A stronger coach-client working alliance goes with better coaching outcomes (association, not cause)
A meta-analysis of 27 samples (3,563 coaching processes) finds a moderate, robust association between working alliance and client outcomes (r = .41), strongest for satisfaction and perceived effectiveness, and a negative association with unintended negative effects (r = −.29).
Finding
What the study measuredAcross 27 samples (3,563 coaching processes), working alliance was moderately associated with client outcomes (r = .41, 95% CI [.34, .48]). Associations were strongest with coaching satisfaction (r = .64) and perceived effectiveness (r = .58), and lower but still moderate with self-efficacy (r = .32) and results such as goal attainment (r = .32, k = 14). Alliance was negatively associated with unintended negative effects (r = −.29). The association held whether clients rated both alliance and outcomes (r = .47), coaches rated both (r = .34), or coaches rated alliance and clients rated outcomes (r = .26). It did not differ by client sample, coach expertise, number of sessions, language or peer-review status, and there was no sign of publication bias.
Authors' interpretation
What the authors concludeThe authors conclude that, as in psychotherapy and mentoring, a high-quality working alliance matters for coaching outcomes, and that the relationship in coaching appears nearly twice as strong as in psychotherapy. They recommend that coaches monitor the alliance throughout coaching, that clients use a first session to judge fit, and that supervisors use alliance as a starting point. They state that the data are correlational and call for cross-lagged designs to test the direction.
| Question type | association |
|---|---|
| Population and setting | Coaching clients in field samples (67% of samples) and university students (33%); professional coaches in 56% of samples, novice coaches otherwise; studies in English (74%) and German; Workplace, executive, life and career coaching; diverse industries, universities, military, manufacturing and telecommunication |
| Intervention | Not applicable: the analysis relates the quality of the coach-client working alliance to outcomes (coaching averaged 8.6 sessions where reported) |
| Comparison | Not applicable (correlational synthesis) |
| Follow-up | First measurement point used where studies had several |
| Outcomes | Affective: coaching satisfaction, perceived effectiveness, self-efficacy; Cognitive (self-reflection, insight); Results (e.g. goal attainment); Unintended negative effects |
| Measures | Working Alliance Inventory adapted for coaching; A German coaching-specific alliance scale; Study-specific alliance scales; Alliance and outcomes rated by clients, coaches or both |
| Effect size and uncertainty | Overall r = .41, 95% CI [.34, .48], k = 27; heterogeneity Q = 94.97, I² = 72.62. Affective r = .53 [.44, .60]; cognitive r = .49 [.19, .71]; results r = .32 [.25, .39]; unintended negative effects r = −.29 [−.42, −.14]. |
| Limitations | Small number of studies and high heterogeneity; all included studies are correlational, so cause cannot be inferred and the direction may run both ways; skill-based outcomes could not be tested; one study supplies about half of the total N (results unchanged without it). |
| Funding and conflicts | No specific grant from any funding agency. No conflict-of-interest statement found. |
| Replication | Not applicable (synthesis of prior studies). |
| Source reuse terms | Publisher copyright (© The Author(s) 2019, SAGE); not open access as checked 2026-10-24. |
Figures reported, with their location in the source
- 23 studies, 27 samples, N = 3,563: Evidence base (p. 44; Figure 1 (p. 43); p. 45)
- r = .41, 95% CI [.34, .48]: Overall association (Table 3 (p. 48); p. 45)
- Q = 94.97, I² = 72.62: Heterogeneity (p. 45)
- r = .64 / .58 / .32: Satisfaction / perceived effectiveness / self-efficacy (Table 3 (p. 48))
- r = .49 [.19, .71], k = 5: Cognitive outcomes (Table 3 (p. 48))
- r = .32 [.25, .39], k = 14: Results outcomes (e.g. goal attainment) (Table 3 (p. 48))
- r = −.29 [−.42, −.14], k = 5: Unintended negative effects (Table 3 (p. 48))
- r = .47 / .34 / .26: Client-client / coach-coach / coach-rated alliance with client-rated outcome (Table 4 (p. 48); p. 49)
- r = .41 [.32, .49]: Overall result without the largest study (p. 49)
Read the source
Graßmann, C., Schölmerich, F., & Schermuly, C. C. (2020). The relationship between working alliance and client outcomes in coaching: A meta-analysis. Human Relations, 73(1), 35–58. https://doi.org/10.1177/0018726718819725
Verification record
- Status
- human-verified
- Access level
- full-text
- Date accessed
- 2026-10-24
- Tool-checked
- 2026-10-25
- Automated retraction check
- 2026-10-24: Crossref record has no update, correction or retraction notices (Crossref carries Retraction Watch data). Owner to confirm.
- Human-verified
- 2026-10-25
- Retraction status
- none-found (checked 2026-10-25); Owner check: "Retraction check: r-0005, none-found"
- Last reviewed
- 2026-10-25
How to cite this entry
consultants.coach. (2026). A stronger coach-client working alliance goes with better coaching outcomes (association, not cause) (Entry r-0005, version 2026-10-25 00:00:00 -0400 EDT). https://consultants.coach/research/r-0005/
Original source: Graßmann, C., Schölmerich, F., & Schermuly, C. C. (2020). The relationship between working alliance and client outcomes in coaching: A meta-analysis. Human Relations, 73(1), 35–58. https://doi.org/10.1177/0018726718819725
Change log
- : Rebuilt from the full text (published version supplied by the editor); grade assigned.
- : First published after human verification.
Related entries
Randomized trials of coaching show a moderate average effect, with signs of publication bias
A meta-analysis of 37 randomized controlled trials (39 samples, n = 2,528) finds coaching improves leadership and personal outcomes, with a best estimate of g = .59, and reports significant publication bias.
Workplace coaching meta-analysis: positive effects overall; the largest effects rest on very few studies
A meta-analysis of 17 workplace coaching studies (2,267 people) finds a positive overall effect (δ = 0.36), larger without multisource feedback and with internal coaches, but the biggest effects come from subgroups of only three studies.
Our commentary
consultants.coach editorial viewThis is the most robust finding in our library so far, and it is about association, not cause. The strongest links are with how clients feel about the coaching (satisfaction, perceived effectiveness), which are close to the alliance itself; the link with goal attainment is smaller but rests on the most studies. Two details make it more convincing: the link held when coaches rated the alliance and clients rated the outcomes, which reduces the worry that one person is rating everything the same way, and results did not change when the very large single study was removed. For sponsor reporting, alliance quality is a useful, confidential process measure to track, but not evidence of impact in itself.