When something you own is at risk of slipping, how and when do you tell the people depending on it?
answer
- your trigger for speaking up
- who hears it first, and how
- headline first, causes after
- propose an adjustment, not just a problem
- the cadence that keeps it current
basics
~20 sProbes whether people hear about problems from you or from the calendar. Give a trigger, an audience and a cadence: speak the moment your estimate stops being true, tell the affected person directly, and bring a proposed adjustment.
how to answer
5 beats- the trigger that makes you speak upOpen with your rule, stated concretely: the moment your forecast stops matching the commitment. A named trigger beats a vague promise to communicate proactively, because it is testable.
- who hears it first and through what channelSay how you pick the audience, usually the person whose commitment is affected, and why they hear it before a group channel does. Mention when you escalate beyond your immediate lead.
- the headline, then the detailDescribe the shape of the message: the impact and the new expectation in the first sentence, cause and context after. Keep it to a few lines so it gets read.
- the adjustment you bring with itExplain that you arrive with a proposed change and a specific ask, and that you keep the picture current afterwards rather than treating the flag as the end of the conversation.
- a short example that proves the habitSpend thirty to forty seconds on one concrete instance, with the interval and one number. Without it this answer is a policy statement, and policy statements are cheap in interviews.
your answer
5 story prompts- Write down your actual trigger for raising a risk, in one testable sentence.
- Recall one time your estimate moved and note the interval between knowing and telling.
- Name who you told first and why that person rather than someone else.
- Have one line ready for the adjustment you proposed alongside the news.
- Note your regular status cadence, if any, since it makes the escalation credible.
draft and rehearse your own answer in a learn session
go deeper
This probes communication discipline and ownership rather than a single dramatic episode. The interviewer wants to know your default: what makes you speak, who you speak to first, and whether the message is actionable. A strong answer shows that stakeholders stay calibrated as work moves, and that raising a risk is routine for you rather than an admission of failure.
My rule is that I say something the moment my own estimate stops being true, not on the day the work was due. On a security platform team of three, my piece of an MFA rollout was the self-service enrollment page, which was meant to lift enrollment share past the halfway mark without the helpdesk touching every user. I had estimated five days. On day two I found that the older internal portal it had to live inside refreshed its session token in a way our new flow broke, and I was going to need an hour from whoever knew that portal. I raised it at standup that morning rather than spending another day trying to rescue it quietly. I kept it to three lines: the page lands Thursday rather than Tuesday, the reason is the token refresh in the old portal, and if I get an hour with its owner today then Thursday holds. My lead pulled that person in the same afternoon. It landed Thursday, and nothing downstream moved, because the announcement to the first department had not gone out yet. Since then I keep my current estimate and my confidence in the ticket itself, so my lead can see a wobble without having to ask me how it is going.
Junior scope handled well: one task, but the trigger is explicit, the interval is short, and the ask is specific enough for someone to act on the same day. Checking what did not move downstream shows you understood who was depending on you. Vague on the ask and it becomes a status complaint.
I try to forecast rather than report, because by the time a number turns red the other person has lost their options. I ran a seven-week MFA pilot for one business unit of about eleven hundred people ahead of the enterprise-wide push. My commitment to their operations manager was 85 percent enrolled by the end, and he had staffed a support desk around that. Four weeks in, the curve was tracking to land near 71 percent, because roughly a fifth of that unit worked shifts and never opened a weekday e-mail. I told him nine days before his desk was due to stand down, not on the day. He got the forecast, the reason, and two adjustments to choose from: run enrollment at shift handover with a laptop cart, or extend his desk by a fortnight while I re-targeted the reminders. He took the handover option and offered a supervisor to run it, and we finished at 88.6 percent. The cadence, not the recovery, is what I would keep. Every Friday he got four lines from me: the number, the forecast, what changed, what I need. By the time I brought him something unwelcome it was the ninth of those, so it read as an update rather than an ambush.
The middle bar shows in forecasting instead of reporting: the news arrives nine days before it matters, with two adjustments the other side can choose between. The standing four-line update is what makes an unwelcome message land as routine. Strip the cadence and this reads as a lucky catch.
Your trigger can be simple, but make it explicit: the moment the estimate you gave stops being true. Say who you told, how quickly, and what specific help you asked for.
Show that you forecast rather than report. You saw where the number would land before it went red, and you gave the other person a choice between adjustments instead of a single fait accompli.
Talk about the cadence your team runs and why it is trusted: what counts as amber, who is on the distribution, and how something gets escalated past you when it needs to move faster than you do.
Frame it as designing how risk surfaces between teams. The default channels, the forum where slips are visible without punishment, and the incentives that make people raise things early instead of absorbing them.
saying these in an interview costs you the question
- Waiting for the next scheduled status meeting when the news cannot wait
- Telling peers but never the person who owns the commitment
- Flagging every wobble until real risks stop standing out
- Describing the problem with no proposed adjustment or ask
- Announcing it in a group channel before telling the person affected
- Claiming your estimates never move so this has never come up
- How do you decide between sending a message and pulling someone into a call?Tie it to consequence and ambiguity. A written note works when the adjustment is small and the reader can act alone; a live conversation is right when someone has to make a tradeoff, when the news affects their commitments to others, or when a written version would read worse than it is. Say you follow the call with a written summary.
- What do you do if you are asked to sit on it for a few days?Show you can hold the tension without either leaking or going silent. Ask what the wait buys, agree an explicit date when it goes out regardless, and make sure nobody is making decisions in the meantime that the news would change. If someone is about to commit resources on stale information, say that plainly.
- How do you avoid crying wolf?Distinguish a risk you are managing from one that needs a decision. The first belongs in your regular update with a confidence level; the second earns an interruption. Mentioning that you close the loop when a flagged risk resolves is a strong detail, because it is what keeps your amber signal meaningful.