B2B SAAS CUSTOMER SEGMENTATION CASE STUDY

How Wave Used Customer Segmentation
to Explain a Falling North Star Metric and Uncover a New Growth Segment

Wave’s tax-filing enrollment had fallen from a long-standing 90% baseline to 78%, and no team could explain why. Customer segmentation revealed that the metric wasn’t broken. Wave’s customer mix had changed.

Wave’s payroll team had roughly 9,000 paying users, and enrollment in automated tax filing had historically been one of its strongest retention signals. When enrollment began falling, marketing, product, engineering, and tax operations could not explain the change.

HPI interviewed five internal stakeholders and 15 customers, analyzed behavioral and revenue data, and tested seven segmentation hypotheses. The analysis identified nine customer segments and revealed that two of the fastest-growing groups were contractor-only businesses that had little need for automated tax filing.

Once those customers were separated from the analysis, enrollment among the remaining customer base was steady. The declining North Star metric was not evidence that the core retention mechanism had stopped working. It was evidence that Wave was attracting a different kind of customer.

Background

Wave was a software company generating approximately $100 million in revenue across three products: accounting, invoicing, and payroll. This B2B SaaS customer segmentation case study focused on the payroll product, which had roughly 9,000 paying users.

One of the payroll team’s most important North Star metrics was enrollment in automated tax filing and payments. The reason was practical: customers enrolled in that service were about seven times more likely to retain than customers who were not.

For years, enrollment held near 90%. Then it fell to roughly 85%, then 80%, and eventually 78%. Because the metric had a strong historical relationship with retention, the decline looked like a serious product and growth problem.

The Challenge

A critical metric was falling, but the aggregate data could not explain why

Wave had already looked for the obvious causes. Marketing could not explain the change through acquisition. Product could not tie it to a clear experience problem. Engineering found no bug. The tax operations team found no operational explanation. The dashboard showed what was happening, but not what had changed underneath it.

The North Star metric had become an assumption
Because automated tax filing had been such a strong retention signal, the organization treated higher enrollment as universally desirable. That assumption made sense when most customers had similar payroll needs. It became less useful as the customer mix changed.

The existing ICP was too broad to diagnose the problem
“Small business” described the market, but it did not distinguish how different businesses actually paid people. A company paying salaried employees, one paying hourly staff, and one paying only contractors could all look similar at the firmographic level while needing very different payroll workflows.

Wave needed a causal explanation before changing the product
Had customers stopped valuing the tax service, or was Wave attracting more customers for whom that feature was never important?

The purpose of the study was to identify which customer differences actually explained product behavior, then use that evidence to determine whether the metric, product, pricing, or go-to-market strategy needed to change.

The Approach

Test the customer model against qualitative, behavioral, and revenue evidence

HPI did not start by inventing personas. The team gathered customer and internal evidence, compared it with product usage and revenue data, and tested multiple ways of segmenting the customer base until a model emerged that explained the behavior Wave was seeing.

  1. Start with internal evidence:
    HPI spoke with five internal stakeholders to collect the stories, hypotheses, and operating context behind the declining North Star metric.
  2. Add the customer voice:
    HPI conducted 15 customer interviews to understand how different businesses ran payroll, which jobs they needed Wave to perform, and which parts of the product mattered in practice.
  3. Connect the conversations to behavior and revenue:
    The team extracted and cleaned product and revenue data, then compared what customers said with the features they actually used and the value of those accounts.
  4. Test competing segmentation hypotheses:
    HPI evaluated seven different ways the customer base could be segmented. Two dimensions consistently explained the behavior best: the type of workers a business paid and the number of workers it paid.
  5. Build the nine-segment model:
    Four worker-type groups – salaried only, hourly only, mixed worker types, and contractors only – were split by smaller and larger worker counts, with owner-only businesses forming a ninth segment. The result was a segmentation model tied directly to payroll behavior rather than company descriptors alone.

What Changed

  • Two contractor-only segments were growing exceptionally fast – approximately 200% and 300% year over year.
  • Those businesses did not need the automated employee tax-filing workflow that sat at the center of Wave’s North Star metric. Their core job was simpler: pay contractors accurately, with less need for the tax automation built for employee payroll.
  • That meant the overall tax-enrollment percentage was being pulled down by a growing group of customers for whom enrollment was not a meaningful success signal.
  • When the contractor-only customers were separated from the analysis, tax-filing enrollment among the remaining customer base was steady. The retention mechanism had not suddenly failed. The composition of the customer base had changed.
  • The product opportunity became specific: a slimmer contractor-focused payroll experience, priced for a lower-complexity use case instead of bundling customers into a full employee-payroll product.
  • The workflow opportunity was equally concrete: contractor-paying businesses needed more flexibility to pay people when work was completed instead of being locked into a fixed biweekly payroll schedule. The segmentation gave product and marketing a clearly defined customer, job, feature set, and buying story to evaluate.
The Result:

The study turned an unexplained metric decline into a specific customer and product decision

In six weeks, HPI turned scattered customer signals into a prioritized segmentation model and an execution plan. The work clarified which customers Wave should focus on, which product moments deserved greater emphasis, and how the team could test the strategy with minimal engineering lift.






Findings
Evidence
Why it mattered






North Star explained
Tax enrollment fell from ~90% to 78% overall, but was steady when contractor-only accounts were separated.
Distinguished a customer-mix shift from a product or retention failure.




Customer model
9 segments based on worker type and worker count, selected after testing 7 segmentation hypotheses.
Replaced a broad small-business ICP with a model tied to real payroll behavior.




Growth segment
Two contractor-only segments were growing at approximately 200% and 300% YoY.
Made the emerging product and marketing opportunity visible.




Product direction
Contractor customers needed simple payments and flexible timing more than full employee tax automation.
Defined what a contractor-focused offer would need to include – and what it could leave out.




Modeled upside
Active paying-user growth modeled from ~10% to ~21%.
Suggested roughly double the user growth rate if the contractor opportunity were executed; this was modeled, not achieved.




Commercial tradeoff
A slimmer contractor offer would likely carry a lower average contract value.
Made the growth scenario more realistic by separating user growth from revenue growth.

The value was not the number of segments. It was the ability to make clearer choices. Wave could see who to prioritize, which product moments created the strongest retention signal, what to test first, and how to measure whether the plan was working.

This is the purpose of an HPI Customer Segmentation Study™ : combine customer interviews and behavioral evidence so leaders can replace broad assumptions with a decision roadmap.

Frequently Asked Questions

Your dashboards may show what changed. A segmentation study can explain why.

See how HPI combines customer interviews, behavioral data, and customer cohort analysis to identify the customers and product moments that deserve priority.
Start a Conversation

Customer evidence. Market clarity. Revenue decisions you can defend.