Price is how you assign risk
Hourly billing says risk shared minute by minute: scope can move and the meter reflects it. Fixed price says risk shifts to the seller unless change orders exist: scope must be stable or the seller eats overruns.
Neither is morally better. They fit different projects and different client relationships. Choosing wrong is how friendly projects turn into silent resentment on both sides.
Choose hourly when
Scope is exploratory. Research spikes, audits, "figure out why performance collapsed" work.
The client changes mind often. Product discovery with a founder who pivots weekly.
You are new to the domain. Unknown unknowns dominate; pretending otherwise is a fixed-price trap.
Trust is still forming. Hourly plus frequent demos lets both sides adjust without renegotiating the whole deal.
Hourly pairs well with transparent time notes and outcome summaries so the client sees value, not just hours (billable versus working hours).
Choose fixed price when
Deliverables are crisp. "Migrate these twelve pages to the new CMS with approved designs."
Inputs are client-controlled and dated. Content, assets, and approvals arrive on schedule or the timeline slips by contract.
You have done this before. Similar projects give you a distribution of hours, not a fantasy single number.
The client wants budget certainty. Finance teams often prefer a cap even if the cap is higher than optimistic hourly.
Fixed price demands buffer in money and explicit change-order rules (how much buffer). When the brief grows mid-project, use a calm change-order protocol instead of absorbing free extras.
Hybrid models that work
Fixed discovery, hourly build. Paid workshop produces a spec both sides sign; implementation hourly or fixed per milestone.
Hourly with a cap. "Up to forty hours at X; stop and re-estimate if we hit eighty percent of cap."
Fixed per milestone, hourly for support. Launch package fixed; post-launch retainer hourly with a monthly bucket.
Hybrids fail when the boundary between phases is vague. Write the deliverable that ends phase one.
Red flags for fixed price
"We will know it when we see it" design. Third-party APIs nobody has integrated. Stakeholders who have not agreed on one owner. "Quick project" with no written acceptance criteria.
Any of those without a paid discovery phase should stay hourly or very small fixed slices.
Red flags for hourly-only
Client asks for detailed estimates every week but refuses to pay planning time. Scope never shrinks but budget must. You absorb meeting sprawl without billing because it feels petty.
Hourly is not permission to skip estimates. It is permission to re-estimate when reality diverges. See estimate freelance work with uncertainty.
Outcomes still beat arguing about hours
Clients buy results. Hours are one way to price them. When you can name an outcome clearly, price the outcome and keep hours internal in your rate model (calculate a freelance hourly rate).
Conversation script
Ask four questions before proposing a model:
- What does done look like in one paragraph?
- What inputs do you owe me, and by when?
- Who can approve scope changes in one reply?
- Is exploration part of success or failure?
Exploration-heavy answers point hourly or fixed discovery. Crisp answers with stable inputs point fixed milestones.
Example: bounded scope fits fixed
Example: "Migrate five blog posts to new CMS template" with acceptance criteria per post can be fixed if revisions are capped in writing.
Example: discovery stays hourly
Example: "Improve checkout conversion" without analytics baseline is hourly with weekly checkpoints until metrics exist.
Switch models without shame
Projects that grow past the original shape need a new agreement, not silent resentment. "This outgrew fixed scope; here is a change order or we move to hourly for phase two" is normal business.
Document the handoff between models. Projects often start hourly for discovery and shift fixed for delivery. Write the trigger: "Fixed quote follows signed spec v1." Without a trigger, clients hear fixed numbers early and you hear hourly flexibility late.
Uncertainty clause
State what happens when scope expands: change order, hourly overflow, or re-estimate. Silence becomes discount.
Revision rounds belong in the contract shape
Fixed price without a revision cap invites infinite polish. Hourly without a scope narrative invites client anxiety. Write how many feedback cycles are included and what triggers a change order.
Maintenance contracts are a third shape
Retainers reward steady availability; fixed projects reward bounded scope. Mixing them without naming which mode you are in creates scope arguments.
Trust and record-keeping
Hourly needs honest time records you would show if questioned. Fixed needs honest scope records you would show if questioned. Both need the same integrity; only the unit of account changes.
When to raise your freelance rate when utilization stays high and deliverables are strong, not when you are desperate.
When clients ask for fixed price because it feels safe, explain what safety costs: explicit scope, change orders, and fewer concurrent demands on your calendar. Safety without boundaries is risk transfer onto you.
Pick the model that matches scope stability and write the assumptions down. Price follows risk; risk follows clarity.
Myth: fixed price is always safer for the client
Fixed price feels safe because the number is on the cover page. Safety requires scope frozen on the same page. When scope is not frozen, fixed price transfers risk to the seller, who either eats hours or turns bitter. The client experiences that as rushed work, change-order fights, or quiet scope cuts.
Hourly feels risky because the meter runs. Risk drops when you pair hourly with weekly demos, written caps, and stop rules at eighty percent of budget. The client sees trajectory early instead of discovering overrun at the end.
Scenario: the pivoting founder
Example: Week one, hourly discovery produces a spec and a clickable prototype. Week two, the founder pivots after three user calls. Week three, a new flow replaces half the prototype.
Under fixed price for "the app," every pivot is a negotiation. Under hourly with a documented cap per phase, pivot is expected cost of discovery. Phase two fixed quote arrives only after sign-off on spec v2.
The hybrid respects that early work is learning and later work is manufacturing.
Scenario: finance needs a cap
Example: A finance lead requires a budget line before fiscal year end. They cannot approve open-ended hourly, but engineering scope is stable: migrate reports, no new metrics.
Fixed milestone with acceptance tests per report fits. Buffer sits in price, not in vague optimism. Change orders cover new metrics explicitly.
Match the model to the buyer's fear. Finance fears uncapped spend. Founders fear slow learning. Write the contract so each fear is addressed without lying about scope.
Price maintenance across models
Hourly rate reviews and fixed-price template updates should happen on the same calendar rhythm. A fixed package you sold two years ago at old scope assumptions is an hourly rate you forgot to raise.
When you renew a fixed client, recompute hours from recent actuals plus buffer, not from nostalgia for the first quote they accepted.
What to put in writing before signature
Scope paragraph, exclusion list, revision cap, change-order trigger, payment schedule tied to milestones, and who counts as approver. Missing any one of those is choosing conflict later.
The model is secondary to clarity. Hourly with muddy scope still hurts. Fixed with crisp acceptance tests can be generous to both sides.
Closing choice in one sentence
If scope, approver, and inputs are written and stable, fixed milestones can work. If any of those three is moving, hourly with caps and checkpoints protects both sides until clarity arrives.
Say that sentence out loud in the sales call before anyone sees a number. The model choice becomes obvious without a pricing theology debate.
Write the chosen model on the first page of the statement of work with the scope paragraph. Later pages can detail milestones; page one prevents "I thought this was fixed" arguments in week three.