I talk with peers inside the customer regularly

The mean rating was 5.5 and the median 6.0 (n = 809 respondents). Ratings of 3 or below, the distressed range, made up 5.93% of the answers. The rating was not associated with cancellation (r = 0.00, 95% CI −0.16 to 0.16, p = 0.957, n = 150 engagements), customer satisfaction among active engagements (r = −0.06, 95% CI −0.26 to 0.14, p = 0.558, n = 94 engagements), the engagement indicator mean (r = −0.08, 95% CI −0.23 to 0.09, p = 0.355, n = 150 engagements) and intent to leave (r = 0.01, cluster bootstrap 95% CI −0.05 to 0.07, p = 0.812 with standard errors clustered by engagement, n = 760 respondents). Each of these associations was equivalent to zero. No segment family differed after Holm correction.

Most of the 48 written responses came from C114, the comment field on the financial institution's organization. Across 27 responses, respondents said the financial institution's people answered quickly or sat with the team, and placed the difficulty in access, data, sign-off or ownership. The eight responses with the strongest negative score came from respondents who never met anyone on the financial institution's side. One of the four written outcome tests reached significance.

Ratings of 3 or below, the distressed range, accounted for 48 of the 809 answers (5.93%). The engagement mean fell in the distressed range in 0 of 150 engagements.

All 48 responses were secondary. Each answered a question outside the indicator's survey section and addressed the indicator in its content. They came from 3 questions. That secondary share ranked 1st from the highest among the 11 Part 4 indicators with 30 or more written responses, tied with seven other indicators, against a median of 100.00%.

Across 27 responses, respondents said the financial institution's people answered the same day, sat with the team, stayed on calls late or made time, and placed the difficulty elsewhere: access requests that took weeks, data that disagreed across systems, a change board that met once a month, or no one who owned the AI work and could approve it. Six described a standing call or a direct chat with staff on the financial institution's side, used to get answers that the formal process did not give, and one said nothing from those chats was logged.

Eight responses, each with the strongest negative score, came from respondents who never met or spoke with anyone on the financial institution's side, in one case over eight months, and several said they could not rate people they had never met. Six responses said the financial institution was two groups, such as the staff who used the tool and the team that owned the systems, and that one rating could not describe both. One respondent with long experience in the industry said the team now talked with the financial institution every day, more than on any earlier project, mostly about basic operational problems, and that the financial institution deferred most decisions.

Of the 48 responses, 9 (18.75%, 95% CI 8.70% to 30.95%) were negative and 8 (16.67%) carried the strongest negative score of −10. Positive responses numbered 38 (79.17%), and 1 was neutral. The mean written score was 2.7 (95% CI 0.8 to 4.3, n = 48 responses), and the median was 5.0.

Of the 48 responses, 46 came from respondents who also rated the indicator. Among them, the written score did not track the rating (Spearman rho 0.06, 95% CI −0.29 to 0.41, p = 0.672, n = 46 responses). Of the 8 negative responses with a rating, 1 came from respondents whose rating was in the distressed range.

Respondents who wrote about the indicator and gave a rating numbered 45, with a mean rating of 5.5. The 764 respondents who rated the indicator without writing about it had a mean of 5.5. The difference, 0.0 (95% CI −0.4 to 0.3), included zero. Ratings in the distressed range made up 4.44% of the ratings from respondents who wrote (2 of 45) and 6.02% of the ratings from respondents who did not (46 of 764).

Every figure in the segment tables is at respondent level. Segments with fewer than 30 respondents carry no interval. Segments with fewer than 10 report counts only.

Segment families recorded at engagement level were tested on engagement means. Respondent role and vendor type were tested on respondents. Where an expected count fell below 5, the test of the distressed share used Fisher's exact test for a two-by-two table, marked exact, or a Monte Carlo test, marked simulated. No family differed after Holm correction.

Written sentiment did not differ by engagement status and lifecycle stage across the 37 engagements with a written response (Kruskal-Wallis H 6.50, df 3, p = 0.090). Of the groups with 10 or more responses, responses from canceled engagements were the most negative (mean −0.3, 16 responses) and responses from active engagements in Transition/Cut-Over the least negative (mean 4.2, 18 responses). Responses from active engagements in Build-Out (6) and active engagements in Production (8) are reported as counts only.

No test across respondent roles was run, because fewer than two roles had 30 or more written responses. Individual contributors wrote 35, engagement managers wrote 10 and portfolio directors wrote 3 responses.

Cancellation rate is the share of engagements canceled by the study date, measured at engagement level (150 engagements, 56 canceled). Canceled engagements were rated after cancellation. The association is therefore a descriptive contrast and not a forecast of cancellation.

Across the study set, the rating was not associated with cancellation (r = 0.00, 95% CI −0.16 to 0.16, Holm p = 1.000, n = 150 engagements). The association held after Holm correction in zero of the 20 segment levels tested.

Customer satisfaction is the percentage recorded against each engagement. The primary estimate uses the 94 active engagements. A descriptive contrast uses all 150 engagements.

Across the active engagements, the rating was not associated with customer satisfaction (r = −0.06, 95% CI −0.26 to 0.14, Holm p = 1.000, n = 94 engagements). The association held after Holm correction in zero of the 16 segment levels tested.

The engagement indicator mean is the mean of an engagement's ratings across the rated indicators, measured at engagement level (150 engagements). The indicator does not enter the mean.

The rating was not associated with the engagement indicator mean (r = −0.08, 95% CI −0.23 to 0.09, Holm p = 1.000, n = 150 engagements). The association held after Holm correction in zero of the 24 segment levels tested.

Intent to leave is the answer to C019, “How likely are you to leave your current job in the next 12 months?”, rated 1 to 10. It is measured at respondent level on up to 798 respondents, with the engagement-level form reported alongside.

The rating was not associated with intent to leave (r = 0.01, 95% CI −0.06 to 0.08, Holm p = 1.000, n = 760 respondents). The association held after Holm correction in zero of the 37 segment levels tested.

Two comparisons set the engagements with at least one negative response against all other engagements. Of the 141 other engagements, 113 had no written response addressing the indicator. These two comparisons therefore do not isolate written sentiment. Of the 9 engagements with a negative response, 6 were canceled, against 50 of the 141 other engagements (35.46%). The difference did not reach significance (Fisher exact p = 0.079). The comparison of customer satisfaction was not computed, because one of the two groups held too few active engagements.

Two correlations test written sentiment directly. Across the 37 engagements with a written response, the engagement mean written score was not associated with cancellation (point-biserial r −0.32, p = 0.052). Across 43 respondents, a more negative respondent mean written score came with higher intent to leave (Spearman rho −0.35, p = 0.021). Of the three comparisons, one reached significance at 0.05.

The four associations were compared in pairs on their engagement-level forms, each oriented so that a higher value is the better outcome, across the 150 engagements with all four measures. After Holm correction, zero of the six pairs differed.

Answers clustered within engagements (ICC −0.01). Clustering reduced the 809 respondent answers to the equivalent of about 809 independent answers. Respondent-level intervals therefore used the cluster bootstrap. Respondent-level tests of intent to leave used standard errors that allow for clustering by engagement.

The association with cancellation, customer satisfaction, the engagement indicator mean and intent to leave were equivalent to zero. The detection floor shown is the study floor at 150 engagements. Intent to leave was tested on its effective sample size.

Each change was computed before rounding and may differ by 0.01 from the difference between the two printed values.

All 22 checks held. A check holds when its test falls on the same side of 0.05 as the primary test. Where the primary test falls below 0.05, the check must also keep the sign of the primary correlation. The thin-base check removes the engagements in any level of outsourcing model or process family with fewer than 30 engagements. Institution type is not used for that check, because every one of its levels has fewer than 30 engagements.

Each of the four measures was computed from data the indicator does not contain. Each result is therefore a finding and not an arithmetic consequence.

Each deviation was computed before rounding and may differ by 0.01 from the difference between the two printed values.

The comparison covered 150 engagements. The indicator departed from the study's typical pattern on all four measures. Its association was weaker than the norm on cancellation, customer satisfaction, the engagement indicator mean and intent to leave.

Among the 11 Part 4 indicators with 30 or more written responses, the indicator's share of negative responses, 18.75%, ranked 7th from the highest, against a median of 33.33%. Its mean written score, 2.7, ranked 8th from the most negative, against a median of 0.8.

The indicator most often present in the same responses was C109, “Rate how quickly the customer responds when this engagement raises issues that need customer action.”, in 23 of the 48 responses (47.92%). C106, “Rate the availability of the customer's subject matter experts when this engagement needs them.” (20 responses), C072, “I talk with engagement management inside the customer regularly.” (10 responses), C073, “I talk with the customer's leadership team regularly.” (10 responses) and C094, “Rate the quality of the customer's data across the systems this engagement depends on.” (10 responses) followed.

The most frequent theme was “The scale or question does not fit”, in 8 responses (16.67%). The second was “The rated entity is not one entity”, in 6 responses.

Fewer than one in ten ratings of regular contact with peers inside the financial institution fell in the distressed range (48 of 809 answers, 5.93%). The rating was not associated with any of the four study measures, and each association was equivalent to zero. Canceled engagements rated about the same as the three active stages (mean 5.5, n = 305 respondents).

The written evidence is consistent with that result. Most respondents who wrote had working contact with the financial institution's staff and placed the difficulty in access, data, sign-off or ownership, matters that contact at peer level could not resolve. Respondents whose written responses were more negative reported higher intent to leave, and the other three written outcome tests did not reach significance.

The evidence is the reported experience of the vendor's delivery teams, collected in a single cross-section. The associations reported here are not causal.

Tests by respondent role and by vendor type treated respondents as independent. The section on clustering reports how far clustering within engagements reduced the information in respondent-level figures.

Customer management executives gave 1 answers, a base of anecdote strength.

Shares of written sentiment are shares of responses and not of the study set of respondents. The written score takes five fixed values, so its mean and median are coarse. Responses clustered within engagements (48 responses on 37 engagements). Intervals for written sentiment therefore used a bootstrap by engagement.