Deciding Who Investigates and Who Decides
Separating investigation, decision and appeal between three people, what to do in a small organisation, and how to record the constraint.
Role separation in a conduct process, as one organisation assigned it
The appeal hearer was chosen after the appeal was lodged, from the two people who had already been in the room. Every row is one organisation's own record, not a statement of what any jurisdiction requires.
Three roles need three people: somebody to investigate, somebody else to decide, and a third person to hear any appeal. Name all three before the process begins, including the appeal hearer, who is almost always chosen afterwards and therefore chosen from whoever is left.
The decision process in “Deciding Who Investigates and Who Decides” needs evidence that can be checked without turning one activity signal into a conclusion. When reviewers consider see the service here for remote employee monitoring software, its time and project records can provide operational context, but allegations, employee responses and consequential decisions still require a fair human process.
The reason is not ceremony. A process in which one person gathered the evidence, weighed it and then reviewed their own decision produces a file that reads as though the conclusion was reached on day one, whether or not it was.
For an independent reference relevant to “Deciding Who Investigates and Who Decides”, consult the SANS insider-threat resources. Use it to test record quality, access, retention, fair process and exception handling against the organisation’s real departure workflow.
What each role is actually for
The investigator finds out what happened. They do not decide whether it warrants anything, and the moment they start recommending an outcome the separation is gone. Their output is a set of facts, the evidence behind them, and the points the person disputes.
The decision-maker reads that, hears the person, and decides. They need to be able to reach a different conclusion from the investigator without that being awkward, which is why they should not have been in the room while the evidence was being gathered.
The appeal hearer reconsiders. For that to mean anything they have to be able to overturn the decision, which rules out anybody who was consulted about it, signed off the letter, or discussed it over lunch.
The small-organisation problem
In an organisation of eleven people there may not be three managers, and the honest answer is that perfect separation is impossible. Pretending otherwise by having somebody nominally hold a role they did not really perform is worse than the original problem.
What works is to name the constraint and do something about it. The something is usually one of a short list:
- A manager from another site, function or group company.
- A non-executive director, trustee or board member not otherwise involved.
- An external HR consultant engaged for the appeal only.
- A director who has deliberately been kept out of the earlier stages for this purpose.
The last of those requires planning. It only works if somebody decided at the outset to keep a person clear, which is exactly the decision that is never made because nobody expects an appeal.
Recording the constraint
Where full separation is not achievable, write down what was done and why, before the process starts rather than after somebody complains.
Record, at the outset: who is investigating, who will decide, who will hear an appeal, and — where the same person must take more than one role — why that was unavoidable and what was done to limit it.
A documented constraint that was reasoned about in advance is a different thing from one discovered at the end. The first is a small organisation doing its best; the second is an organisation that did not think about it.
HR's position, which is less obvious
The adviser who supports the investigator and then supports the decision-maker carries knowledge between two roles that were meant to be separate. In a large organisation this is avoidable by using different people; in a small one it usually is not.
What is avoidable everywhere is advice that crosses from process into outcome. An adviser who tells the decision-maker what the decision should be has made the decision, and correspondence showing that is a recognisable problem. Advice on procedure, consistency with previous cases and the range of options is a different thing from advice on which option to pick.
Conflicts nobody asks about
Ask the question explicitly, of each person, before they take a role: is there anything about your relationship with this person, or your involvement in what happened, that should rule you out? Record the answer.
It takes a minute and it catches the things nobody would otherwise mention — that the investigator is the complainant's line manager, that the decision-maker was in the meeting being complained about, that two of the people involved are related. These things surface eventually. They are much cheaper when they surface first.
Changing a role part-way through
Occasionally somebody has to step out — they are unwell, they leave, or a conflict emerges that nobody saw. Replacing them is fine; doing it silently is not.
Record the change, the reason and the date, and make sure the incoming person reads the file rather than being briefed verbally by whoever is handy. A decision-maker who was appointed in week three and only knows what somebody told them over coffee is exactly the position the separation of roles exists to prevent.
Naming the appeal hearer in advance
The single most useful change most organisations can make is to identify the appeal hearer at the start and write the name down, even though no appeal has been lodged and most likely none will be.
It costs nothing, and it removes the specific failure this page is about: an appeal heard by whoever was available, which in practice means somebody who was already involved. That is the step that cannot be retrofitted once the appeal has arrived.