The essentials in 30 seconds:
→ Likeability is not a bad signal in hiring. A team has to work on a human level.
→ The problem is the asymmetry: likeability is experienced thoroughly, competence is often tested only at the surface.
→ Teams assembled on likeability alone end up thinking alike. Teams built on competence alone cannot hold the friction.
→ Three steps close the gap: the same questions for everyone, a task that genuinely relates to the job, and standards defined in advance.
→ After that, gut feeling still gets a say. It just no longer decides on its own.
Likeability is not a bad signal
This is worth stating first, because the professional debate often suggests the opposite. When a managing director says someone fits them better, she has perceived something real. Whether a person can work inside an existing structure, whether they listen, whether they can take disagreement: none of this is a side issue. A team that does not work on a human level does not deliver.
The mistake is not that likeability plays a role. The mistake is that it ends up being the only thing that was genuinely tested.
Why competence stays at the surface
A conversation in which everyone gets along is pleasant. A work sample in which someone can visibly fail is not. It is uncomfortable for the candidate, and it is uncomfortable for the person who has to set it and then assess it.
So it is often dropped or watered down. A case question is asked that has little to do with the job. A presentation is requested that mostly measures presence. Or the interview is relied on and called experience.
This is the moment where I become inconvenient in a mandate. Not on principle, but because it is part of the job: if nobody takes the opposite position, the more comfortable route wins. Every time.
What happens when only one side counts
A recent study in Personnel Psychology shows this in founding teams. Chen and colleagues interviewed 36 real founding teams, successful and failed ones alike, and found two patterns.
Teams assembled on likeability alone had little conflict and a great deal of trust. And over time they all thought alike. One founder put it this way: the company eventually developed its own "DNA," and the longer it ran, the harder it became to turn. New directions were not rejected; they simply stopped occurring to anyone.
Teams built on competence alone had the diversity, but not the trust to hold it. Professional disagreements turned personal. One founder describes the formation of camps: two groups, and no way back between them.
The second part of the study is a four-week business simulation with 274 student teams. The statistical effect there is real, but small. I find the interviews with real founding teams considerably more convincing, and this text rests on those.
What transfers is the pattern, not the number. The study is about how founding teams are assembled, not about filling a single position in an established SME. What does transfer is the mechanism: where only one of the two dimensions is properly tested, things break at exactly the point where they become uncomfortable. And nobody notices on the day of hiring. They notice months later.
Three steps that make the difference
None of this is complicated or expensive. It is craft.
1. The same questions for every candidate
Same questions, same order. It sounds bureaucratic and it is the single most effective step: structured procedures predict job performance considerably better than open conversations (Sackett et al., 2023). At the same time they protect against mistaking interview routine and presence for capability. Anyone who communicates differently drops out of an open conversation before their abilities have been seen at all.
2. A task that genuinely looks like the job
Not some generic case, but something the person will actually do in the role. Costing a real quotation. Replying to a difficult customer email. Preparing a team meeting. If the task cannot be derived from everyday work, the role has not yet been described properly. That is the real finding.
3. Define the standards in advance, not afterwards
What counts as a strong answer, an average one, an inadequate one? If that is decided only after the conversation, the standard adapts to the person you happened to like. Defined beforehand, the person is measured against the standard.
And gut feeling?
It stays. It is information about working together, and that counts. After these three steps it simply no longer decides on its own, and it no longer distorts the assessment of competence and potential that has already taken place.
What this means for Swiss SMEs
In smaller organisations a bad hire weighs more heavily than in a corporation. There is no second department to move someone into, and depending on the position the cost quickly reaches five or six figures once onboarding, lost productivity and refilling the role are added up.
For leadership positions a second effect is often overlooked: the person you hire will go on to hire others. A selection error at that level multiplies, because the pattern carries downward.
The structural advantage of SMEs is the short distance. Where management, leadership and the team sit in the same building, these three steps can be introduced in weeks rather than quarters. No software is needed, and no certification. What is needed is someone who does not fold the first time the process gets uncomfortable.
References
Chen, L., Ling, C.-D., Zhang, M. J., Wang, N., & Huang, W. (2026). Team Dual Formation Strategy, Team Ambidexterity Trajectory, and Entrepreneurial Effectiveness. Personnel Psychology, 79(3), 327–351. https://doi.org/10.1111/peps.70030
Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2023). Revisiting the design of selection systems in light of new findings regarding the validity of widely used predictors. Personnel Psychology, 76(4), 1113–1139. https://doi.org/10.1111/peps.12572
Is there a hire coming up where it must not go wrong?
30 minutes, no sales pitch. You tell me which role is coming up and how you select today. I listen, then straight talk. Afterwards you will know whether a structured process is worth it in this case, even if we do not end up working together.