For decades, collections programs have been built on a quiet but powerful assumption:
collection vendors are the experts, and they know best how to work the debt.
Under this model, clients assign delinquent accounts, monitor high-level results, and intervene only when revenue declines become obvious. Vendors are trusted to optimize outreach, staffing, and strategy based on their experience.
It’s a reasonable assumption. But it’s also incomplete. This case study shows how data science revealed the limits of collection agency deference, including the natural risk of cherry-picking debt. By shifting to active performance management, Trellint unlocked $4.05 million in measurable revenue improvement for our client.
The Hidden Incentive Problem
Most collection vendors are paid based on recoveries. That creates a rational incentive: prioritize accounts that are easiest to collect. Vendors naturally focus on their return-on-investment, concentrating time and energy on newer debt, higher balance accounts, and accounts that respond to digital outreach.
Meanwhile, harder-to-collect accounts may receive less consistent effort, fewer calls, or delayed notice activity. Over time, this creates uneven performance across the portfolio, even if overall revenue appears acceptable in the short term.
The problem is not vendor intent. The problem is lack of visibility and failure to communicate expectations.
What Data Science Made Visible
A centralized, performance-managed approach to delinquent accounts and past-due balances requires both a detailed understanding of collections and deep analysis of data.
Therefore, Trellint data scientists studied the performance of firms, looking beyond revenue alone and developing leading indicators of effort and behavior, including outbound call volumes, channel mix, outreach, and consistency.
Our data review showed that effort was not always aligned with portfolio needs.
Cherry-Picking Without Malice
In some cases, outbound calling had quietly declined by double digits at certain vendors, even as vendors expressed confidence in their strategies. Text messaging was being overused because it was efficient. Noticing was inconsistent from month to month.
None of this necessarily indicated negligence. It reflected local optimization: vendors improving their own workflows and economics in the absence of portfolio-level guidance. They focused on easier debt and efficient outreach.
A New Management Theory: Optimize the Portfolio, Not Just the Vendor
The insight that followed was critical: Vendor expertise optimizes locally. Active management optimizes globally.
Collection agencies and law firms are very good at optimizing within their own four walls. They make rational decisions based on their staffing models, cost structures, tech stack, and internal KPIs. From a vendor’s perspective, “doing a good job” often means focusing effort where returns are highest, leaning into scalable channels, prioritizing accounts that convert quickly, and controlling costs.
However, what’s optimal locally is not always optimal for the entire, global portfolio.
When vendors optimize independently, they risk focusing on “easier debt”, neglecting collection efforts on older accounts and prioritizing efficiency over effectiveness. Aggregate revenue may look acceptable, but it often hides performance gaps.
Active Management Optimizes Globally
You can’t balance effort, detect drift, or counter cherry-picking if you only look at monthly revenue, self-reported activity, or individual vendor dashboards. To be successful, collections management must be transparent and connect behaviors to outcomes.
Trellint data scientists applied centralized performance management to gain insights across the entire portfolio, including all account segments, vendors, and stages of delinquency. We studied vendor efforts, activities, balances, and leading indicators for work drift and revenue lags.
This global optimization approach provided insights into the health and performance of the whole portfolio, not just its individual parts or vendors. Data science allows us to reframe vendor management from deference to accountable partnerships and deliver greater results for our client.
Turning Insight Into Action
When data revealed sustained declines in outbound calling, an early indicator that portions of the portfolio were being underworked, we engaged collaboratively. The Trellint team revisited outreach strategies, clarified expectations about minimum activity levels, and rebalanced the channel mix.
Outbound calls shifted from decline to double-digit growth without increasing inventory, fees, or enforcement scope.
The vendors didn’t change. The incentives didn’t change. The management lens changed.
Ensuring Harder Accounts Aren’t Left Behind
The Trellint team also found that outreach volumes were stagnant despite stable account volumes, suggesting that effort was being concentrated rather than distributed. By identifying internal volume caps and aligning staffing and dialing strategies with the full portfolio, outreach activity more than doubled.
Balancing Efficiency With Effectiveness
Data science also challenged the assumption that automation alone improves outcomes. Some vendors leaned heavily into text messaging because it scaled efficiently, but data showed that certain segments required voice outreach to remain productive.
By actively managing channel mix, outbound calls increased from a few thousand calls per month to more than 250,000 calls per month, improving engagement depth and supporting the program’s cumulative improvement.
Consistency as an Antidote to Drift
The same principles apply to notice-based outreach. Without active oversight, variability crept in. By treating noticing as a managed process, with clear expectations and monitoring, average monthly notices increased from approximately 1,900 to over 3,100. This reduced variability and strengthened downstream collections.
The Real Shift
The real change wasn’t vendor competence or commitment, but the belief that expertise alone could ensure fair and effective effort across the portfolio.
Data science showed how rational incentives can sometimes lead to uneven outcomes. Active management corrected for that reality, not by replacing vendors, but by guiding effort intentionally.
This shift unlocked $4.05 million in measurable revenue improvement in 2025 and provides the foundation for further growth in 2026.
Why This Matters
Any program that relies on third-party execution will face the same challenge: without oversight, effort gravitates to what is easiest, not what is optimal.
Data science doesn’t replace expertise; it augments it. It keeps collection vendors honest, balanced, and aligned.
And when active management is driven by data, revenue follows.
Data-Driven Stewardship
Cities are under constant pressure to do more with limited resources without eroding public trust. Improving revenue doesn’t require harsher enforcement or higher fines; it requires managing existing programs more intentionally.
By applying data science and active performance management to delinquent accounts, cities can uncover revenue that is already owed, already authorized, and already in the system but not fully realized. The difference is visibility, balance, and governance.
If your city is looking to strengthen financial stewardship while maintaining fairness and accountability, contact us today to discuss how this can be applied to your program.