Every size of organisation can benefit from psychometric assessments. But the benefit is different at each size, and the mistakes are different too. This piece is a field guide for HR Directors trying to work out what to buy now, what to wait on, and what to never buy at all.
15 people: one team, one founder, one problem at a time
At this size, the company is essentially the founding team's personalities playing out on customers and early hires. The HR function is the founder's calendar and one part-time operator. The question is rarely "do we have succession coverage?"; it is "why does the CTO go quiet in strategy meetings?"
What helps
- DISC for the founding team. Cheap, fast, and produces the conversation everyone has been avoiding about communication styles. Run everyone, not just the loud ones.
- Work Values for the first five hires. Values mismatches at this stage are existential. You are not looking for agreement — you are looking for readable differences so you know what to respect.
What doesn't
- Team-role instruments like Belbin are premature. You do not have enough role coverage for the model to say anything interesting.
- 360-degree feedback. Not enough raters, no anonymity.
- Anything labelled "leadership index". You are one leader with four direct reports. You know your leadership index.
The mistake founders make
They buy one premium assessment for themselves, get a forty-page report, and then don't run it for anyone else because the optics feel weird. Run it for the whole team or don't run it at all.
150 people: the first real HR job
Somewhere between 80 and 150 people, the HR function becomes a function rather than a person. There is usually a first dedicated HR Director or People Lead, often recruited from a much larger company. That person walks in and realises that the systems from the big company do not fit, and the lack of systems from the small company does not either.
This is also the size at which management quality becomes the binding constraint on growth. First-line managers were promoted from individual contributor seats eighteen months ago and nobody has ever taught them to manage. Psychometric reports, used well, are one of the most leveraged things HR can do at this size.
What helps
- DISC for everyone, including managers. Running DISC across the whole of engineering or the whole of sales produces a team map that pays for itself the first time a conflict surfaces.
- EQ for the management layer. The gap between a high-EQ and a low-EQ line manager is the single biggest lever on retention at 150 people. EQ reports give you a development map per manager.
- Conflict Style for the leadership team. Under growth pressure, leadership-team conflict becomes the throttle on every other decision. Five profiles and a facilitated debrief is a one-day session with a multi-quarter payoff.
- Work Values before each senior hire. The cost of a mis-hire at this size is measured in quarters of lost momentum, not weeks. Spend the £129.
What doesn't
- Company-wide psychometric screening. You do not yet have the HR bandwidth to handle the debriefs or the political machinery to handle the fallout.
- Using reports as a hiring gate. The instrument's predictive validity for job performance is real but modest. Use them as a conversation in onboarding instead.
The mistake HR Directors make
They buy a premium assessment suite from the incumbent vendor because they used it at their last job. Six months later, half the licences are unused because the price-per-seat made it a rationed resource, and rationed resources do not get used at the bottom of the org. Better to buy a cheap credit-based instrument and run it widely.
1,500 people: the portfolio question
At 1,500 people the HR function is itself a multi-team operation with its own internal politics. There is a talent team, an L&D team, a DEI function, probably an internal coaching capability. Psychometric assessments are no longer an experiment — they are part of the operating model.
The question shifts from "does this instrument work?" to "what is our portfolio?" You need a small number of instruments you use a lot, not a long tail you use rarely. Integration and data hygiene matter more than the validity coefficient of any single tool.
What helps
- A single instrument as the shared language of leadership development. Pick one — typically a premium forced-choice DISC or a Big Five — and run it through every leadership programme. The consistency is worth more than any individual report.
- A second instrument layered on for specific situations. EQ for coaching, Conflict Style for leadership-team offsites, Derailers for senior succession, Culture Preferences for post-M&A work.
- Team composition reports at the top three layers. The dynamics of the top team, the top-two-layers combined, and the top-three-layers are the levers that move the whole business.
- Annual benchmarking. Retake the same instrument every 18-24 months for senior leaders. People genuinely change, and your succession model should reflect it.
What doesn't
- Bolt-on assessments that are not in the portfolio. They generate reports nobody reads.
- Rolling out any instrument without a trained internal debriefer per 100 people. Reports without debriefs are not development; they are homework.
- Using assessment data as a performance-management input without a bright line between the two. You will lose the trust of your population and the instrument will become useless.
The mistake Chief People Officers make
They let each business unit buy a different psychometric from a different vendor, because each business unit HR lead has a favourite. Three years later there are seven vendors in the spend report, no-one can run a consolidated leadership report, and renewals happen on vendor auto-renewal instead of talent strategy.
One rule that applies at every size
Never run a psychometric assessment for someone without booking the debrief conversation at the same time. The report is one input. The debrief is the product. Companies that treat it the other way round end up with filing cabinets of unread PDFs and people who remember being tested but not being developed.