Customer story · Checkr

How Checkr uses Rippit to turn feedback from 120,000 customers into a product edge

Most of Checkr’s roadmap used to go to its top 20 customers. Now problems from every customer are ranked by how much they hurt, and the ones driving churn reach product and leadership in hours.

By Rippit Team · Last updated Oct 29, 2026

With Rippit, Checkr
  • Found its biggest churn risk

    Pinpointed the billing problem that would most reduce churn and took it to the Chief Product Officer with customer evidence

  • Knew which fix to prioritize

    Found why small business customers leave, which its structured data alone couldn’t show, and put fixes on the roadmap

  • Got answers by morning

    Answered a late-night CEO question about a problem in one state with a detailed report by the next morning

Overview
Industry
HR tech
Customers
About 120,000
Company size
About 1,000 employees

Checkr, a data platform best known for background checks, has about 120,000 customers, but most of its roadmap went to the top 20. With Rippit, the Growth Data Platform, it reads every customer conversation, joins it with warehouse data and ranks the problems that hurt satisfaction and churn. Product and leadership get that evidence in hours or days instead of months.

The stakes

Why did Checkr need to hear all of its customers?

Checkr has about 120,000 customers. Like most companies, it built for the biggest ones: about 75% of roadmap items went to its top 20 customers. The loudest and largest customers got the most of the roadmap.

Checkr has a hundred and twenty thousand customers. Seventy-five percent of our product roadmap items are for our top twenty customers, because you end up focusing where your revenue is. There’s a lot of value left on the table.
Eric BurdullisEric BurdullisSVP of Operations, Checkr

Checkr’s leaders see better products as how they pull ahead. That means knowing which problems matter across the whole customer base, not only at the top.

“The reason we want to fix customer issues is to be not just a little bit better than our competitors, but much, much better than our competitors,” said Eric Burdullis, SVP of Operations at Checkr.

The problem

Why couldn’t Checkr see which problems mattered most?

Checkr’s measures disagreed. The quality score said 99%. Customer satisfaction said 72%. Only about 7% of customers answered surveys, and mostly when they were very happy or very upset.

“We have a ninety-nine percent quality score and a seventy-two percent CSAT. There’s just a fundamental gap that we need to close,” said Eric.

The categories Checkr used to sort customer issues ran four levels deep and still didn’t say enough, and changing them meant retraining every agent. Even Checkr’s data and ML teams couldn’t find a clear reason in their structured data for why small business customers left.

The shift

How did Checkr compress years of work into months?

Eric Burdullis and Kristin Johnson had built a system like this before, at Lyft. It took five to seven years. At Checkr they asked whether they could do it in twelve months with newer technology.

“Can we do what we did at Lyft, but in the next twelve months? Technology has improved. Can we do it just differently and faster?” said Eric.

Instead of rebuilding the old approach faster, they rethought it. Their first hire was an analyst who specializes in working with AI models, not another quality reviewer. Then they built measures around what customers actually experience.

How Checkr uses Rippit

How Checkr puts its data to work for growth

Use case 1: Product priorities

How Checkr ranks the problems worth fixing

Checkr calls them Atomic Problems. AI classifiers read every conversation and sort each issue by two things: how the customer felt and whether the problem was solved. Low satisfaction and low resolution means fix it now. Problems that get solved but still leave customers unhappy point to the silent sufferers, people who had to reach out when they shouldn’t have.

The results go to the C-suite and product teams in a regular report. The loudest complaint is no longer the default priority. The biggest problems across all customers are.

“We produce a report that goes to the C-suite and product teams. It’s how we talk, measure and explain what’s actually going on,” said Kristin Johnson, Director of Shared Services at Checkr.

“Those roadmap items should not be vibes-based on what we think is cool to build. They should be solving real problems that inflect churn, retention or revenue, and we actually have that data now,” said Eric.

Related use case: Taxonomy Builder →

Use case 2: Churn

How Checkr found two hidden causes of churn

Billing. It kept coming up in the Atomic Problems report: about 600 tickets with a 47% predicted satisfaction score. Leadership was surprised, since Checkr thought its billing was clear. Customer quotes showed three or four specific problems underneath.

Eric took it to the Chief Product Officer with a three-page report on what customers were actually saying, linked to the specific customers leaving over it. “If you want to fix one thing that will positively impact churn, it’s this specific theme within billing,” he said. Kristin added, “Within less than a week, we had new information to people, and people are moving on it. That would have taken months before.”

Small business customers. This churn was harder to see. Checkr’s data and ML teams couldn’t find a clear reason in their structured data, so an analyst on Kristin’s team brought account data from Snowflake into Rippit, matched it to customer conversations and analyzed both together.

“It was the first time we were able to say we can’t get this data in a structured way. Let’s use our unstructured information to point us, give us that clarity,” Kristin said.

The report spread across the company, and fixes to prevent this churn are now on the product roadmap, though further out.

It’s this customer saying this thing at this time that lets our ML team go back and say, here are all the other data points. Marrying unstructured data and structured data is a really powerful match.
Eric BurdullisEric BurdullisSVP of Operations, Checkr

Related use case: Churn-Risk Bug Watch →

Use case 3: Leadership

How Checkr answers the CEO overnight

Late one night, Checkr’s CEO asked whether there was a problem with driving records from one state. Nobody knew. By the next morning, the team had a detailed report, and there was a real problem.

The Atomic Problems report is now linked in Checkr’s weekly business review, and Eric sends it to product leaders while they plan their roadmaps.

“It’s really on demand. It’s really fast, and I think it’s going to help us tremendously to make change in our product for the better, for our customers,” said Eric.

Use case 4: Customer experience

How Checkr measures how every customer feels

Checkr predicts satisfaction for every conversation instead of relying on the 7% who answer surveys. People talk to a chatbot differently than to a person, so Checkr measures each separately, then combines them into one score. Surveys are now used only to tune it, and the two agree about 85% of the time.

The same predicted score makes up half of each agent’s quality score. The other half asks whether the agent understood the problem, solved it and left the customer feeling good.

“We still survey our customers, but we don’t actually look at that. We look at this one hundred percent. And we report it all the way up to the C-suite,” said Kristin.

Related use case: Predictive CSAT →

Across the company

How is Checkr bringing every customer touchpoint together?

Checkr is building one place for everything it hears from customers. Sales and customer success calls from Gong now sit next to support conversations and Snowflake account data. When a problem customers report also comes up with sales prospects, the case for fixing it gets stronger.

“Really looking at the customer holistically, anywhere that we’re touching the customer or having conversations with them, we want ingested in the same place so we can mine those conversations and make better decisions,” said Eric.

Results

What results has Checkr seen with Rippit?

Checkr now builds from what all of its customers say, not only its biggest accounts. Product and leadership get ranked problems with customer evidence in hours or days, and fixes that affect churn are reaching the roadmap.

“We’re getting to use our time in so much more meaningful ways, to connect deeply to customers and deliver not only delight, but revenue outcomes that often aren’t coming from operations teams,” said Kristin.

Churn

#1 churn fixThe billing problem worth fixing first

A billing theme with a 47% predicted satisfaction score went to the Chief Product Officer with customer quotes and the accounts leaving over it.

Churn

RoadmapSmall business churn finally explained

Conversations joined with Snowflake data showed why small business customers leave, and fixes are now on the product roadmap.

Leadership

OvernightExecutive questions answered by morning

A late-night CEO question became a detailed report by the next morning, confirming a real problem.

Customer experience

85%Every customer measured, not 7%

Predicted satisfaction covers every conversation and matches survey results about 85% of the time.

FAQ

Frequently asked questions

What is Checkr?

Checkr is a data platform that helps people make better decisions, from hiring to dating apps to finding a babysitter. It is best known for background checks and has about 120,000 customers.

What are Atomic Problems at Checkr?

Checkr’s way of ranking customer problems. AI classifiers read every conversation and sort each issue by customer satisfaction and whether it was resolved, so the problems that hurt most rise to the top.

How does Checkr measure satisfaction without relying on surveys?

It predicts satisfaction for every conversation, separately for people and the chatbot, then combines them. Surveys are used only to tune the model, and the two agree about 85% of the time.

How did Checkr find why small business customers churn?

An analyst joined Snowflake account data with customer conversations in Rippit. That showed reasons the structured data alone couldn’t, and fixes are now on the roadmap.

How fast can Checkr answer an executive question?

Often overnight. A late-night CEO question about a problem in one state had a detailed answer by the next morning.

What data does Checkr bring into Rippit?

Support and chatbot conversations, Gong sales and customer success calls, and account data from Snowflake.

Find your next edge

Connect a source, set a growth goal and launch an agent that keeps working toward it. Your competitors are already copying. Start building what they can’t.

Start in Rippit

Create an account, write a goal like “Tell me which accounts are at risk, and why,” and Rippit guides you the rest of the way.

Try Rippit free →

Start from your coding agent

Copy this line into ChatGPT, Claude, Gemini or your coding agent. It takes it from there.

Set me up: rippit.com/setup.md