Skip to content

Roadmap

From Code to a Product That Runs Without You

Small software, sold to people who already told you what to build

Starting point
You can build software, or can learn to, and you have no product
Where it takes you
Recurring revenue from software you own and never had to raise money for
Realistic duration
18–36 months
Capital required
No capital · $0
  1. Phase 1Months 1–6

    Get paid to learn a domain

    Goal: Be inside a real industry long enough to see what it is missing

    What to do

    • Take client work in one industry rather than whatever arrives, so the second project teaches you more than the first
    • Write down every problem the client works around with a spreadsheet, a manual process or an intern
    • Ask what they already pay for and what it costs — the price of the alternative is your price
    • Refuse the temptation to build a product for an industry you have only read about

    Techniques in this phase

    Milestone Three projects in the same industry and one problem that appeared in all three

  2. Phase 2Months 4–10

    Ship something small enough to finish

    Goal: Complete and publish one thing, because finishing is the skill being learned here

    What to do

    • Build the smallest useful version of the problem you found — a plugin, a script, a template, a single-purpose tool
    • Set a deadline in weeks and cut scope to meet it rather than moving the date
    • Publish it where the people with that problem already are, and answer every support message yourself
    • Read what people ask for. The requests are the specification for the product that follows

    Techniques in this phase

    Milestone One tool live, a hundred people using it, and a support inbox you answer personally

  3. Phase 3Months 8–16

    Charge from the first day

    Goal: Prove people pay, before building anything that takes a year

    What to do

    • Price it from day one. A free product with lots of users teaches you nothing about whether it is a business
    • Sell to ten customers by hand, one conversation at a time, before automating anything about signup
    • Ship weekly to the people paying and ignore every request from those who are not
    • Do not add a second product. The first one is not finished, it is barely started

    Techniques in this phase

    Milestone Ten customers paying every month, none of whom are friends or former clients

  4. Phase 4Months 12–24

    Solve distribution, which is the actual job

    Goal: Make new customers arrive without you finding each one

    What to do

    • Pick one channel and work it for six months — search, integrations, a community, or partnerships. One, done properly
    • Write the things your buyers search for while looking for the workaround they use today
    • Get listed wherever the software they already use lists compatible tools
    • Measure how many customers arrived without you doing anything that week, and grow that number specifically

    Techniques in this phase

    Milestone Half of new customers arriving from a channel that keeps working when you stop

  5. Phase 5Months 24–36

    Remove yourself, or sell it

    Goal: Turn a product you operate into an asset that has value without you

    What to do

    • Write the documentation and the onboarding that answer the questions you keep answering
    • Hand over support first — it is the most interruptive and the easiest to hand over well
    • Keep revenue, churn and costs in a form someone else could read in ten minutes
    • Decide honestly whether you want to run this for a decade or sell it, and get a valuation either way

    Techniques in this phase

    Milestone A month where the product gains customers, ships fixes and handles support without you doing all three

Why this does not start with an idea

Ideas are not the constraint. Every developer has a list, and almost all of it is software for problems nobody has confirmed out loud.

Client work is unglamorous and it is the fastest legal way to be paid while looking at the inside of an industry. You see the spreadsheet holding the whole operation together, the process everyone complains about, the tool they hate but renew every year. That is worth more than any amount of ideation, and it pays while it happens.

The rule underneath the first three phases is the same one: never build anything large for a problem you have not already been paid to look at.

Distribution is the job

Most technical founders build a good product and then discover the actual work had not started yet. Software is close to free to copy, which means the hard part was never making it — it is being found by the two thousand people who would pay for it.

That is why phase four is as long as the building phases and gets its own milestone. A product with one channel that works is a business. A better product with no channel is a side project with a maintenance burden.

The channels here are slow and durable on purpose: search, integrations, partnerships, communities. Each one keeps producing customers on weeks when you do nothing, which is exactly what paid ads and posting do not do.

Where people stall

Rewriting instead of selling. The framework is wrong, the code is untidy, the design needs work — all true, and none of it is why nobody is paying.

Free forever. Ten thousand free users feels like validation and is not. Price it and find out in a week what a year of free would never tell you.

Building for developers. They are the customers you understand best, expect the most for free, and churn fastest. The unglamorous industry from phase one pays better.

Waiting for the product to be finished. It never is. The first ten customers should arrive while it is visibly incomplete, because their complaints decide what to build.

Other roadmaps