Cancellation, lower reported customer satisfaction, less favorable indicator ratings, and greater intent to leave occurred together across the engagements studied. These relationships describe substantial delivery difficulties, but they do not identify independent causes. Structural attributes provided mixed evidence: some estimated relationships fell within the study's prespecified small-association bounds, others were supported, and others remained inconclusive. The analysis examines these findings alongside engagement maturity and commercial arrangements, distinguishing full-sample cancellation contrasts from estimates restricted to active engagements.
“We are four months into this engagement and customer satisfaction has collapsed. My team corrects client data almost every cycle before it can be used, legal research inputs arrive incomplete and inconsistent, and by the time it reaches the model it is not fit to work from.” -Engagement Manager, KPO, Regulatory, audit, tax and legal
The four core measures were strongly correlated. The engagement indicator mean summarizes respondents' ratings across the 100 engagement indicators, averaged within each engagement. Indicator quintiles are five equal groups ranked by that mean. However, each describes a different aspect of the engagement. The section “The Engagement Indicator Mean and the Interdependence of the Four Measures” examines their shared variance, the pattern across indicator quintiles, and the limits of attributing outcomes to individual indicators.
Decision rights and change control sit at rock bottom here… the joint sign-off process for exception handling takes weeks and breaks cannot wait that long, so the team ends up making the call itself under real time pressure. I have flagged this upward more than once and it has not moved, and with cancellation now confirmed, I am fairly sure it never will.”
ENGAGEMENT MANAGER · BPO, Trade, settlement and reconciliation · “
when i joined the playbook was already written and we follow it step by step, this is the strongest part. besides the QA people do check the agent output properly, i did not expect that. training is the weak part here. i learned the monitoring stack from my lead and from an old runbook, not from any course. so i have marked training low, but the rest i am satisfied with.”
INDIVIDUAL CONTRIBUTOR
ITO, Technology engineering and operations