Case Study — 01
Fixing fragmented data and saving $9M in ARR — by convincing leadership to rebuild the infrastructure
At Vendasta — a B2B SaaS platform with $93M revenue and 10M+ active users — churn was rising and no one could explain why. The answer was buried in six disconnected data systems. I found the gap, built the case, and led the platform rebuild that changed it.
Context
A $93M platform where nobody could see what was happening
Vendasta is a B2B SaaS company that helps digital agencies manage local business clients. By the time I joined as Senior PM, the platform had grown to serve 10M+ active users — but that growth had come at a cost. Customer data was fragmented across six separate systems, with no coherent view of how accounts were actually performing.
The immediate symptom was churn. Accounts were cancelling at rates that made the sales team nervous and the finance team alarmed. But nobody had a clear model of why — because nobody had a clear picture of what was happening to customers after they signed up.
Problem
The real issue wasn't churn. It was visibility.
Most of the organization was treating churn as a sales or success problem. More follow-up calls, more check-ins, more QBRs. I spent time with the data team trying to understand what we actually knew about customers — and the picture was worse than expected.
Usage data lived in one system. Billing data in another. Support tickets in a third. Feature adoption and campaign performance each had their own home. When a customer success manager wanted to understand whether an account was healthy, they had to manually piece together data from multiple tools — and most of the time, they didn't bother.
"The problem wasn't that we didn't care about customers. It was that our infrastructure made it structurally impossible to see them."
This meant we had no early warning system for at-risk accounts. Churn was being diagnosed at the point of cancellation, not 60 days before it. And without a unified data layer, any attempt to build proactive retention tooling would be built on sand.
Approach
Building the business case, then building the platform
The infrastructure rebuild I had in mind wasn't small. It would require engineering resources, cross-team coordination, and a pause on some feature work. Leadership needed to be convinced that this was the right investment — and that meant making the cost of inaction concrete.
Quantified the gap
Worked with the data team to model what a "data-complete" view of customer health would look like vs. what we had. Estimated the percentage of at-risk accounts we were identifying too late to act on. Tied it to ARR numbers that leadership could react to.
Mapped the architecture needed
Collaborated with engineering to design a unified customer data layer — one source of truth for usage, billing, support, and engagement data. Scoped a phased approach that could deliver early signal while the full platform was being built.
Aligned stakeholders across CS, Data, and Finance
Customer success, data engineering, finance, and marketing all had a stake in how customer data was structured. Spent significant time ensuring the reporting schema served all of them — not just the version that was easiest to build first.
Launched the reporting and analytics platform
Shipped the unified analytics layer with dashboards for CS managers, health score models for account teams, and automated early-warning signals. Ran a structured rollout: pilot with 200 accounts, measure, adjust, then full release.
Followed on with financial reporting infrastructure
Extended the platform to include structured financial reporting via SQL-based ETL pipelines into NetSuite. This alone saved $450K annually in manual reporting overhead and reconciliation errors.
Outcome
89% churn reduction. $9M ARR retained.
Within the first year of the analytics platform launch, churn dropped by 89%. The mechanism was straightforward: CS teams could now see at-risk accounts 60–90 days before cancellation and intervene. The signal existed before — it was just scattered across six systems that nobody had the time to manually correlate.
The $9M ARR figure represents the delta between projected churn under the old model and actual churn post-platform. It doesn't include the compounding effects of renewals, upsells enabled by better account visibility, or the downstream impact on CAC from improved retention metrics.
The NetSuite financial reporting system saved an additional $450K annually. Combined with other platform efficiency initiatives, total annual savings from the infrastructure work exceeded $9.45M.
What I Learned
The hardest part of infrastructure work is making it legible to the business.
Infrastructure projects die because they feel abstract until they're done. The only way to sustain organizational support through a multi-quarter rebuild is to translate the technical problem into financial terms early — and then keep that translation current as the project evolves.
The second lesson: unified data isn't just a technical problem, it's a political one. Every team that owns a data silo has reasons — historical, organizational, sometimes contractual — for why it exists. The rebuild required as much stakeholder navigation as engineering work.