How Sift Science Helps On-Demand Businesses Grow
On-demand businesses are growing, attracting more than 22.4 million consumers annually and $57.6 billion in spending. But as they grow, they face a number of unique challenges and a myriad of fraud problems. Sift Science has had a lot of success with on-demand companies, and new features we’ve introduced over the last year make it even easier to grow your business.
Here’s an overview of how Sift Science gives on-demand companies the power to grow.
1. Live Machine Learning for real-time results
On-demand businesses need to be able to act in real time to block fraudulent orders and make sure as many good orders as possible go through. Sift Science is the only fraud prevention tool powered by Live Machine Learning, automatically learning new fraud patterns and adapting to changing fraud attacks without extra integration work. For example, if you are able to catch a new fraudster attempting to use a stolen card, we’ll learn other users like that fraudster (same mobile device, similar email, etc.) are likely to be fraudulent.
We apply Live Machine Learning to custom events and fields, too, so we can pick up on fraud specific to on-demand businesses. Our integration guide has some recommendations like pickup location, time to delivery, and minutes until service.
2. Multiple predictions
While chargebacks from payment fraud are painful, on-demand companies face many other challenges, too. For example, it’s common to dole out promo codes as a way to get more customers to sign up for your service. While that certainly attracts new, recurring customers, it also presents an easy way for people to take advantage of the system by using the same code over and over again or referring themselves. You might end up losing more than you make if promo abuse runs rampant.
Additionally, there are many on-demand companies that also operate as marketplaces, where fraudsters may create fake accounts to scam legitimate users. If you want your good users to have confidence in your site and keep coming back, they need to be able to trust the person walking their dog or driving them to a new location. Fake accounts damage user trust and your brand.
When you’re scaling quickly, you don’t want to worry about managing different tools to fight different types of fraud. Last year, Sift Science rolled out new fraud and abuse prevention products – all on a single platform – to address the varied concerns of modern online businesses. We have different models for Account Abuse, Payment Abuse, Content Abuse, Promo Abuse, and Account Takeover – so you can get more accurate scores for each problem and act accordingly. For example, you might not want to ban someone who commits promo abuse, but you do want to ban someone who commits payment fraud. The different scores we generate for each fraud type helps you make informed decisions.
3. Automation
On-demand businesses need to meet customer demands in a very short amount of time, meaning they can’t hold or review orders the way many other companies do. Our Workflows tool makes automation easy. You can build custom flows based on different scores and specific attributes, automatically blocking or accepting users as well as routing them into review queues. You can also trigger Workflows on orders to stop them from going through, or even block at the point of account creation in instances of Promo or Account Abuse.
Let’s say, for example, you want to block deliveries when a user has used multiple phone numbers in a short period of time. You can easily create a Workflow to block orders that meet this criteria and adjust whenever you like without developer intervention.
Read more about how machine learning can power business growth in our Kickstart Your Fraud-fighting Strategy ebook!
