How Peer-to-peer lending company uses data to serve more consumers
LendingClub is an online financial community that connects creditworthy borrowers with intelligent investors. Since its inception in 2006, the organization has processed approximately $18 billion in loans, including personal, commercial, student, and medical loans.
LendingClub’s Product Analytics team is building infrastructure to enable the company’s over 1,000 workers to be as self-sufficient as feasible. The team also handles A/B testing, web analytics, surveys, and SEO. Director of Product Analytics Alan D’Souza and Senior Product Analyst Amanda Rosenberg lead LendingClub’s analytics implementation and strategy.
The team has made a tremendous impact after using Heap. They discovered little friction spots in the consumer experience, allowing them to serve thousands more clients. Check out Bridge Payday for more loan offers.
Looking for a better way
LendingClub hired Alan for site analytics two years ago. The team has used a solution from a well-known business provider for years but undertook minimal route analysis. Concerns regarding data accuracy kept it out of the team’s efforts. Delays in adding new LendingClub pages occurred due to rapid product and page evolution.
Amanda said it was challenging to see user activity over several pages. “Either we couldn’t discover anything to guide our judgments, or it was impossible, so we didn’t have answers to inquiries like, ‘How many people clicked on View Agreements on this page?’ How many of those who clicked View Agreements signed up?’ Our primary tool couldn’t assess that formula, and the calculations were opaque.
When Alan joined, he had to decide whether to continue with the existing product or transition to something new.
Because we intended to undertake sophisticated analysis, the existing system wasn’t flexible enough. ‘How many individuals did this in this order?’ We wanted to trace every click, segment, and group them. We wanted a more versatile tool.”
List of analytic tools
Alan and his colleagues recognized they needed a tool that could:
- Event-driven vs. pageview-driven
- Easy to use. They needed substance, not just beauty, in an analytics tool. One essential worry was speed and ease of implementation. Is the device ready to use?
- They have added additional data. “It wasn’t feasible to foresee every conceivable question,” Alan said. ‘How many users click on this agreement popup in the footer of this specific URL?’ Those are minor details I’d never notice.” They sought to make event tracking easy by not having to pick what to tag ahead of time.
- I have made raw data accessible. No matter how cool the new tool’s UI was, LendingClub needed access to their raw data to do A/B tests, merge numerous data sources, and run predictive modeling.
“Simultaneously deployed the Mixpanel, Amplitude, and Heap scripts,” Alan says. After ten minutes, Heap has all data. I didn’t want it to work.
Improving client experience
After choosing Heap, the team dug through their backlog. Amanda wanted to look at loan request friction spots, notably validation problems. When a user doesn’t fill out a form correctly, the system prompts them to retake it rather than moving them on to the next stage. We should identify them, count the affected persons, and decide which engineers to address first.
With Heap, “I was like a detective,” Amanda continued. The Heap data allowed us to track how many validation failures we received, target specific ones, and see whether a user had converted. For example, a checkbox was left blank. If a user didn’t check a box, they’d likely return, do so, and go on. That may mean we need to make the box bigger. If they don’t go forward, we need to eliminate some additional friction in the process.”
To distinguish between validation failures and user mistakes, Amanda could use a user’s email, and user ID to search Heap for how many times they had errors, attempted to repair them, and where they got stuck. Then she could see how many people had the issue and didn’t proceed.
“If 300 individuals per day encounter the same issue, we know it’s worth addressing. “We multiply how many individuals were touched by the average conversion rate,” Amanda explained.
The advances we’ve achieved are tremendous. They help thousands of visitors have a better experience on our site. They help thousands of visitors have a better experience on our site. It means a lot to them and us.
Using Redshift to combine data sources
The team also uses Heap for A/B testing and persona generation. Alan and Amanda use Redshift to combine data from multiple sources with Heap data to gain deeper insights. They may use A/B testing to enhance Heap data, export it to Redshift, and model it how they like.
“You can look at any down-funnel and non-test page activity, not only A/B testing data. “We’re only touching the surface of A/B testing analytics on top of our raw Heap data,” Alan remarked.
They’re developing comprehensive user personas by merging Redshift Heap data with user data from their database (e.g., credit score, location, age). Because mobile traffic is increasing, the team utilizes Heap to improve the mobile loan application experience.
“We wanted to know whether users preferred beginning and finishing applications on desktop, mobile, or both. ”That’s critical because it determines whether you should optimize an experience for people switching devices or making each flow the best possible,” Alan says.
Increasing data literacy
Using Heap has changed Alan and Amanda’s workflows in unforeseen ways. Fewer meetings and quicker decisions. The team is creating lightweight dashboards for the corporation and intends to allow everyone to access raw data in Heap.
“Existing measures like page summary conditioned people,” Alan says. “But they can’t be done. Precise inquiries enhance things for our consumers, therefore we want to spend less time on high-level, feel-good figures and more time on the specific things that produce value every day. Only Heap allows everyone to instantaneously answer business questions.”