The Challenge
A UK-based HR software provider serving mid-market businesses on annual contracts wanted to bring structure to a renewal process that relied heavily on instinct and last-minute effort.
Their customer success team managed every account with the same formula, with no visibility of actual renewal likelihood. Warning signs surfaced late, often within 30 days of contract expiry, leaving little room for anything beyond a defensive discount to salvage the deal. Margin was being routinely traded away to save accounts that a timelier value-led conversation could have kept at full price.
The client wanted a predictive framework to isolate at-risk accounts early, allowing the renewal team to concentrate effort and budget where they could change the commercial outcome.
The Solution
Consolidating the client’s siloed account data was the critical first step, bringing together product usage logs, support ticket history, billing records, and CRM activity under a single account-level identifier.
We then built a propensity model to evaluate every account’s renewal likelihood, which was refreshed monthly and surfaced 90 days ahead of each contract date. Declining seat activity, unresolved support tickets, and the departure of the account’s original internal sponsor emerged as the primary churn signals, carrying far more weight than account size or tenure.
These health scores were deployed into the client’s CRM, feeding a practical three-tier approach: safe accounts moved to light-touch automated outreach, mid-scoring accounts received standard milestone check-ins, and high-risk accounts triggered an immediate intervention by a senior customer success manager.
The Impact
This 90-day warning window fundamentally shifted renewal conversations from defensive to proactive. Reaching at-risk accounts to resolve underlying product adoption issues before contract expiration lifted the renewal rate by 6 percentage points, protecting an estimated £380k in annual recurring revenue.
The strategy also delivered vital margin protection, cutting unplanned discounting by a third and replacing eleventh-hour panic responses with strategic, planned concessions.
The model now operates as a standard monthly workflow, embedding the data-driven three-tier prioritisation directly into the customer success team’s quarterly planning cycles.
Key Takeaways
- Renewal Propensity Modelling - Scoring account-level renewal likelihood 90 days ahead of contract expiry provided a reliable early-warning window.
- Early Warning Signals - Seat activity, support tickets, and sponsor changes predicted risk far better than account size or tenure.
- Tiered Prioritisation - Automated CRM routing channelled customer success resources toward high-risk accounts where intervention moved the needle.
- Commercial Results - Renewal rates improved by 6 points, protecting £380k in recurring revenue and reducing margin-eroding panic discounts by a third.
Tools and Techniques
- Propensity modelling
- Exploratory data analysis - profiling usage logs, ticket histories, and CRM inputs
- K-Nearest Neighbours classification (KNN) - statistical classification for risk tiering
- SQL - data extraction, cleansing and transformation
- R - data handling and predictive modelling
- HubSpot - operational CRM integration and activation
Related services
No matching items