Utah Transit Authority | TVM Failure Sequences & Service Recovery

Executive Summary

Failed purchase sessions reached 44.1K in the latest month while vendor-alert coverage surfaced only 25.3K, leaving a persistent gap between what riders experienced and what support teams could see. Session reconstruction across 250+ TVMs shows one timeout-heavy path accounts for 31% of failed attempts, with the steepest drop occurring after payment initiation. The highest-value opportunity is not fare-policy modeling; it is service recovery: classifying the repeat failure paths, reconciling them against vendor alerting, and prioritizing the fixes most likely to recover revenue and rider confidence inside a $90M+ annual fare context.

Latest Alerting Gap

42.6%

of failed sessions not surfaced by vendor alerting

Slice the data

Machine Group

Failure Class

Window

Section 1

Failure Signal & Monitoring Gap

Failed Purchase Sessions

44.1K

↑ 12.8% YoY

Vendor Alerts Captured

25.3K

↑ 14.0% YoY

Uncaptured Failure Gap

18.8K

42.6% of failures unseen

Recoverable Leakage

$0.69M

Annualized from top failure classes

Dominant Failure Path

31%

Timeout-heavy session sequence

Figure 1

Failed Sessions vs. Captured Alerts — 36-Month Trend

The monitoring gap between rider-visible failures and vendor visibility

Failure volume and alert volume both rise over time, but the unresolved gap persists into late 2024. That gap is where the rider experience degrades before support teams can respond.

Section 2

User-Flow Sequence Analysis

Figure 2

Top Failure Paths by Share of Failed Sessions

The sequence classes most responsible for rider abandonment

Failure Path	Share	Primary Issue
Timeout after payment initiation	31%	Session expires before confirmation returns
Card-reader retry loop	22%	Repeated auth attempts without ticket issuance
Ticket issuance interruption	15%	Payment accepted but ticket or receipt flow breaks
Cash acceptor incomplete close	11%	Physical device or change-return failure

Figure 3

Completion by User-Flow Step

Healthy sessions vs. the two most common failure patterns

The sharpest fall-off happens after payment initiation, where timeout-heavy and retry-loop sessions start diverging materially from healthy completion.

STAR Case Study

Want the interview-ready UTA version of this story?

This dashboard shows the operational model. The companion case study turns it into a clean STAR story: repeated ticket-machine failures, the user-flow and alert-gap analysis behind them, and the prioritized remediation work tied to about $0.69M in annualized recoverable leakage.

Read the STAR case study Browse all four stories