The engagement is not achieving stated KPIs

Distressed answers on whether the engagement was achieving its stated KPIs came from a minority of answers (241 of 782 answers, 30.82%) and a minority of engagements (32 of 150 engagements). On this item a higher figure marks an engagement closer to its stated KPIs (each figure is 11 minus the raw answer). Engagements rated further from their stated KPIs were canceled more often (r = −0.59, 95% CI −0.69 to −0.48, p < 0.001, n = 150 engagements). Engagements rated closer to their stated KPIs showed higher customer satisfaction on active engagements (r = 0.62, 95% CI 0.48 to 0.73, p < 0.001, n = 94 active engagements) and a higher engagement indicator mean (r = 0.79, 95% CI 0.72 to 0.84, p < 0.001, n = 150 engagements). Respondents who rated the engagement further from its stated KPIs reported higher intent to leave (r = −0.63, cluster bootstrap 95% CI −0.69 to −0.56, p < 0.001, n = 734 respondents). Ratings differed after Holm control by engagement status and lifecycle stage (Welch F 29.64, Holm p < 0.001, omega squared 0.338), by respondent role (Welch F 7.11, Holm p 0.008, omega squared 0.010), and by vendor type (Welch F 4.47, Holm p 0.004, omega squared 0.023).

The indicator had 14 written responses coded to it, from 13 of 850 respondents on 12 of 150 engagements. Negative responses numbered 12 of the 14 scored responses (85.71%). The strongest negative score, −10, was carried by 5 of the 14 scored responses (35.71%). Secondary responses, written in answer to questions outside the indicator’s own survey section, numbered 10 of the 14 responses (71.43%). Respondents who wrote rated the indicator lower on the goodness view than respondents who did not (mean 4.3, n = 12 respondents, against mean 4.8, n = 770 respondents), a difference reported with no interval, because fewer than 30 respondents who wrote gave a rating.

The study set mean was 4.8 on 782 answering respondents, with a median of 5.0. 241 answers, 30.82% of the 782, were at the distressed end. A distressed engagement mean was recorded in 32 of 150 engagements.

Of the 14 responses, 4 (28.57%) were primary and 10 (71.43%) were secondary. The secondary responses came from 10 respondents on 10 engagements, in answer to 4 questions outside the indicator’s own survey section (C029 with 6 responses, C030 with 2 responses, C135 with 1 response and C156 with 1 response). One respondent wrote both a primary and a secondary response.

Negative responses made up 8 of the 10 secondary responses (80.00%), against all 4 primary responses. The difference did not reach significance at 0.05 (Fisher exact p = 1.000, n = 14 responses).

Of the 14 scored responses, 12 (85.71%) were negative, 5 (35.71%) carried the strongest negative score and 2 (14.29%) were positive. The responses had a mean written score of −5.4 (n = 14 responses on 12 engagements). The base of scored responses was under 30, so no interval was reported for the mean.

Respondents whose rating stood at 3 or below on the goodness view wrote 7 of the 11 negative responses with a rating. Written score and rating showed a strong positive correlation that reached significance at 0.05 (Spearman rho 0.65, p = 0.017, n = 13 responses on 11 engagements). The base of responses with a rating was under 30, so no interval was reported for the correlation. Where a written score diverges from the same respondent’s rating, the rating is the measure of the indicator.

Respondents who wrote about the indicator rated it 0.5 points lower on the goodness view than respondents who did not (mean 4.3, n = 12 respondents, against mean 4.8, n = 770 respondents). The base of respondents who wrote was under 30, so no interval was reported for the difference. Ratings of 3 or below on the goodness view came from 6 of the 12 respondents who wrote and gave a rating (50.00%), against 235 of the 770 respondents who did not write and gave a rating (30.52%). Respondents who wrote formed a self-selected group and not a random sample of the answering base, because written responses were voluntary.

Figures are respondent level. Means are on the goodness view. A level under 30 prints no interval and a level under 10 prints counts only.

Engagement families were tested on engagement means, and respondent role and vendor type were tested on respondents. A distressed p marked exact is a Fisher exact test for a two by two table, and one marked simulated is a Monte Carlo test run until its p is stable to the third decimal, both used where an expected count is under 5. Engagement status and lifecycle stage differed after Holm control (Welch F 29.64, Holm p < 0.001, tested on engagement means). In Descriptive Statistics by Segment, Active, Production had the highest mean, 5.8 (n = 93 respondents), and Canceled the lowest, 3.4 (n = 294 respondents). Respondent role differed after Holm control (Welch F 7.11, Holm p 0.008, tested on respondents). In Descriptive Statistics by Segment, Portfolio Director had the highest mean, 5.8 (n = 24 respondents), and Individual Contributor the lowest, 4.7 (n = 608 respondents). Vendor type differed after Holm control (Welch F 4.47, Holm p 0.004, tested on respondents). In Descriptive Statistics by Segment, Consulting firm had the highest mean, 5.3 (n = 112 respondents), and BPO specialist the lowest, 4.2 (n = 179 respondents). No difference held after Holm control for outsourcing model, institution type, process family and class of work.

Scored responses came from 8 engagements in the Active, Transition/Cut-Over group, 1 engagement in the Active, Production group and 3 engagements in the Canceled group. No scored response came from the Active, Build-Out group. A test of mean written scores across the groups, with each engagement counted once, showed no difference that reached significance at 0.05 (Kruskal-Wallis H 2.95, df 1, p = 0.086, n = 11 engagements). The test compared the Active, Transition/Cut-Over group and the Canceled group. It left out the Active, Production group, which held fewer than two engagements. Groups under 10 responses are reported as counts only. Negative responses numbered 6 of 8 in the Active, Transition/Cut-Over group, 2 of 2 in the Active, Production group and 4 of 4 in the Canceled group.

Across respondent roles, no test was run, because fewer than two roles reached 30 scored responses. Roles under 10 responses are reported as counts only. Among individual contributors, negative responses numbered 7 of 9. Among engagement managers, negative responses numbered 3 of 3. Among portfolio directors, negative responses numbered 2 of 2.

The share of engagements canceled on the study date, measured at engagement level (base 150 engagements, 56 canceled). The association is a descriptive contrast, inflated because canceled engagements were rated after cancellation, and is not predictive.

The study set r was −0.59 (95% CI −0.69 to −0.48, Holm p < 0.001, n = 150 engagements). Of the 20 tested segment levels, 12 held after Holm control.

The satisfaction score recorded against each engagement on a 100-point scale. The primary estimate uses the 94 active engagements. A descriptive contrast uses all 150 engagements.

The study set r was 0.62 (95% CI 0.48 to 0.73, Holm p < 0.001, n = 94 engagements). Of the 16 tested segment levels, 13 held after Holm control.

The mean of an engagement’s 100 indicator ratings, measured at engagement level (base 150 engagements). Where this indicator is one of the 100, the mean is recomputed without it.

The study set r was 0.79 (95% CI 0.72 to 0.84, Holm p < 0.001, n = 150 engagements). Of the 24 tested segment levels, 23 held after Holm control.

The respondent’s stated likelihood of leaving the current job within 12 months, rated 1 to 10 (C019). Measured at respondent level on up to 798 respondents, with the engagement-level form reported alongside.

The study set r was −0.63 (95% CI −0.67 to −0.59, Holm p < 0.001, n = 734 respondents). Of the 37 tested segment levels, 36 held after Holm control.

The cancellation contrast compared engagements with a negative written response against all other engagements. The comparison group comprised 140 other engagements, of which 138 had no scored written response coded to the indicator and 2 had only scored responses that were not negative. The contrast therefore did not test written sentiment by itself. Cancellation was recorded for 3 of the 10 engagements with a negative response (30.00%), against 53 of the 140 other engagements (37.86%). The difference did not reach significance at 0.05 (Fisher exact p = 0.744). The customer satisfaction contrast was not computed, because only 7 active engagements had a negative response, fewer than 10. The association of written sentiment itself with the study measures was tested by two correlations. The association of the engagement mean written score with cancellation did not reach significance at 0.05 (r = −0.45, p = 0.139, n = 12 engagements). The association of the respondent mean written score with intent to leave reached significance at 0.05 (rho = −0.70, p = 0.011, n = 12 respondents). Of the three outcome comparisons, one reached significance at 0.05: the correlation of the respondent mean written score with intent to leave. Both comparisons with cancellation were descriptive, because responses on canceled engagements were written after cancellation. Every comparison rested on a thin base, because fewer than 30 engagements had a scored written response.

Pairs were compared on the engagement level form of every measure, each turned so that a higher value was the better outcome, on the 150 engagements where the indicator and all four measures existed. Of the 6 pairs, 5 differed after Holm control.

Clustering reduced the 782 respondent answers to about 209 independent answers. Respondent level intervals therefore used the cluster bootstrap, and respondent level tests on intent to leave used cluster-robust errors.

No measure showed a tested null. No result was inconclusive. The detection floor column is the study floor at 150 engagements. Intent to leave was tested on its effective n.

Each change was computed before rounding, so it can differ by 0.01 from the difference of the two printed values.

Every check held. A check holds when its test lands on the same side of 0.05 as the primary test and, where the primary test is below 0.05, its r keeps the primary sign. Thin-base rule: engagements in any level under 30 engagements of outsourcing model or process family removed. Institution type is not used, because every one of its levels is under 30.

Each of the four measures was computed from data this indicator does not contain, so each result is a finding rather than an arithmetic consequence.

Each deviation was computed before rounding, so it can differ by 0.01 from the difference of the two printed values.

The comparison ran on 150 engagements. The indicator followed the study’s typical pattern on every measure.

The peer set comprised 101 indicators of the four Parts with 30 or more scored responses. The indicator was outside the peer set, because it had fewer than 30 scored responses. It was added to the ranking only and did not enter the peer median. Its share of negative responses ranked 37 of 102 from the highest. Its mean written score (−5.4, n = 14 responses) ranked 38 of 102 from the most negative. The peer median of the mean written score was −4.6 across the 101 indicators.

The co-coded indicators tied as most frequent in the responses coded to the indicator were “The dashboard and performance reports given to the customer are accurate” (C052, 4 responses) and “The vendor manipulates the dashboard and performance reports given to the customer” (C053, 4 responses). The only themes of the coding rulebook coded in the same responses were “Positive with an embedded negative,” “SLA definitions do not fit agent work” and “Skills learned,” each in 1 response.

Canceled engagements had the lowest KPI ratings (mean 3.4, n = 294 respondents), and Active, Production had the highest (mean 5.8, n = 93 respondents). Individual Contributor respondents gave the lowest ratings by role (mean 4.7, n = 608 respondents), and Portfolio Director respondents gave the highest (mean 5.8, n = 24 respondents). BPO specialist respondents rated lowest by vendor type (mean 4.2, n = 179 respondents), and Consulting firm respondents rated highest (mean 5.3, n = 112 respondents). The associations with the four study measures held under every robustness check run on this item (22 of 22 checks).

The written responses corroborated the rating at respondent level, with a strong positive correlation between written score and rating. Written responses were voluntary. They recorded what the respondents who wrote reported, and how common it was is measured by the ratings. The analysis of the written responses rested on 14 responses, a thin base.

The evidence is reported experience: what the delivery team said about its own engagement, gathered in one cross-section. The figures showed associations, not causes. Indicators about the financial institution were reported at second hand by vendor staff. Cancellation is a descriptive contrast, because canceled engagements were rated after they ended.

Tests across respondent role and vendor type were run at respondent level and treated respondents as independent. The section Clustering and Effective Sample Size states whether, and by how much, clustering within engagements reduced the information in respondent level figures.

Answers clustered within engagements, which reduced the effective number of independent answers (ICC 0.65, 782 answers worth about 209 independent answers). No Customer Management Executive respondent answered this item (n = 0).

Shares of written sentiment were shares of responses, not of the study set of respondents. The written score took five fixed values, so its mean and median were coarse. Responses clustered within engagements. With fewer than 30 scored responses, no interval is reported for written sentiment. Coding was performed by AI coders under a fixed rulebook. Agreement on whether a response was coded to the indicator rested on 1 response second-coded under rulebook 0.5, with the coders agreeing on 1 of 1. Agreement on the score rested on 1 response second-scored under rulebook 0.5, with the scores agreeing in 1 of 1. These bases of second-coded responses are too small for an indicator-level alpha.