Published

The Option Nobody Asks About

Some choices stay hidden because asking about them costs standing. Whether they stay hidden is partly an interface decision.

A job offer lists a salary, a start date, a title, and a signature line. It does not list four days a week.

Nothing forbids asking. But a new hire who asks about part-time work before the first day is telling the employer something about their commitment, and they know it may be heard that way. So the question goes unasked and the option is never priced. Years later they may say they never considered it. That is accurate, and it was never a decision.

That gap, between what someone has been doing and what they believe they are allowed to consider, is where a lot of product design operates without saying so. Most interfaces narrow it: they ask the question the business wants answered and offer the options the business sells. An interface can also widen it. It can make an alternative visible, make the question ordinary, make the consequences calculable, and stop before the conclusion.

Each of those is a decision about a screen. Here is what each looks like in a general case, and in something I built.

Show the option that isn’t on the menu

A mortgage calculator asks how much house you want. Every answer it can give is a house. A calculator that also prices renting for five more years and investing the difference has not argued for renting. It has put a row on the screen that the first question excluded.

Show the calculator

Mortgage calculator

How much house?Enter a price
  • A bigger house
  • The house you asked about
  • A smaller house
  • Rent five more yearsand invest the difference
The same calculator with one row added. The row argues for nothing; it puts renting where the user can compare it.

The decision is what the screen contains before the user asks for anything. People read the options in front of them as the whole set, so the set is the design.

I built Ambit for a different subject: the AI agents a person runs and the tools each one can reach. It reads their configuration files and draws them as a tech tree. Beside what the setup can already do, it draws the edge: for each next step, what that step would unlock. Nobody has to want the next step for it to be worth seeing.

Make the question ordinary

Take two ways to ask about reduced hours on a benefits page. One is a checkbox: “I’d like to discuss a reduced schedule.” Ticking it is a disclosure, made alone, under your name. The other is a table every employee sees, with rows for full time, four days, and three days, and pay and days off filled in for each. Nobody ticks anything. Reading the third row says nothing about you.

Ask about reduced hours

Schedule request

Name
Manager

Your manager will see this, with your name on it.

Schedules

SchedulePayDays off a week
Full time100%2
Four days80%3
Three days60%4

Every employee sees the same table.

The same question, asked two ways. The table costs more to build and maintain, and nobody has to volunteer anything to read it.

What changed is who has to raise it. When asking costs standing, the interface can ask for everyone, so that no one person has to.

Tenant Tools applies that move to a rental building. A tenant who reports a broken heater alone is one complaint from one identifiable unit. The app lets a neighbor join an existing issue with “me too,” without writing anything, and the joint letter asking for repairs carries a count of households, not their names. The heater is the same. Raising it is now something the building does.

Price the alternative

“Should I go part time?” is a question about who you are. “At four days, what happens to my take-home pay and the year my loans are paid off?” is arithmetic. Arithmetic is easier to look at, because the answer says nothing about the person asking.

Ask about part time

Should I go part time?

YesNo
  • Full time3 years to spare
  • Four days1 year to spare
  • Three days2 years past

Years until the loans are paid off, on each schedule.

The same decision, asked as a verdict and as arithmetic. The person sets the target; the screen only measures against it.

The move is to turn the identity question into the numbers it depends on, run on the person’s own figures, and to show what would have to change for the answer to change.

Whether does this for startup leaders deciding when to hire, spend, or raise. Every Monday it scores U.S. funding conditions and, for each of five areas, gives an action, a line that would pause it, and a line that would reopen it. Running it taught me that people use those three answers, not the forecast underneath them. “Should we hire?” becomes “what would have to be true?”, and that version can be checked every week.

Stop before the conclusion

A tool that ends with “you should take this offer” has answered a question nobody asked it. It has also shrunk the user’s next move to agreeing or disagreeing with software, which is a smaller decision than the one in front of them.

End the review

Non-compete, section 7.2

Physician will not practice within 15 miles of any Practice location for two years after this Agreement ends.

Covers the hospital across town.

You should sign.

Sign

Pro-employer reading

Protects the practice’s patient base.

Pro-physician reading

Rules out the nearest other job.

Your call

AcceptPush backAsk a lawyer
The same clause, ending two ways. Both state the fact; only one takes the decision.

At Andwise, the contract tool I built for physicians drew that line in the product. It highlighted the clauses worth arguing about, explained each in plain language, and showed the physician-friendly and employer-friendly readings side by side. It did not advise. Accepting a clause, pushing back, or calling a lawyer stayed the physician’s call.

Grade the facts, not the preference

Refusing to advise has a cost: a tool that grades nothing is easy to trust and easy to stop opening. The answer is to be exact about what gets graded. A tool can say the non-compete radius covers the hospital across town, and that the tail-coverage clause assigns the premium to nobody. Those are facts, and it should state them as bluntly as a good attending would. Whether to stay in that city is a preference.

The same moves can steer

An option can be made visible because it carries the margin. A question can feel ordinary because every user is being asked to upgrade. A price can be framed so that one row looks reckless.

The test I use is whether the interface would show an option that earns the business nothing. Andwise kept a free directory of more than 70 financial advisory firms and took no payment for a listing. It was a small page.

Success, for this kind of design, rarely shows up as a conversion. It shows up as a person who kept the job they had, having looked at the others.

Case study Reading a physician’s employment contract before they signed it Andwise · 2022–2024

Revised