanic left holding the bill.

“The Customer Service Manager used to be on my back -- with Sift in place, he hasn't complained to me in a good few months.”

With fraud falling under the purview of the small Customer Service team, the company needed a proactive and preventative fraud solution that wouldn't bog down their lightweight and agile systems. When suspicious activity first began appearing on the site, the customer service team began investigating and manually created spreadsheets of bad users and heat maps of scammy users. However, this practice wasn’t scalable, with team members constantly creating and updating rules, while mechanics were getting frustrated at the less-than-optimal experience they were having with the marketplace. In order to keep chargebacks low while their business needs grew, ClickMechanic sought out a more efficient and accurate solution to layer on top of their in-house systems.

Solution

Accelerating fraud management with accuracy

The Customer Service team, led by Chris Glover, needed a way to prevent fraud that didn't require endless hours of manual review. ClickMechanic Co-Founders Andrew Jervis and Felix Kenton found Sift Science through a simple Google search. It was easy for the ClickMechanic team to locate API documentation and start implementing Sift Science themselves without ever talking to a salesperson. At first, they opted for a lightweight integration, followed by a deeper integration a few weeks later. In a matter of days, ClickMechanic was tracking new user accounts details, booking/order tracking information, and payment tracking data. The Sift Science solution merged seamlessly with ClickMechanic’s existing customer relationship management system, providing visual cues and auto-flagging suspicious users. ClickMechanic was able to automate and streamline the review process – even more importantly, they were reviewing fewer orders and processing legitimate orders even faster than before.

“Sift Science allows us to weed out high-risk customers, which translates to time savings, cost savings, and creating a much better user experience for mechanics and good users.”

Every time a suspicious user arrives on ClickMechanic’s site, the CRM pulls the customer's Sift Score in real time. With just a click, the customer service team can dive into the user's details in the Sift Science Console, and quickly determine if they're legitimate or should be blocked.

Results

Happy mechanics, happy motorists, happy ClickMechanic

Six months after completing their integration, ClickMechanic uses Sift Science confidently. While the team used to spend a couple of hours every week dealing with chargebacks and potential freindly fraud issues, they now spend just a few hours per month even though order volume has doubled. With the full integration, manual review time has also decreased dramatically, dropping from 10 to 2 minutes. Having the Sift Score API plugged directly into ClickMechanic’s CRM system allows for efficient order reviewing and highlights the most relevant data for the Customer Service team.

For ClickMechanic, the most useful Sift Science features have been the device fingerprinting and IP address tools, both of which have – in Andrew's words – “undoubtedly caught hundreds of suspicious bookings that shouldn't go ahead.” By automating fraud review based on Sift Score, ClickMechanic uses these data points to preempt chargebacks, which had been costing them considerably more than just the value of the mechanic's service or the stolen parts. As Chris explains, “When you do get a chargeback, the amount of customer service input required is enormous. Sift allows us to preempt the fraud before the charge.”