Imagine you're discussing a piece of work with your team. You need it by Friday, so you ask, “We can have this finished by Friday, can't we?”
Someone says yes. You put the date in the plan.
What have you learned? Perhaps the work really can be finished by Friday. Perhaps someone has pictured a clear week with no interruptions. Or perhaps you've received agreement with the date you wanted, without discussing what it will take.
A useful deadline conversation separates three things: when you need the work, when the team currently expects to finish it, and what would have to change if those dates don't match.
You can still be ambitious. You just need to know which part of the plan is an estimate and which part is a request.
Ask what “finished” includes
Before asking how long something will take, agree what you're asking someone to complete.
Does “finished” mean the first draft is written, the figures have been checked, the client has approved it, or the final version has been released? A date attached to different definitions isn't much of an agreement.
For a hypothetical customer guide, you might define completion as: the content is checked by the product owner, the layout is accessible, the final file is approved and the support team knows where to find it.
That definition reveals work a shorter instruction might leave out. It also shows whose involvement needs confirming. The person writing the guide cannot simply promise when another team will approve it.
Try asking: “What needs to be true before we can honestly call this done?”
Then separate the work the team controls from the responses, materials or decisions it needs from others.
Look at what happened last time
A plausible plan can be a very tidy story. Draft on Tuesday. Review on Wednesday. Final changes on Thursday. Done on Friday.
Before relying on that story, look for comparable work you've actually completed. What were the start and finish dates? Where did the work wait? How much checking or revision was involved?
Research by Roger Buehler, Dale Griffin and Michael Ross examined optimistic predictions across academic and everyday tasks. In one experiment with 123 students, explicitly connecting relevant past experience to a new assignment reduced optimistic bias. Simply recalling the past did not have the same effect. Importantly, the intervention did not improve absolute prediction accuracy or make participants finish earlier. These were student tasks, not a test of this management conversation. Read the original 1994 paper.
The practical question is not just, “Have we done this before?” It's, “What does that experience tell us about this estimate?”
If the last three guides waited several days for review, don't remove that waiting time merely because this guide feels urgent. Check whether review capacity has actually changed.
You might have good reasons to expect a different result. The template may already exist, the scope may be smaller, or a reviewer may have reserved time. Write those reasons down. “This time we'll be more focused” needs a little more detail.
HM Treasury's guidance on optimism bias also recommends using evidence from past or similar projects and adjusting for the project's particular characteristics. That is UK public-project appraisal guidance, not a ready-made formula for an Australian team's weekly workload. The transferable idea is to make the comparison explicit, rather than borrowing its percentage adjustments. Read the guidance.
Put the conditions beside the date
“Friday” is easy to repeat. The conditions that make Friday possible can disappear from the conversation.
Ask the person estimating to explain the important dependencies alongside the date. For example:
We expect to have the guide ready on Friday if the product details arrive on Tuesday and the reviewer can respond on Thursday. Neither is confirmed yet.
This is an illustrative estimate, not a suggested commitment for every project. Its value is that you can see what needs checking.
When uncertainty is substantial, a range may be more honest than a single date. But don't decorate a guess with precise-looking probabilities. If you say you are 90 per cent confident, be clear about what supports that number. A range based on a few comparable jobs is still limited evidence.
You can also be explicit that you don't yet have a useful estimate: “We need to inspect the existing material before we can say.” Agree what will be checked and when you'll return to the question.
The aim isn't to add a generous cushion to everything. It's to make uncertainty visible enough to manage.
If the date is fixed, discuss the trade-off
Sometimes Friday really is fixed. A launch, conference or external commitment may leave little room to move.
That doesn't mean the forecast should quietly become more optimistic. It means there is a decision to make about the work.
For the hypothetical guide, you could ask:
- Which sections are essential for Friday, and which could follow?
- What other work could be paused, and who has authority to decide that?
- Can the necessary review time be confirmed?
- Is there a smaller, usable version that still meets the requirements?
- If none of those options works, who needs to know now?
Don't remove required quality, safety or compliance checks to make the dates agree on paper. Use the relevant specialist process for work where those checks matter.
The conversation may end with a revised scope, a different priority or an escalation. It may also confirm that the original date is workable. Either way, you have more than a reassuring yes.
Agree when to look again
An estimate made before the work begins won't contain everything you'll learn while doing it.
Choose a review point tied to useful information: after the material arrives, after the first section is tested, or when the reviewer confirms availability. Ask what would justify changing the estimate before that point.
If a dependency slips, return to the conditions you recorded. “The product details haven't arrived, so the Friday estimate needs reviewing” is clearer than waiting until Thursday to announce a surprise.
When the work is complete, keep a short record of what you expected and what happened. Note scope changes separately from estimation mistakes. Over time, those records can give the next conversation something firmer than memory to work with.
This is a practical routine to try, not a validated forecasting system. For complex projects, use the planning and risk expertise the work requires.
Start with a different question
At your next planning conversation, replace “You can do that by Friday, can't you?” with:
Friday is when we need it. What does the evidence suggest is achievable, and what would need to change to get there?
Then listen for the difference between a plan, an assumption and a confirmed arrangement.
If several people hold different parts of the information, you can also use a process for getting independent views before a team decision. Here, the immediate task is narrower: make the reasoning behind the date visible.
If your leaders would benefit from practising clearer thinking and more useful everyday conversations, explore the Leadership Mindset Workshop. Bring the situations your team is working through so we can discuss whether the format fits.