Estimate vs quote, and which number sticks.
Send a range and the client keeps the bottom of it. That habit costs more than the legal difference between the two documents ever will, and here it's priced out across an eight-week project where the argument at the end runs to $6,900.
An estimate is your approximate price for work that hasn't been described closely enough to price properly, usually written as a range. A quote is one fixed number attached to a scope both sides have already read. The legal difference is that an accepted quote settles the price and an estimate leaves the final invoice open, which every article on this subject will tell you. The commercial difference is quieter and costs more, because a client who receives a range remembers the bottom of it, and from that moment on the bottom of your range is the budget you're working against.
That's the part worth planning around. Below is one eight-week project written both ways, with the same reasonable request landing in week five of each version, so you can see where the money goes and where the written record either exists or doesn't.
What's the difference between an estimate and a quote?
An estimate is an approximation offered before the scope is settled, and a quote is a fixed price offered after it is. Everything else follows from how much you knew at the moment you sent the number. An estimate leans on comparison, meaning you look at three past projects of roughly this shape and say the new one probably lands somewhere in that band. A quote leans on description, meaning you've written down what you're building and priced the parts.
Most of the pages that rank for this question are written for the trades, where a plumber prices a bathroom against countable materials and hours. Agency work behaves differently, because your materials are a font license and a couple of software seats, and the variable that actually moves the total is how many times somebody changes their mind about the thing you're building. That's why the estimate vs quote choice matters more for an agency than the same choice matters for a contractor. The contractor's uncertainty gets resolved by opening a wall, while yours gets resolved by writing the scope of work well enough that the request in week five is recognizably outside it.
There's a third document people fold into this comparison, and it does a different job. A proposal argues for the approach and the value before either number appears, which is why a quote and a proposal aren't interchangeable even when they carry the same total at the bottom.
Is an estimate legally binding?
Generally it isn't, because an estimate is treated as an informed opinion about price rather than an offer a client can accept into a contract, so the final invoice can land above it without anyone having broken a promise. A quote sits on the other side of that line, because a client who accepts a written quote has accepted a price, and you're expected to deliver the described work for that number even if it takes longer than you thought.
Most comparisons stop right here, and stopping here is what makes them useless. Not binding doesn't mean not consequential. The client who received your estimate has already forwarded it to whoever controls the budget, that person has written a number into a spreadsheet, and the spreadsheet is now the reality you're negotiating against. Whether a court would enforce the number is a question nobody in the room is asking. The question they're actually asking is why the invoice doesn't match the email. Exact wording and local rules do change the legal answer, so have your own estimate and quote templates read by a lawyer in your state before you rely on them, and treat what follows as commercial advice rather than legal advice.
The document that does carry weight is the one both sides sign, which is why the relationship between what you proposed and what you contracted decides more disputes than the pricing label ever does.
How far above an estimate can you actually invoice?
In several regulated consumer trades the answer is exactly 10 percent, written into statute. California Business and Professions Code section 9884.2 caps an auto repair bill at 10 percent over the written estimate with no exception unless the customer authorizes the excess in writing before the extra work happens. Virginia's section 59.1-207.3 uses the same 10 percent ceiling, Maryland forbids charging more than 10 percent over a written estimate without consent, and Minnesota expresses it as a bill that can't exceed 110 percent of the estimate unless the shop issues a revised estimate and gets approval.
None of those statutes covers a business to business design or development engagement, and you shouldn't quote them at a client as though they did. They matter for a different reason. That 10 percent number has leaked into how ordinary people think about estimates, so the client who asked you for a ballpark is carrying an unspoken assumption that a ballpark means the real figure will land within about a tenth of it. Send a range of $24,000 to $32,000 and a fair number of clients hear a ceiling somewhere near $26,400, which is the low end plus the tenth they believe they're entitled to.
What does the same $27,600 project look like as an estimate and as a quote?
Take Sablewood Analytics, a reporting tool for property managers whose signups keep stalling on the first run through the product. They want the onboarding rebuilt and a set of in-app help screens to go with it, and on the first call nobody can say how many screens that is. So the first number is an estimate, and it reads like this: projects of this shape usually land between $24,000 and $32,000, with the spread depending on how many onboarding screens survive the audit and whether the help system needs new illustration.
Two weeks later the audit is done, the flows are drawn, and the same engagement gets quoted at $27,600 fixed:
- Discovery and onboarding audit: $4,200
- Onboarding flow, 12 screens at $950 each: $11,400
- In-app help system: $6,800
- Design system updates and handoff: $5,200
The two documents describe the same work and land in the same neighborhood, and they behave nothing alike from here. The estimate names a band and stays deliberately silent about what's inside it, which is the honest thing to do when you don't know yet. The quote names a price and, because it had to name a price, was forced to name 12 screens, one help system and one round of design system updates. That forced description is the whole benefit, and it's the reason the per-screen rate exists at all. When something new arrives, there's already a unit price to attach to it.
What happens when week five adds work to each version?
In week five Sablewood decides enterprise accounts need their own onboarding path, which is six more screens. Under the quote that request has an obvious price, because six screens at the $950 rate the quote already established is $5,700, and the engagement moves from $27,600 to $33,300. The client sees the arithmetic, approves it in writing before anyone opens a file, and the record of that approval sits with the project rather than in somebody's inbox. A written change request takes about four minutes to raise when the unit price is already agreed.
Under the estimate there's nothing to change, and that's the problem. No fixed number was ever set, so no formal revision is triggered, and six screens of work quietly get built because they're obviously part of making onboarding good. The conversation about money happens once, at the end, when the invoice arrives. If you bill the real figure of $33,300, the client points at the $32,000 top of your range and asks why you went past your own number. Settling at $32,000 costs you $1,300 and a slightly colder relationship. If the client anchored on the low end instead, you're arguing against roughly $26,400, and the gap you're trying to close is $6,900 on a project that was never worth more than $33,300 in the first place.
Neither outcome involves anybody behaving badly. The client genuinely believes they asked for something small, you genuinely believe you did $33,300 of work, and there's no document in between that says which of you is right. That's the same mechanism behind almost every dollar scope creep costs an agency over a year, and the estimate is what removed the checkpoint where it would have been caught.
Should an agency send an estimate at all?
Yes, but only for one job, which is finding out whether the client's budget and your price live in the same universe before either of you spends real time. Used that way an estimate is efficient and honest, and refusing to give one makes you look evasive on a first call. The failure is letting the estimate be the last number before the work starts, because a range was never built to carry a project and it can't be enforced, revised or approved.
Three things make an estimate safe to send. Keep the range genuinely wide, which in practice means 25 to 30 percent between the ends, because a narrow band like $27,000 to $28,000 is a quote in disguise and will be treated as one. Name what moves the number, in the same sentence, so the top of the range is attached to a reason rather than to your mood. And put a date on it, since an estimate written against this quarter's availability shouldn't still be circulating in November.
Then convert it into something enforceable, either by quoting the work once the scope is written or, when the scope genuinely can't be written yet, by selling a paid discovery phase whose only output is the document that lets you quote. Charging for discovery also solves the version of this problem that shows up when agencies reach for hourly billing instead, where the total stays open by design and the approval conversation gets deferred rather than removed.
Why the low end of your range becomes the client's budget
An estimate is a range you own for about a week and the client owns for the rest of the project, and the number they keep is the bottom of it. That single sentence is the reason the $24,000 in this piece did more damage than the $33,300 ever did, and it's why the sequence that works is a wide range with an expiry, then a scope, then a fixed number with a unit price inside it, in that order and without skipping the middle step.
The unit price is the piece agencies most often leave out, and it's the cheapest insurance in the document. Once $950 a screen exists on paper, week five stops being a negotiation about whether the request is fair and becomes a short calculation both sides can do. Getting there means the scope has to be specific enough that a new request is visibly outside it, which is the same discipline behind deciding what's in scope and what isn't before the work starts rather than after the invoice goes out.
Frequently asked questions
What's the difference between an estimate and a quote?
An estimate is an approximate price, usually a range, given before the work has been described closely enough to price properly. A quote is one fixed number attached to a scope both sides have already read. The estimate vs quote distinction is really a distinction about how much you know: you estimate when the shape of the job is still moving, and you quote once it has stopped moving.
Is an estimate legally binding?
Generally no, because an estimate is treated as an informed opinion about price rather than an offer the client can accept into a contract, which is why the final invoice can land above it. A quote is the opposite, because a client who accepts a written quote has accepted a price, and you're expected to deliver the described work for that number. Wording and context change the answer, so have your own template reviewed by a lawyer in your state.
How much can a final bill exceed a written estimate?
In several regulated consumer trades the answer is exactly 10 percent. California Business and Professions Code section 9884.2 caps an auto repair bill at 10 percent over the written estimate unless the customer authorizes the excess in writing first, and Virginia, Maryland, Michigan and Minnesota all run versions of the same rule. None of those statutes covers a business to business design or development engagement, but the 10 percent figure is the instinct a client brings to the conversation anyway.
Should an agency send an estimate or a quote?
Send an estimate only to check whether the budget is in the same universe as the work, and quote once you can describe what you're delivering. The mistake that costs money is letting an estimate be the last document before the work starts, because then the range is doing a job it was never built for and every addition gets argued against a number nobody committed to.
How wide should an estimate range be?
Wide enough that the top and the bottom describe genuinely different versions of the project, which in agency work usually means at least 25 to 30 percent between the two ends. A range of $24,000 to $32,000 tells the client something real. A range of $27,000 to $28,000 is a quote wearing a disguise, and the client will hear the lower number and hold you to it.
Does a quote need a change order when the work grows?
Yes, and that requirement is the reason quoting protects you. Because the quote fixed a number against a described scope, anything outside that scope has to be priced and approved on its own before it gets built, which produces a dated record of what was added and what it cost. An estimate has no fixed baseline to change, so the same addition arrives silently and gets discussed for the first time when the invoice does.