FounderWiseDecisions, not feeds
← All articles FounderWise · Long-form

They already know, and they have guessed worse

You are deciding when to tell the team. They have been reading your calendar, your face and who was in which meeting for three weeks, and they have reached a conclusion.

15 Oct 2026 12 min read By Joshua Pi’Rwot
Share X LinkedIn

The round is not closing. You have decided to tell the team once you know what the alternative is, which you expect to know in two weeks.

Two weeks is a long time. Your team has been running an inference on you since the day your calendar changed, and silence is one of its inputs rather than an absence of them.

Why these three models

The decision is when to tell your team something bad and how much of it. The features that fire are a group observing you through signals you did not intend to send, an inference that travels between them faster than any announcement, and a collective read that is more accurate than any individual’s.

Three lenses. Monitoring noise produces a complex answer about what they can actually observe. Spread produces a cycle answer about how their conclusion propagates. Aggregation produces a random answer about how good that conclusion is. Together they invert the usual framing: the question is not whether to tell them, it is whether your account or theirs becomes the operative one.

1. They are reading a noisy signal, and they cannot stop

Start from their position rather than yours. They cannot see the cap table, the bank balance or the investor emails. They can see a set of noisy proxies and they are reading them continuously, because that is what people do when the thing that matters to them is unobservable.

The proxies are specific and you should know what they are. Which meetings appeared in your calendar and with whom. How long the finance conversation ran. Whether the hire that was approved has gone quiet. Whether you answered a direct question with a general answer. Whether two people were asked to stay behind.

Two signals are louder than any of those and founders underweight both, because they look administrative rather than strategic. Payroll timing is the first. A salary that has landed on the 28th every month and lands on the 2nd once has told everybody something, and mobile money makes the exact minute visible to each person separately, so there is no version of it that goes unnoticed. Supplier behaviour is the second. A delivery that arrives with a demand for cash before offloading is read by whoever signs for it, and that person tells the team before you have decided what to say.

Note also what a quiet team is not. A team that asks no questions has not failed to notice. Asking the founder directly is risky in most companies and unthinkable in some, so an absence of questions is entirely consistent with a very active inference, and it should reassure you about nothing. The tells are behavioural rather than verbal: leave taken that was not planned, somebody asking for a letter of employment, a request for a reference from a person who has never needed one.

Every one of those is a low-quality signal about the underlying state, which means their inference is noisy in both directions: they will sometimes conclude disaster where there is none, and occasionally miss real trouble entirely.1 But it is never blank. There is no configuration of your behaviour that transmits nothing, and the attempt to transmit nothing is itself among the loudest signals available, because it is unusual.

The practical consequence is that you are not choosing between informing them and not informing them. You are choosing between a precise account and an imprecise one, and the imprecise one is already running.

Silence does not buy you time. It buys you a version of the story you did not write.

2. Their conclusion travels faster than your announcement

The second lens is about speed, and it is why the two-week plan does not work.

A conclusion moves through a team the way anything moves through a network of people who talk: by transmission between individuals already in contact, at a rate set by how connected they are and how much they want to talk about it.2 Bad news is highly transmissible, because it is useful to the receiver and interesting to the sender.

An all-hands is a single simultaneous exposure. The inference is a spreading process that started three weeks ago. You are not racing from a standing start, you are entering a race already in progress, and by the time you announce, most of the team has heard a version and formed a reaction in a conversation you were not in.

That changes what the announcement is for. It is not the delivery of information, because most of the information has arrived. It is the establishment of an authoritative version, which is a different job and needs different content: specifics, numbers, dates, and above all what is not true, because the circulating version has usually acquired things that are not.

The operational instruction is to move first on the parts you can state. You do not need the resolution to say what is happening. The round is taking longer than planned, here is our runway to the day, here is what we are doing, here is the date by which I will know more, and here is what I will tell you then whatever it says. That is deliverable today and it terminates the spreading process, because it replaces speculation with a dated commitment.

The dated commitment has one failure mode and it undoes everything else. If the date arrives and you say nothing, you have taught the team that your dates are decorative, and every number you give them afterwards is discounted. So the commitment is to report on that date, not to resolve by it. Those are different promises and only the first is inside your control. When the date arrives with nothing resolved, say so, say what changed and what did not, and set the next date. A founder who reports three times running that the position is unchanged is more credible than one who reports once and then goes quiet.

The cheaper partial version, for a founder who cannot face an all-hands this week, is four written lines to everybody on the same day. Written beats spoken here. A spoken version is retold and drifts with each retelling, and a written one is forwarded intact, which is the whole point of establishing an authoritative account.

3. Their collective read is better than you think

The third lens is the uncomfortable one and it should change your estimate of what you are protecting them from.

Aggregated judgements outperform individual ones when the individual judgements are roughly independent and their errors point in different directions.3 A team is a partial version of that: several people, each seeing a different fragment, each with a different reason to over- or under-read it. Your salesperson knows the pipeline is thin. Your engineer knows the infrastructure spend was questioned. Your operations lead knows a supplier asked about payment terms.

Individually those are weak. Assembled over lunch they are a fairly good estimate of the company’s position, frequently better than the founder assumes and occasionally better than the founder’s own, because the founder’s read is anchored on a fundraise they are personally invested in believing.

The condition matters, though, and it is what makes the aggregation unreliable in the direction that hurts you. Once the team starts discussing it, their reads stop being independent, and a group converging on one confident interpretation has the appearance of consensus with the information of whoever spoke first. That is why the accurate collective estimate degrades into a shared and often wrong story if you leave it running.

Confidence is the wrong tell to watch for either way. People draw firm conclusions from very few observations, especially vivid ones, so the certainty in the room says nothing about whether the read is accurate.4 A team can be completely sure and completely wrong, and both halves of that cost you something different.

Two things follow. Do not calibrate your disclosure to what you think they can handle, because they have already handled a worse version. And ask them what they think is happening, individually, before you present. The answers tell you what the circulating account contains and therefore what your announcement actually has to correct.

There is a second-order effect of asking, and it is the reason to ask individually rather than in a room. The question is itself a signal, and putting it to a group starts the discussion that destroys the independence you are trying to measure. Asked one at a time, you get four separate reads and can see where they diverge. Asked in a meeting, you get whichever read the most confident person states first, and you have manufactured the false consensus yourself.

The divergence is the useful output, not the average. Where three of four agree, the account has already settled and your announcement has to address that specific claim by name. Where they disagree, nothing has hardened yet, and that is the part you can still write.

What the three say together

  • Accept that the inference is running. The choice is between your account and theirs, not between telling and not telling.
  • Say what you can today: the situation, the runway to the day, the date you will know more, and a commitment to report whatever it says.
  • Ask three people individually what they think is happening, before you announce. That is your brief.
  • Use the announcement to correct what is false, not to deliver what they have already worked out.

Where they disagree

The aggregation argument and the spread argument disagree about how much to say.

Aggregation says say a great deal: they can handle it, they have assembled most of it, and precision is what your version adds. Spread says that whatever you say becomes highly transmissible immediately, including to people outside the company, and that in a small market a detailed internal account of a difficult month reaches customers and candidates within days.

The resolution is to separate the facts from the interpretation rather than to compress both. Be exact about the situation, the runway and the dates, because those are the parts they are guessing at and the parts that stop speculation. Be sparing about interpretation, blame and hypotheticals, because those are what travel worst outside the room and add nothing to the decisions anyone has to make. A team can hold a hard number. A speculative aside about whether an investor was serious becomes a fact in the market within a week.

What none of them contain

None of the three prices what the disclosure does to the people who leave because of it. Some will, and the ones with options leave first, which is the same selection that operates on customers when you raise a price. The models describe the information and not the consequences of acting on it.

None of them handles the case where you genuinely do not know. All three assume you hold information and are deciding whether to release it. The harder and more common case is that you are as uncertain as they are, and the honest version of this article’s advice is to say that, which most founders find harder than delivering bad news.

None of them covers the person who is told and cannot act on it. A team member who now knows the runway is four months, and who has a family and no savings, has been handed a problem with no lever attached. Disclosure without options is not transparency, it is transferring anxiety, and the honest version of this article pairs every hard number with the specific thing that person can do about it, even when that thing is start looking.

And one property the ensemble will not produce: the team’s reaction changes the outcome. Telling people the round is slow can make it slower, because attention shifts and the best person starts taking calls. Every lens here treats disclosure as reporting on a state rather than as an intervention in it, and it is both.

The one action that survives the ignorance: before you plan the announcement, ask three people separately what they think is going on. You will learn what the circulating version contains, which is the only thing your announcement genuinely has to fix, and you will usually find they are closer to the truth than you expected.

Who has to move

The founder, and the intervention is earlier than feels comfortable rather than more elaborate. The instinct is to wait until you can pair the bad news with a plan, which is generous in intent and expensive in effect, because the waiting period is when the alternative account gets written. The cheapest first test is the three conversations, which cost twenty minutes each and which most founders avoid precisely because they expect the answers to be uncomfortable. That expectation is the finding.

Sources and notes

  1. Jeffrey Carpenter and Andrea Robbett, Game Theory and Behavior, MIT Press. Imperfect monitoring, and the consequence that inferences drawn from noisy signals are wrong in both directions rather than merely absent, are developed in the chapters on repeated games and asymmetric information. Used in section 1. The application here inverts the usual direction: the standard treatment concerns a principal monitoring an agent, and this article concerns a team monitoring the founder.
  2. Albert-László Barabási, Network Science, Cambridge University Press. Spreading processes on networks, and the dependence of propagation on existing contact structure rather than on broadcast, are developed in the chapters on spreading phenomena. Used in section 2 for the claim that an inference already circulating cannot be overtaken by a single simultaneous announcement. No claim is made about the topology of any particular team.
  3. Jeffrey Carpenter and Andrea Robbett, Game Theory and Behavior, MIT Press. The conditions under which aggregated judgements outperform individual ones, and the failure of aggregation once individual estimates become correlated through discussion, are treated alongside information cascades and social learning. Used in section 3, including the important qualification: the team’s collective read is good while the reads are independent and degrades into a shared story once they are not.
  4. Sanjit S. Dhami, The Foundations of Behavioral Economic Analysis, Oxford University Press. The tendency to draw confident conclusions from small and vivid samples supports the observation in section 1 that a team’s inference from a handful of proxies will be confident regardless of its accuracy. Used for the direction of the effect only, with no magnitude claimed.

A note on a number this article does not give. There is no correct number of days to wait. It depends on how visible your signals are, how connected your team is and how much of the picture they already hold. What transfers is that the clock started when your behaviour changed, not when you decided to speak, and that you can find out how far it has run by asking three people.

Joshua Agonya Pi’Rwot, Founder.

Lock in your calls.

You’ve marked 0 of 5. Now choose how often you want the signals.

Step 1 · Pick your cadence

The DispatchWeekly · your Monday 5 callsFreealways

Step 2 · Where to send it

Personalize your BriefThe Brief

Tune every edition to the markets and industries you actually act on.

🔒 Unlock personalization — The Brief, $19.99/mo →
Free Dispatch forever · upgrade anytime · we never share your details.
Know where you stand?
The Dispatch tells you what changed. Knowing what to do about it is a different question, and it is the one FounderWise answers. Start with the free Traction Audit: 12 questions, about 3 minutes, scored out of 100.
Find out where you stand →

For teams, syndicates & programs

Recommended
Team
$15/seat · mo
Daily Brief for the whole team (min 3 seats).
  • Everyone on the same signal
  • Admin + shared watch-list
  • One invoice · ~25% off solo
Get Team →
Cohort License
$1,500/cohort
One cohort of up to 30 founders, for accelerators, funds & programs.
  • MVP Studio and GTM Studio for every founder
  • Your logo, your cohort
  • A funder report on what each founder did
Talk to us →
Pass the Dispatch on
Know a founder making these calls blind? Send them this week’s five — free, every Monday.

Decisions, not feeds. · Curated by Joshua Pi’Rwot · FounderWise · Free Audit · Store · parent of Business Growth Accelerator

Call committed. We’ll hold you to it.
Know where you stand →