How-to guide · Pricing

How to price a project you've never done before

By the WorkKitly Team · Last updated: November 2026 · 7 min read

You need not have done this exact project before to price it well — you need a decomposition. Break the job into the pieces you have done, price those from your rate and records, bound the new pieces with a range and quote its pessimistic end — then choose a payment structure that protects you where the estimate can't. Experience prices the familiar bulk of the job; structure, not hope, carries the rest.

Break the job into familiar parts and new parts

Almost no project is entirely new. A website with an integration you have never touched is still mostly familiar pages, setup and handover, plus one unfamiliar piece; a rebrand in a new industry is the same design process with different references. The genuinely new part of a "never done it" job is usually two or three components. The rest is your normal work.

So the first move is a list in two columns. Left: every part you have done before and can size from experience. Right: parts you have never done, have only watched someone else do, or cannot yet picture hour by hour. The columns get priced completely differently — the left from your records, the right with a range and a protective structure. Blend them into one guess and new work gets underpriced, because the comfortable parts make the whole job feel familiar.

One rule for the list: a part you cannot yet describe step by step belongs in the right column, however small it sounds. "Connect the payments thing" stays there until you can write down how it is done.

Price the familiar parts from your own numbers

The left column is priced the way you price repeat work: honest hours for each part, from your last similar jobs rather than your fastest day, times your hourly rate. If you do not have a rate you trust yet, build one first with the hourly rate calculator — every figure after this point stands on it. The project price calculator runs this build-up and shows the floor underneath the quote — the line that matters if the client negotiates.

Resist the optimism discount: the temptation to shave the hours because you will be faster this time, or the client seems easy. Don't. Honestly priced familiar parts are the solid ground the quote stands on, and the new parts will draw on your contingency heavily enough by themselves. If a part took six hours on the last two jobs, it goes in at six.

Bound the new parts: research, a range, the pessimistic end

A part you have never done cannot honestly be estimated as a single number — anyone who offers one is guessing. It can be bounded. Start with an hour or two of research: read the documentation, watch the task actually being performed, check what the tool demands. Better still, run a spike: a small, time-boxed test — build the tiny version, connect the test account, export the sample file. A two-hour spike turns "somewhere between easy and impossible" into a range you can defend.

Then estimate each new part as a range, best case to realistic worst case, and quote from the pessimistic end of it — not the absolute worst case, where everything fails at once (when that is the honest picture, the answer is a discovery phase, covered below), but the plausible end where the ordinary surprises happen. Here it is on one fictional job: a web developer asked for her first online shop — store pages and CMS work familiar, the payment and stock integration new.

Part of the jobHoursAt $90/hr
Discovery and setup (familiar)3$270
Store pages and product templates (familiar)9$810
CMS and content setup (familiar)4$360
Testing and handover (familiar)2$180
Familiar parts, totalled18$1,620
Payment and stock integration (new — range 6 to 14 hours after a spike, quoted at 12)12$1,080
Total quote30$2,700

Read the range honestly in both directions. If the integration lands at its optimistic six hours, the job takes 24 hours and still pays $2,700 — the $540 gap is the risk premium, earned for carrying an unknown. If it lands at fourteen, the job takes 32 hours and the quote comes up $180 light: two hours short, an annoyance rather than a disaster. The bad case is survivable and the good case fairly rewarded — the reverse of what quoting the hopeful end arranges.

Choose the structure that carries the risk

The estimate is only half the protection. The other half is the payment structure — how the quote is framed, split and sequenced. Four structures cover nearly every first-time job:

Fixed fee, with the buffer inside the number. The method above as one price, with the range and the pessimistic end already inside it, invisible to the client. Right when the new parts are bounded and modest next to the familiar whole. Our guide to fixed fee vs hourly pricing covers how buffers scale with repeatability and how to keep a fixed fee safe from scope creep.

Paid discovery first. When a new part cannot even be ranged — you honestly cannot say whether it is a day or a month — do not quote the project yet. Sell a small, fixed-price discovery phase instead — a few bounded hours at your rate, ending in a deliverable: a written plan and a real estimate for the main job. It is the honest default for genuinely unknown work: the client pays a little to learn what the job is; you quote the rest from knowledge, not theatre.

Milestones, with the unknowns early. Split the price so the unfamiliar part is delivered, and paid for, early — while familiar work is still ahead. On the shop above: an $810 deposit, $1,080 when the integration demonstrably works on the test store, $810 at launch — the risky piece proved and settled before the polish weeks begin. The deposit calculator splits any price into stages, and our guide to milestone payments covers how to size and sequence them.

Hourly, with a not-to-exceed cap. Some clients insist on hourly. Then hourly it is — with a ceiling agreed in advance ("billed hourly at $90, not to exceed $3,000 without written approval") and a check-in where you report hours against progress. The client keeps the honesty they asked for; the cap answers the open-endedness they fear, and trouble surfaces while budget remains to fix it.

Put the unknowns in the scope, not just in your head

Everything above lives or dies on one page: the scope. The new parts, their assumptions and your chosen structure belong in writing, in the document the client agrees to — build it with the scope of work generator if you lack a template. The single most valuable sentence in it is the assumptions line: "This price is based on the assumption that the payment provider's standard checkout meets your needs. If a custom integration turns out to be required, that part is re-estimated and the price reopens before the work begins." Agreed in advance, it converts your worst case from an argument into a procedure — the price reopens because the signed document says it does.

Name the exclusions beside it: the data migration you are not doing, the product entry the client is doing, the third-party fees the client pays. On new work they are the boundary between the job you priced and the larger job the client may be picturing.

Red flags: when no structure can save the quote

The client refuses discovery and demands a firm fixed price on undefined work. That combination hands you the entire risk of their unknown job, at a number they can hold you to. The professional answers are a discovery phase, hourly with a cap, or no. A firm price on genuinely undefined work is not a quote; it is a donation with paperwork.

"It should be simple," from someone who has never done it. Simplicity claims about work the speaker cannot perform are not information — they are a hope about the price, dressed as a fact about the job. Treat every "should be simple" as a right-column item until your research says otherwise.

After delivery: record the real hours

When the job ends, write down what the new part actually took, next to the range you quoted. Five minutes, no ceremony: estimated 6 to 14, quoted 12, actual 9. That habit turns a first-time job into priceable experience: the part joins your left column with real numbers attached, and the second quote of this kind is ordinary. Skip it, and unfamiliar jobs stay unfamiliar forever, priced from vibes each time.

Questions

Break the job into parts and price the two kinds differently. Familiar parts get honest hours from your past jobs, multiplied by your rate. New parts get researched first — documentation, examples, a small time-boxed test — then estimated as a range and quoted at its pessimistic end. Then choose a structure that protects the estimate: a fixed fee with the range already inside it, a paid discovery phase when the work cannot be ranged at all, milestones that pay for the unknown part early, or hourly with a not-to-exceed cap. Write your assumptions into the scope, so a false one reopens the price by agreement rather than by argument.
Not as a standing rule. The client buys an outcome and a carried risk, and the quote already makes room for your level: new parts are priced at the pessimistic end of a researched range, which is where the learning cost lives. If you choose to absorb some learning yourself — the initial reading and research, say — make it a deliberate one-off for this job, never presented as your rate. A discounted first job the client believes is your normal price has not won full-value work; it has set an anchor you will argue against on every later quote.
It depends which way it fails, and the structure decides how much it matters. Inside the range you quoted, pessimistic-end pricing means the overrun is small or already covered. If an assumption underneath the estimate proves false — the data is in a different format, the platform cannot do what its brochure implied — the assumptions line in your scope does its work: that part is re-estimated and the price reopens before you continue, as agreed in writing. If you simply misjudged, finish well, record the real hours, and let the next quote be smarter. The outcome to avoid is silently absorbing a large overrun, which turns one bad estimate into unpaid weeks.
Yes. For work that genuinely cannot be sized in advance it is the standard professional answer, and serious clients recognise it as one. Keep it small and deliverable-shaped: fixed price, bounded hours, and a concrete output — a written plan, the risky piece tested and working, a real estimate for the main job. What it is not is a free consultation renamed: a client who will not pay a small amount to define the job was asking you to absorb the cost of finding out — the estimation risk itself, arriving early.

Related: the pricing toolkit

Set the rate every quote stands on with the hourly rate calculator, build the number part by part with the project price calculator, split it into protected stages with the deposit calculator, and fix the boundary with the scope of work generator. Unfamiliar work stops being a pricing problem once it is a decomposition, a range and a structure.