Teams Turnover and project drift
Turnover is treated as a human resources subject, and project drift as a steering subject. On a fixed-price project carried by a small team, they are nevertheless two sides of the same fact: a single unplanned departure tips the margin, and it takes with it a knowledge of the project that nothing replaces in a few weeks.
What a departure costs, and what it costs on top on a project.
These figures describe the cost of replacement. On a fixed-price project, you have to add what the studies do not count: knowledge of the client context, the history of decisions, the reasons behind technical choices. That part cannot be bought back on the labor market. The details of the studies are on the value lost in projects.
How a departure becomes a drift.
The chain is always the same, and it unfolds over a few weeks.
The replacement takes time
Recruiting or reassigning takes weeks. Meanwhile, the team's capacity has dropped while the commitment has not moved.
The team absorbs
The remaining people take on the load. They hold, and this overload in turn becomes a factor in further departures.
The handover costs
The replacement has to understand the project, and the person best placed to explain it to them is no longer there.
The margin tips
Consumed rises, progress stalls. On a fixed-price project with a small team, a single departure is enough.
The signal existed. It was written down nowhere.
Let us say it upfront: pylotik does not forecast departures, and does not claim to detect them. That would be false technically, and it would be a wrong toward the people concerned.
What we observe is more modest and more useful. In hindsight, after an unplanned departure, the signals almost always existed: an overload that had lasted six weeks, a weekly report whose tone had changed, a blocking point raised three times without an answer. What was missing was not someone's perceptiveness. It was a place where these elements were written down, dated, and read by someone other than the person concerned.
The difference between a weak signal and a readable signal therefore does not come from an algorithm: it comes from the existence of a trace and of a ritual. This is what we call operational commitment management, applied to the human thread as much as to the contractual one.
No one is rated, nothing is scored.
The product has no flight-risk index, no list of people to watch, no analysis of the tone of a text. Such mechanisms exist elsewhere on the market; we do not build them, and it is not a question of feasibility but of principle.
A mechanism that singled out people as being at risk would produce exactly the opposite of the intended effect: no one would write sincerely any more, knowing that their text feeds a ranking. The processing of personal data is described in the privacy policy.
What exists, and is enough.
Three traces, and a ritual that gets them read.
Workload, over time
A week over capacity can be made up. Six consecutive weeks are a state, and it shows in the workload plan.
The weekly report
What got stuck, and in what tone. Reread over six weeks in a row, it says what no single week said alone. See the weekly report.
The two-level reading
The manager and the human resources department read the same thing. The case where the manager is the subject stops being a blind spot.
The overall frame of the human side is on team engagement. The full tour is on the product page.
What we are asked about departures.
Does pylotik foresee departures?
No, and it never will claim to. There is no risk index and no list of people to watch. What exists are dated written records and a ritual that gets them read, which is something other than a prediction.
How do you anticipate a departure in a project team?
By looking at what is observable and dated: an overload that lasts, a blocking point raised without an answer, a listening ritual that exists and that someone responds to. It is not anticipation in the predictive sense, it is attention with tools.
What is the impact of a departure on the margin of a fixed-price project?
It adds up: replacement cost, capacity down for several weeks, time for the replacement to pick up the context, and overload of the remaining people. On a small team, a single departure is enough to tip a fixed-price project.
Are employees rated?
No one is rated, nothing is scored. No figure is attached to a person, and no ranking exists. It is a design rule.
What should you do about an overload you cannot resolve?
Make it visible and date it. An overload that is accepted and discussed is a trade-off; an overload that is endured and left unsaid becomes a delay, then sometimes a departure. The second costs far more.
Is it a human resources matter or a steering matter?
Both, and that is precisely why the subject so often falls between two stools. Workload is project data, climate is team data, and a departure hits them together.
Going further.
- Team engagement: two accounts of the same week.
- The weekly team report: the template and its use.
- Lost value: the sourced figures, including those on disengagement.
- Operational commitment management: the definition of the category.