Costs
Part of Why your own first hundred days rarely go in a straight line
First 100 Days Examples: Seven Founder Scenarios
First 100 days examples for founders: seven hypothetical early-business situations, what each may indicate, and a practical next move.
These first 100 days examples are seven hypothetical situations set in the first weeks after money starts changing hands. For each, the article describes what it looks like, what it may indicate, and a possible next move.
All of them are invented for this page. They are here because the difficulty in this period is almost never a shortage of events. It is that each event has an obvious reading and a correct one, and they are usually different.
What to take away
- Nearly every early situation is evidence about a narrower thing than it appears to be evidence about.
- The next move is more often a question than a change, and questions are cheap in exactly the weeks that changes are expensive.
- When a situation has two readings, the cheapest observation that separates them is usually a conversation with somebody who stopped.
The first week is excellent and the third is silent
What it looks like. The product worked and then stopped working.
What it is evidence of. A one-time audience, spent. The launch reached everyone reachable through routes that do not renew: your contacts, an announcement, a moment of attention. Week three is the actual baseline and week one was a stock, not a flow.
Next move. Build one route a stranger could find without knowing you exist, and read weeks five onward as the real number. Do not touch the product. Choosing a route deliberately rather than announcing once is the substance of the promotion and sales guidance.
Everyone signs up and nobody comes back
Steady arrivals. Almost nobody active in the second week.
Diagnose the retention drop
Ask five people what they expected
Frequency problem: promise matched delivery
Offer sheet is lying: promise did not match
What it looks like. A retention problem.
What it is evidence of. Either a promise that did not match delivery, or a problem that recurs less often than your billing does. Those want opposite responses and they are easy to tell apart: ask five people what they expected. If the answers describe your marketing, it is a frequency problem. If they describe something else, the offer sheet is lying.
Next move. Five short conversations before any product work. The technique for hearing what they expected without prompting them is what customer interviews is about.
One customer is enormous and everything bends toward them
They pay well, they ask for things, and it is hard to say no.
Keep the whale, stay a market
- Keep serving the large customer
- Record their requests in a separate file
- Check monthly for a second customer asking
- Unasked requests are commercial, not product
What it looks like. Traction.
What it is evidence of. One relationship. That is a real asset and it is not a market. The specific risk is that six weeks of work goes into requests only one organization has, and the requests are indistinguishable from feedback while you are doing them.
Next move. Keep serving them, and record their requests in a separate file from everyone else's. Then check monthly whether anything in that file has been asked for by a second customer. Anything that has not is a commercial decision, not a product signal.
The support inbox is full and nothing is broken
Lots of questions, no defects, people generally succeeding.
Close the comprehension gap
- Count questions per active customer per week
- Group the questions
- Fix the largest group at its source
- Watch whether that group shrinks
What it looks like. Engagement.
What it is evidence of. A comprehension gap, and it is measurable. Count questions per active customer per week and group them. The largest group is where the product or the page is unclear, and it is usually not the group that feels loudest.
Next move. Fix the largest group at its source, in the page or the first-run experience, and watch whether that group shrinks. If it does not, you fixed something adjacent to the actual confusion.
The numbers look good and you cannot say why
Retention holding, arrivals steady, nothing obviously wrong.
Aggregate vs cohort view
Whole population
- What it shows
- Newest people dominate
- When growth grows
- Stays flattering
- What it tells you
- Nothing reliable
Split by arrival week
- What it shows
- Each cohort down the weeks
- When growth grows
- Shapes diverge
- What it tells you
- Whether it is working
What it looks like. It working.
What it is evidence of. Possibly that, and possibly that your whole population arrived recently and has not had time to leave. A whole-population figure while arrivals are growing is dominated by the newest people, and it stays flattering right up until growth flattens.
Next move. Split by arrival week and read one cohort down the weeks. If cohort one and cohort three have different shapes, the aggregate number was never telling you anything. The arithmetic for that split is in the market metrics.
People cancel and say it is too expensive
Several cancellations, the same phrase in each.
What it looks like. A pricing problem.
What it is evidence of. One of three different things wearing one phrase. They wanted it and could not justify the amount, which is a proof or positioning problem. They did not want it much, which is a value problem no lower price fixes. Or they are smaller than the segment you wrote down, in which case the price is right and the audience is wrong.
Next move. One follow-up question separates them: would you have paid that for something that solved this completely. Then decide. The full separation, and what to do with each answer, is worked through in the pricing questions.
A customer writes something glowing and you want to use it
An enthusiastic email arrives in a slow week.
What it looks like. A testimonial.
What it is evidence of. One person's genuine opinion, given privately, in a context you did not disclose. Using it publicly is a decision with rules attached: the endorsement has to reflect what the person actually experienced and thinks, and any connection between you and them has to be disclosed, which the FTC endorsement guidance sets out in detail. Editing it into a stronger claim than the person made converts it into a claim of your own, and claims need a basis before they run.
Next move. Ask permission, quote it unedited, and disclose anything they received. If the quote is not usable that way, it was not a testimonial.
Reading across the seven
Across these scenarios, an event arrives with an obvious reading attached, while another reading may be about reach, timing, or who arrived rather than the product.
That bias has a cause: product work is available at any hour and feels productive. Reach work and conversations with people who left are neither.
The habit that corrects it is to write both readings down before acting, then name the cheapest observation that separates them. It is a version of the same discipline the pillar on the first hundred days applies week by week.
Common questions
Are any of these real companies?
No. All seven are constructed for this page, and no counts came from anywhere.
Which situation is most common?
This page does not establish which is most common. The launch-spike scenario is useful to examine because it separates one-time attention from repeatable demand.
What if two of these are happening at once?
If two overlap, examine the reach issue first: everything else is easier to read once arrivals come from a route that repeats.
Should I ever change the product in the first month?
When something is broken, immediately. When something is disappointing, ask first. The distinction is whether the thing fails or whether nobody wants it, and those look identical on a chart.
