It is a familiar situation. A quarterly budget review is underway, and collections revenue has come in below expectations. Not dramatically, but enough to raise questions. The collections vendor is on the call. Finance wants an explanation.
Someone suggests citation issuance was down. Enforcement disagrees and points to steady activity levels. Attention turns back to the numbers and eventually to the same question every city asks when revenue falls short:
Why?
Too often, nobody can answer with confidence.
The problem is not a lack of technology or data. Most cities already have access to both. The issue is that reporting alone cannot explain what is happening across a collections program.
Collections vendors are typically hired to recover outstanding citations and report on results. They are not hired to build predictive models, evaluate system performance, or determine what revenue should have been generated under normal conditions. Those are different capabilities, requiring a different perspective.
A collections agency manages accounts. A deep generalist manages the entire system.
Why Collections Vendors Optimize for Their Own Performance
Collections vendors are generally paid based on recoveries. Naturally, that creates an incentive to focus on accounts most likely to pay.
That often means prioritizing newer debt, optimal fine balances, certain violation types, and debtors who respond well to digital outreach. More challenging accounts can receive less attention over time, whether through reduced call activity, delayed notices, or generic communication strategies that are less effective for specific debtor groups.
This is not a criticism of collections agencies. It is simply how incentives work.
The challenge for cities is that local optimization does not always produce the best outcome across the entire portfolio. Revenue may look acceptable overall while performance declines within specific account segments, as easier debt attracts more attention and difficult cohorts drift further from recovery.
Reports show payments collected, aging balances, recovery rates, and contact activity. They do not show whether the current strategy is maximizing the portfolio's potential.
To answer that question, cities need an independent model that establishes expected performance, benchmarks results, and provides an objective measure of success.
Parking Collections Reporting Shows What Happened. Modeling Shows Why.
Reporting is essential, but it is descriptive. It tells you where you ended up.
Modeling estimates what should happen based on historical patterns, debtor behavior, citation characteristics, and operational factors. When performance falls short, it helps identify why.
For parking collections, four capabilities are especially valuable:
Together, these approaches move cities beyond reporting and toward evidence-based decision-making.
Using Cohort Analysis to Improve Parking Collections
Not all citations behave the same way.
A meter citation issued in a busy downtown corridor may have a very different likelihood of payment than an expired registration citation issued to a vehicle with multiple previous violations. Yet many collections programs treat these accounts similarly once they enter recovery.
Cohort-based analysis takes a more realistic approach. Rather than applying a single assumption across an entire portfolio, it groups citations with similar characteristics and tracks how recoverability changes over time.
Techniques such as survival analysis allow cities to estimate the probability that a citation remains unpaid as it ages while considering factors such as violation type, location, seasonality, vehicle history, and payment behavior.
The result is a clearer picture of which accounts are most recoverable, which are becoming harder to collect, and when intervention is most likely to succeed.
This insight helps cities allocate resources more effectively and understand whether collection efforts are aligned with the accounts that matter most.
Why Propensity Modeling Matters in Citation Recovery
One of the most common questions in collections is whether a debtor will pay.
However, a more valuable question is: what intervention is most likely to produce payment, and when should that intervention occur?
Propensity models help answer that question by estimating the probability of recovery under different collection strategies.
For example, some accounts may respond best to text messaging. Others may require traditional notices. Applying the same strategy to every account often results in unnecessary costs and missed revenue. The value increases when cities work with multiple collection vendors.
Instead of treating agency placement as a static process, cities can use propensity modelling to route accounts based on expected outcomes. Different firms may excel with different account types, debt levels, or debtor profiles.
Used effectively, the model helps ensure each account is handled through the channel most likely to generate a recovery.
Understanding Revenue Variance in Parking Collections
When collections revenue declines, several explanations are possible.
Citation volume could be lower. Collection rates may have fallen. The mix of violations could have changed. Debtor behavior may have shifted. Or vendor performance may not be meeting expectations.
Without a performance model, it is difficult to determine which factor is driving the outcome.
Variance attribution solves this problem by breaking performance into its component parts. Instead of relying on assumptions or opinions, cities can quantify how much of a revenue shortfall is linked to issuance levels, debtor characteristics, operational changes, or collections performance.
This turns difficult conversations into productive ones.
Rather than debating whether a vendor is underperforming, cities can quantify the gap, identify the cause, and focus on corrective action.
That level of visibility is difficult to achieve through reporting alone because reporting does not establish an expected baseline. A model does.
Forecasting Parking Collections Revenue More Accurately
Many cities still forecast collections revenue using a simple percentage-based approach. Last year's recovery rate is applied to an estimate of future citation volume, and the result becomes the forecast.
It is straightforward, but it often misses important drivers of performance.
More advanced forecasting methods account for cohort behavior, enforcement activity, economic conditions, seasonal trends, payment patterns, and expected changes in portfolio mix.
Scenario modelling allows cities to understand how revenue may change under different conditions. What happens if enforcement increases by 10 percent? What happens if citation issuance shifts toward higher-value violations? What happens if a new vendor strategy improves recovery rates for aging debt?
A forward-looking model helps answer these questions before the quarter ends rather than after performance has already fallen short.
That gives cities more control over budgeting, vendor management, and long-term program planning.
Why Generic Collections Strategies Underserve Cities
There is another challenge that receives far less attention.
Most collection agencies optimize across their entire book of business, not around the unique characteristics of a single city.
Their staffing models, outreach strategies, and operational processes are designed to work across many clients. While understandable, that approach does not necessarily reflect local conditions.
Collection behavior varies significantly between jurisdictions due to differences in violation types, seasonality, vehicle ownership trends, demographics, and economic conditions.
A strategy that performs well in one city may deliver very different results in another.
Models built around local data are better positioned to reflect those differences. They can identify where recovery efforts are most likely to succeed and support more effective resource allocation.
In many cases, they can also support fairer outcomes by ensuring collection activity is driven by evidence rather than broad assumptions.
Closing the Parking Collections Intelligence Gap
The techniques discussed here are not new. Survival analysis, propensity modelling, variance decomposition, and scenario forecasting have been used for years in industries such as finance, insurance, and healthcare.
The opportunity is not adopting new methodologies but applying proven ones to parking collections.
The cities achieving the best outcomes are not necessarily those working with the largest collection agencies, but those with the clearest understanding of their portfolios and the ability to measure performance independently.
That is where a deep generalist adds value.
Rather than replacing collection agencies, a deep generalist sits between the portfolio and its vendors, building models, evaluating performance, designing strategies, managing multi-firm waterfalls, and preserving institutional knowledge.
Most importantly, they help cities move from reacting to results to understanding what results should be.
At Trellint, we help cities close the gap between reporting and insight through independent analysis, collections expertise, and data-driven decision support that improves revenue and accountability across the collections lifecycle. Contact us here if you would like to explore this topic more and a member of our team will reach out.