How to Estimate a Creative Project

Key Takeaways

• Estimate in phases, not as a single number. A single number cannot be defended, adjusted, or partially cut.
• Price in days, not hours. Hours invite a conversation about your speed; days describe the shape of the job.
• The making is usually half the project. Briefing, waiting, revising, managing and delivering are the other half and they get left out.
• Add contingency as a named line, not as padding hidden inside the phases. Hidden padding gets negotiated away invisibly.
• State the exclusions in the estimate itself, because everything unnamed is assumed included.
• Track actuals against estimates on three projects and you will estimate better than any formula can make you.

Estimate the Phases, Not the Job

A single number is indefensible. If a client thinks it is too high, there is nothing to discuss except whether you will lower it, which is a bad conversation with one outcome.

Break the project into phases you can name. For a video: brief and concept, pre-production, shoot, edit, revisions, delivery. For a design project: discovery, concept directions, refinement, production, handover. For a campaign: strategy, concepting, asset production, adaptation, delivery.

Price each phase separately. Now the estimate is a structure. If the client's budget is short, you have a real conversation: one concept direction instead of three, one shoot day instead of two, a shorter discovery. That is a scope conversation rather than a discount conversation, and it protects both your margin and the work.

Phases also expose the parts people forget to buy. A client who sees a "revisions" line for the first time learns that revisions are work. A client who sees "delivery and formats" learns that eleven cutdowns are not free. Neither of those is obvious from a single figure.

Keep it to five or six phases. A twenty-line estimate reads as an invoice and invites line-by-line negotiation of things that are not separable.

Price in Days

Hours are the wrong unit for creative work and using them creates a specific problem.

An hourly figure invites the client to think about your speed. If you say twelve hours, someone will privately wonder whether a faster person would take eight, and the whole conversation shifts to your efficiency rather than the size of the job.

Days describe the job instead. A day is a recognisable unit of professional attention. Half-days work for small pieces. Nobody argues about whether a concept direction should take 5.5 hours instead of 6, and everyone can reason about whether it is a one-day or a two-day job.

Days are also more honest about how creative work actually happens. The edit does not take six continuous hours. It takes a day with a break in it during which the problem gets solved in the shower, and that is not billable padding, it is how the work gets good.

Keep an internal day rate and do not publish it. You need it to build estimates. The client should see phases and a total, because a published day rate turns every scope conversation into a rate negotiation.

The Half You Forget

When people picture a project, they picture the making. The making is usually about half of it.

Briefing and alignment. Kickoff calls, reading the material, the back-and-forth to establish what is actually wanted. Real hours, always, and almost never estimated.

Waiting. Feedback that takes four days, assets that arrive late, an approval that needs a legal review. Waiting is not billable and it is why calendar time and working time are different numbers. Estimate both, and say so: eight working days across three calendar weeks.

Revisions. Included rounds are work. Two rounds on a video is not a rounding error, it is often a fifth of the total effort.

Project management. Emails, scheduling, chasing, status updates, invoicing. On a small project this is minor. On anything involving more than two other people it is a real percentage, and freelancers absorb it silently because it does not feel like craft.

Delivery. Exports, formats, captions, file naming, uploads, handover documents, the last-minute request for a different aspect ratio. Consistently underestimated because it feels like a formality and it takes half a day.

A rough sanity check: if your making time is X, the project is usually somewhere near 1.7 to 2 times X. That multiplier is not a rule to quote at clients. It is a check on your own estimate before you send it.

Contingency, Named

Every experienced estimator adds slack. The difference between the ones who keep it and the ones who lose it is whether the slack is visible.

Hidden padding gets negotiated away without anyone knowing. You quietly add twenty percent to each phase, the client asks you to trim, you trim what looks trimmable, and you have removed your own buffer while believing you protected it.

A named contingency line survives. Ten to fifteen percent, on the estimate, described plainly: contingency for the ordinary unknowns of a project at this stage. Clients accept this far more readily than people expect, because everyone who has commissioned anything knows projects move.

Scale it to how much you know. A project with a locked script and confirmed locations needs less than one where the brief says "something like the reference, but for our brand". If the brief is vague, say that the estimate is a range and the range narrows once the brief is locked, which is true and also a gentle push toward locking it.

Do not use contingency for scope changes. It covers the known-unknowns of agreed work. New deliverables are new work and priced separately.

Write What It Excludes

An estimate that lists only what is included leaves everything else ambiguous, and ambiguity is resolved in whichever direction costs you more.

Add a short exclusions block. Not adversarial, and clients generally appreciate it because it tells them what they will need to budget separately.

For a video estimate: includes one 16x9 master, two revision rounds, and delivery in the agreed format. Excludes additional aspect ratios, subtitle and caption files, additional language versions, music licensing beyond the agreed library, travel and accommodation, and additional shoot days.

Your exclusions list writes itself from history. Anything that has ever arrived late and unpaid on a similar project goes on it. That list is short and it is the same short list every time.

State the validity period and the assumptions. Valid for thirty days, assumes the brief as of this date, assumes feedback within three working days. Those assumptions are what your timeline depends on, and naming them means a slipped approval is a shared fact rather than your problem.

A Worked Estimate

Abstract advice is easy to agree with and hard to apply, so here is the shape on a real job: a two-minute brand film, one shoot day, small crew, delivered in three formats.

Making time in that table is the shoot and the edit: four days. The total is thirteen and a half. That is the 1.7 to 2 times multiplier arriving in practice, and it is why an estimate built by picturing the shoot and the edit comes in at roughly a third of what the project costs.

Calendar time is a separate number. Thirteen and a half working days across four to five calendar weeks, assuming feedback inside three working days. Say both, because the client is planning against the calendar and you are planning against the days.

What that estimate excludes, stated on the same page: additional aspect ratios beyond the three agreed, subtitle files, additional language versions, music licensing beyond the agreed library, travel and accommodation, and a second shoot day.

Notice what is not in the table: a day rate. The client sees phases, days and a total. The rate is yours.

Get Better By Tracking, Not By Formula

Every estimating rule on this page is a starting point. The thing that actually makes you accurate is boring.

Record what you estimated and what it took. Per phase, roughly, in whatever you already use. Not a timesheet system, a note at the end of each project.

Three projects is enough to see your pattern, and everyone has one. Most people are approximately right on the making and consistently wrong in the same place: usually revisions, or delivery, or the number of stakeholder calls. Once you know your own bias, you correct for it, and that correction outperforms any generic multiplier.

Track the calendar duration too. The estimate that said three weeks and took seven is usually not a work estimate failure, it is a waiting failure, and knowing which one you have determines whether you should change your prices or your process.

Do not do this retrospectively for old projects. Start with the next one. Reconstructed data is guesswork wearing a spreadsheet.

What Tooling Helps

Honestly, very little, and it is worth saying plainly rather than recommending software.

An estimate is a document. A spreadsheet with phases, days, a rate, a contingency line and an exclusions block is a complete solution and most working studios use exactly that. Proposal tools such as Qwilr or Better Proposals make it look better and manage acceptance, which matters if you send many. Project management tools track actuals if you already use one.

The variable that determines estimate quality is whether you record what actually happened, and no tool makes that happen.

Storyflow is our product and it does not do this. It has no invoicing, no time tracking, no rate cards, no proposal builder and no estimating templates, so build the estimate in a spreadsheet or a proposal tool. The adjacent job is upstream: holding the brief, the references and the open questions on one canvas, which matters here only because the biggest source of estimating error is a brief too vague to estimate against. Paid-only during early access, with the Free plan landing before the end of 2026, and anyone a paid member invites to a board joins free now.

The Bottom Line

Estimating is not a pricing skill, it is a modelling skill, and the model most people use leaves out half the project.

Break it into five or six phases you have done before, so the estimate can be discussed rather than only accepted or refused.

Price in days, so the conversation is about the size of the job rather than your speed.

Add the half that is not making: briefing, waiting, revisions, management, delivery. Check your total against roughly 1.7 to 2 times your making time before you send it.

Name the contingency and name the exclusions, because hidden slack gets trimmed and unnamed work gets assumed.

Then record what actually happened. Three projects of honest tracking will teach you more than any formula, because you will find your own consistent bias, and everyone has one.

You do not estimate the work. You estimate the work plus everything that happens around the work.