How Nextdoor improved NPS results and saved 80+ hours a month with Kapiche

Feedback analytics at scale

When your business has  “millions and millions” of customers and “mountains of text feedback,” how do you separate what’s trivial from what’s important?

With close to 600,000 open-ended comments collected and just sitting there, Nextdoor needed to process colossal volumes of feedback.

While the primary goal—creating a feedback analytics approach that could quickly process a vast amount of customer data—was straightforward, Maryam Mohit, Research Lead at Nextdoor, knew there was more to consider. 

She calls it  “the problem of the baby and the bricks. The baby is really attracted to shiny new objects, like disco balls. They're glamorous and filled with possibility.” On the contrary, bricks don’t seem that exciting. Bricks can be boring and mundane. But when you’re a business, bricks might be the things that deliver value to your users. 

So when Maryam sat down to think about what to do, she realized that her goal wasn’t just processing lots of data.  “Our end goal had to be identifying, quantifying, and understanding the importance of the stuff that emerges out of the data.” 

Nextdoor needed to get actionable feedback and uncover the customer needs that remained hidden—in Maryam’s words, “getting the baby to focus on the bricks, too”. 

Generating actionable data insights without the human resources

Nextdoor knew that customer feedback was a gold mine for their product team, so they started by having their neighborhood operations department tag incoming feedback manually. With over twenty thousand responses each month, manual tagging required over 80 hours of work each month. 

As if this wasn’t enough Maryam also had over half a million NPS survey responses sitting in a database, waiting to be used somehow.  They’d been faithfully sending out an NPS survey for years, but hadn’t done anything with the responses.

With that feedback, honestly, we had no way to deal with it,” she recalls. “We just did not have the workforce. No one was coding it. It was just sitting there.”

Nextdoor’s core challenge was that their customers were providing so much feedback they didn’t know how to process it. They were overwhelmed. Yet they knew they needed to understand what their customers were saying and which critical bugs they needed to prioritize.

Understanding drivers of NPS

Nextdoor deployed Kapiche to help them quickly process their massive backlog of customer feedback.

One of the first things the analysis uncovered was that login issues were becoming a big problem. Things they considered  “small bugs” internally were actually creating big problems for customers. 

How big?

Enough to drag their overall NPS down by 2.3 points!

Using Kapiche, we could clearly see that if we were able to address these issues, we would instantly improve the company’s overall NPS by more than 2 points.

When the product team understood how significantly these issues were impacting NPS, it forced the team to tackle them. But “login issues” is a broad term, and they weren’t sure where to focus. Kapiche enabled them to drill down further into the data, to understand subthemes and easily view customer verbatims tied to each subtheme, giving them a clear course of action to fix the login issues and bring up their NPS.

Drilling down to what matters

More often than not, customers use different vocabulary than product managers. One such example was  “ease of use.” Customers were complaining about it, and the big negative impact on NPS at Nextdoor was clear. Yet their product managers were at odds over what it actually meant. 

“The problem with ‘ease of use’ is that it's just so vague,” Maryam explains. Teams don’t find it appealing to solve something that’s broad and undefined. “It’s not shiny,” she says, “every product has problems with ease of use. That's just life in technology.” 

But she wasn’t ready to let it go by the wayside.

After using Kapiche to analyze the verbatims connected to the “ease of use” theme, Maryam and the research team were surprised to find out that they had missed the real problem all along. 

Customers were growing frustrated when they found interesting posts, but couldn’t go back and easily find them again. 

With this insight, Nextdoor had something concrete for their design and product teams to work on. “Without Kapiche, we wouldn't have known that it was an issue that mattered,” sums up Maryam.

Nextdoor also uses Kapiche to enable a process of continuous discovery and to distribute insights to teams across the organization. The research team sends an automated survey to a segment of customers each week, and responses are automatically funelled through to their data lake. 

The insights uncovered are loaded onto dashboards, which the research team then regularly shares with other teams and the broader company executives, including the CEO. 

Leaning into their strengths

Leveraging Kapiche for feedback analytics uncovered more than just issues. Nextdoor also quickly learned more about why their customers loved using their app. 

“What neighbors love about Nextdoor, it turns out, is community and information,” says Maryam. “What we saw from using Kapiche was that community is something that people love and it makes them love Nextdoor. That's really motivating for our internal teams,’ she says.