When One Side Is Deliberately Optimistic
The incentives that make each side of a capacity account shade its numbers, how to tell shading from error, and the structural fixes that work better than exhortation.
Both sides of a capacity account have reasons to shade their numbers, and the shading goes in opposite directions. The supply side has an interest in a lower available figure; the demand side has an interest in a lower claim.
The reconciliation in “When One Side Is Deliberately Optimistic” is strongest when assumptions and current work records can be examined separately. A team considering using project time recording in workforce forecasts for project time recording can add time and project context, with transparent settings, limited access and a clear correction route around consequential decisions.
Neither is dishonesty. Both are rational responses to how the numbers get used, and the fix is structural rather than moral.
Additional context for “When One Side Is Deliberately Optimistic” is available from AWS operational-excellence guidance. Use external material to prompt a specific validation check rather than replace local judgement.
Why supply shades downward
Because a high availability figure becomes a commitment, and because the cost of being wrong is borne by the team. A manager who publishes an optimistic figure will be held to it.
A small private buffer is the natural response and is invisible inside a shrinkage rate nobody audits.
Why demand shades downward too
Because a small claim is more likely to be approved. A sponsor asking for two people when they need three improves their chances of getting the two, and the third arrives later as an overrun.
This is the more damaging of the two, because it moves the shortfall into the quarter rather than into the plan.
Telling shading from error
Error is random: some lines high, some low. Shading is systematic: every line in the same direction, cycle after cycle.
Keeping the variance series by claim and by team makes the difference visible within three cycles and requires no accusation to interpret.
The buffer that should be explicit
Supply-side buffers are legitimate. A plan with no contingency is a plan that fails on the first absence.
The fix is not to remove the buffer but to name it: a contingency line in the descent, with a figure and a reason, rather than a shrinkage rate quietly set two points high.
Making the contingency a decision
Once it is a line, somebody can ask how big it should be and the answer can vary with the period: more in winter, more during a migration, less in a stable quarter.
A hidden buffer cannot be tuned, because nobody can see it. Named, it becomes one of the few genuinely adjustable things in the account.
The demand-side fix
Ask for the claim and the confidence separately: how much, and how certain. A sponsor who can say "two for certain, probably three" has given more information than one forced to produce a single number.
The book then carries two figures, and the provisional portion goes in the likely total rather than the committed one.
What punishes honesty
A process that treats an honest larger claim as poor planning, or an honest lower availability figure as a performance problem, will get shaded numbers forever.
The behaviour follows the incentive with complete reliability, and no amount of insisting on accuracy changes it.
The variance record as the corrective
Where actual usage is recorded against each claim at closure, the organisation accumulates a record of who estimates well.
Used gently — as a conversation about estimating rather than a scorecard — that record is the only mechanism that reliably improves claims, because it replaces assertion with history.
The one number to watch
Total claims against total capacity, over several cycles. If the organisation is consistently committing above its capacity and delivering anyway, something is absorbing the difference, and it is usually people's own time.
That series is also the strongest possible argument for the account itself, because it shows a pattern no individual quarter can.
The three places a buffer can sit
- Inside the shrinkage rate, where it is invisible and untunable.
- As a named contingency line in the descent, where it can be argued about.
- As unallocated capacity in the plan, where it is visible to everybody and will be consumed by the first arrival.
The second is the only one that behaves. The first hides the decision; the third exposes it to whoever asks fastest, which is not the same as whoever needs it most.
What an honest account costs the first time
Usually one uncomfortable quarter. Supply buffers come out into the open, demand claims grow as people stop understating them, and the difference gets worse before the figure stabilises.
Saying that in advance is worth doing, because an account that gets worse in its first cycle and was predicted to will survive, and one that gets worse unexpectedly will be blamed for the deterioration it merely revealed.
What to say when you suspect shading
Nothing about the person. Show the series, ask what would make the figure easier to give honestly, and fix that.
In most cases the answer is a guarantee that an honest number will not be used against them, which is cheap to give and, once given, usually believed on the second cycle rather than the first.