Job hunt or self-employment? A 90-day test with numbers to decide

Why 90 days, and why run both tracks at once

The Bureau of Labor Statistics publishes how long unemployed people have been looking for work. In September 2026 the median was 11.5 weeks and the mean was 24.8 weeks, and 27.1 percent of unemployed people had been looking for 27 weeks or more. Read it this way: a search of about three months is ordinary, and more than a quarter of searches run past six months.

That makes 90 days a fair length for both tracks. It is long enough for applications to turn into interviews and for a first client to turn into a second. It is short enough that you find out before your savings run dry.

Running both at once matters because they protect each other. The job search keeps a salary in play if the business stays quiet. The business gives you income, a story for interviews and something useful to do with the dread between applications.

It also matches how most people actually start. The Federal Reserve's survey of household finances for 2024 found that 20 percent of US adults did some gig activity in the month before the survey. Of those, 70 percent spent less than five hours a week on it, and only 21 percent called it their main job. Most people start small and on the side, and this test assumes you will too.

Day 0: three numbers to set before you start

Write these down before you send a single application or pitch. Every checkpoint compares against them.

1. Your monthly floor. Rent or mortgage, utilities, food, insurance, phone, transport and minimum debt payments. Leave out anything you would cut in an emergency. This is the amount one of the two tracks eventually has to cover.
2. Your runway in months. Savings plus any severance, divided by the floor minus income you can count on, such as a partner's pay or unemployment benefit. If the runway is under three months, the test still works, but the self-employment track has to bring in cash within 30 days, and the job track should include part-time and temporary roles from the start.
3. Your hourly floor. Self-employed income carries costs that a wage hides. Self-employment tax in the US is 15.3 percent, charged on 92.35 percent of net profit, which works out to about 14.1 percent of profit before any income tax. You also pay for your own tools, health cover and the unpaid hours spent finding work. Put your monthly floor into the RichTactic freelance hourly rate calculator and treat its answer as the least you can charge.

If you were laid off, sort out two more things on day 0. If you claim unemployment benefit, find out how your state treats self-employment earnings before you invoice anyone, because reporting rules differ by state. RichTactic's guide to working while on unemployment explains how people report earnings without losing the claim. And if you lost job-based health insurance, HealthCare.gov gives you a Special Enrollment Period of 60 days from the loss to pick a Marketplace plan. Self-employed people with no employees buy through the same Marketplace, so the plan can follow you whichever track wins.

The job-hunt track: what to do and what to count

Keep the search going, but make it measurable. Each week:

• Send 10 applications tailored to the posting, with your resume wording matched to the job description.
• Send 5 messages to people who work where you want to work, each asking a specific question about the team or the role, never asking for a job.
• Log every application in a spreadsheet: date, company, role, where you found it, and the outcome when it arrives.

The number that matters is your response rate, which is interviews divided by applications. Count a phone screen as an interview. If you have sent 40 applications by day 30 and have no screens, the cause is usually the resume or the roles you are aiming at, and both can be fixed. Check how your credentials appear with how to list certifications on a resume, and read why certifications get filtered out at screening before you decide the market has rejected you.

If you have already been out of work for months, what to do after six months unemployed covers which explanations recruiters accept. If you have a severance package, the severance runway maths shows how to stretch it across a test like this one.

The self-employment track: one offer, one price, 50 asks

Pick one service that uses something you can already prove: the certification you passed, the job you used to do, or the thing friends already ask you to help with. Write it as one sentence a buyer understands, with a scope and a price. "I set up and secure home and small-office Wi-Fi, printers and backups for a fixed price per visit" is an offer. "IT consulting" is a category, and buyers do not know what they would get.

Then count asks instead of hours. In the first 30 days, make 50 direct asks: a message to someone you know, a LinkedIn message to a stranger in your niche, a conversation with a local business owner. A social media post counts as one ask, however many people see it. Landing your first freelance client has the scripts and the order to work through them.

Leave Upwork and Fiverr out of the first 30 days. AI took much of their entry-level work (writing freelancers on Upwork earned 5.2 percent less a month after ChatGPT, according to a study in Organization Science), and the jobs that remain go to freelancers with reviews, so a month there tests the platform more than your offer. If you add one later, build its fees into your price: Upwork charges freelancers 0 to 15 percent per contract, each Connect costs $0.15, and Connects are not returned when a client turns you down. Fiverr credits sellers with 80 percent of each order. RichTactic's guide to why Upwork and Fiverr stopped working for beginners covers what to do instead.

Expect little money in the first month. The same Federal Reserve survey found that 49 percent of people doing gig work wished the pay was more consistent. Month one tests whether anyone pays at all, and the amount comes later. How much side hustles actually pay breaks down realistic income by business model, which helps you judge whether your day-60 number is on track.

Checkpoints at day 30, 60 and 90

Score both tracks on the same day and write the numbers down without rounding up.

• Job track: 40 or more tailored applications sent, and a response rate you can calculate.
• Self-employment track: 50 asks made and at least one paid job of any size. If 50 real asks produced nothing, the offer, the price or the audience is wrong. Change one of them and run the next 30 days.

• Job track: at least one interview. With none after 80 applications, rework the resume and the target roles before day 90.
• Self-employment track: revenue in the month equal to at least a quarter of your monthly floor, and at least one client who came back or referred someone.

• Job track: are you in final rounds or holding an offer?
• Self-employment track: is your monthly net profit, after platform fees and about 14.1 percent set aside for self-employment tax, at half your floor or more, with work already booked for the next month?

These thresholds are the rules of the test, and they are deliberately strict. A quarter of the floor at day 60 and half at day 90 means the business is growing fast enough that another three months could close the gap, if your runway allows it. They do not predict what the business will earn.

Reading the result: four outcomes

• Job offer, weak business. Take the job. Keep any client who pays well as evening work, after checking your employment contract for moonlighting and conflict-of-interest rules. The months you spent are now a resume entry instead of a gap, and how to put self-employment on a resume shows how to write it.
• No offer, business on track. Extend the self-employment track to six months if your runway covers it, and narrow the job search to roles you would genuinely leave the business for. Register properly and start quarterly tax payments, both covered in first-year self-employed taxes.
• Both strong. Ask the employer whether a contract or part-time arrangement is possible. If it is not, accept the job and raise your prices for the clients you keep, so that two or three of them are worth your evenings.
• Both weak. This is the hardest result. Look at which track failed first. A response rate near zero points to the resume or the targeting. Asks with no buyers point to the offer. Fix the single weakest thing and run another 60 days, taking whatever part-time or temporary work your runway requires in the meantime.

Whichever outcome you get, the decision rests on your own numbers.

What government data says about working for yourself

Three sets of figures are worth knowing before you commit past 90 days.

Survival. The BLS follows private-sector business establishments from the year they open. Of those that opened in the year to March 2024, 77.9 percent were still operating a year later. Of those that opened in the year to March 2020, 51.4 percent were still open five years later. These figures cover businesses with employees, so a one-person service with no payroll is not counted, but they are the best public picture of how often new businesses last.

Pay. In the BLS survey of work arrangements from July 2023, full-time independent contractors had median weekly earnings of $949 on their main job, against $1,132 for full-time workers in traditional arrangements. The BLS notes that part of the gap reflects differences in who does contracting and in which occupations.

Preference. In the same survey there were 11.9 million independent contractors, 7.4 percent of everyone employed. Of them, 80.3 percent preferred their arrangement and 8.3 percent would rather have had a traditional job.

None of this predicts your result. It does tell you that working for yourself is common, that it often pays less per week than a salary, and that most people who do it would not swap back.