You draw before you are paid. Everyone in this trade does. A client describes a kitchen, a pantry, a wall of lockers, and you put hours into a layout and a take-off so you can quote a number you are willing to stand behind. Most of those hours go to people who never sign.
That is the free design problem. It is not a marketing problem and it is not a closing problem. It is a production problem sitting in front of the production.
What is published, and what is not
Before any arithmetic, here is what you can actually look up, and what nobody has measured.
Published: what a lead costs you. Among the design-build firms on Qualified Remodeler's 2026 Top 500, the median cost per issued lead was $416.50, and referrals, repeat business and company websites together supplied about 60 percent of leads (Qualified Remodeler, 17 August 2026). In the kitchen and bath segment of the same list, the median cost per issued lead was $400 and marketing ran at 9.2 percent of remodeling revenue, against 8.7 percent for the Top 500 overall (Qualified Remodeler, 17 August 2026).
Published: what an hour of woodworking labour costs at the national median. US woodworkers had median pay of $21.79 an hour, or $45,310 a year, in May 2025 (Bureau of Labor Statistics, Occupational Outlook Handbook, page last modified 27 August 2026).
Not published: how many hours a set of drawings and a take-off takes in a small custom shop. There is no survey of this. Trade forums carry individual answers and nothing more. Not published either: the close rate for custom cabinetry specifically. Broad contractor benchmarks circulate, but they come from marketing vendors rather than from named research, and the segments they measure are not yours. So both of those numbers are chosen inputs below, and they are labelled as chosen every time they appear.
An example, with inputs chosen by us
This is an illustration, not a finding. Change the inputs to your own and the shape of the answer holds.
Chosen inputs: twelve hours to produce a layout, a take-off and a priced estimate for one project. Thirty enquiries a year that get far enough to deserve that work. A close rate of twenty percent on those thirty.
Twelve hours times thirty projects is 360 hours a year spent on drawings and take-offs. At a twenty percent close rate, six of the thirty sign and twenty-four do not. Twenty-four times twelve is 288 hours. That is the figure worth sitting with. 288 hours is a little over seven working weeks, every year, produced to a standard you would be happy to build from, and given away.
Put the published wage against it only as a floor. 288 hours at the May 2025 median woodworker wage of $21.79 is about $6,275 (Bureau of Labor Statistics). That is a wage floor and not a cost. It excludes your overhead, your software, your burden, and the fact that the person drawing is usually the owner, who is also the only estimator and often the only salesperson. The real number in your shop is yours to work out, and it is higher.
The acquisition cost stacks on top
Those thirty enquiries were not free either. At the published median of $416.50 per issued lead for design-build firms, thirty leads represent about $12,495 of marketing spend, of which roughly $9,996 attaches to the twenty-four that do not sign (Qualified Remodeler, 17 August 2026). That money is spent whatever happens. The drawing hours are the second payment you make on the same lead.
Why it lands hardest on the smallest shops
Because in a small shop the person drawing is the person who cannot be replaced. NAHB's remodeler members have a median of five employees, median annual revenue of $1.7 million, and complete a median of fifteen jobs over $10,000 a year (NAHB Eye on Housing, 9 September 2025). Below that size, there is no drafter and no estimator. There is the owner, in the evening.
So 288 hours is not 288 hours of somebody's time. It is 288 hours of the one person who also has to price, buy, schedule, install and answer the phone.
What a design fee actually is
The trade already has a standard form for this. The NKBA's Standard Form of Agreement for Design and Consultation Services sets out three fee types a shop can charge before it designs anything: a measure fee for visiting the house, a retainer fee set as a flat amount or a percentage of budget, and an hourly fee (NKBA business form BMF15, copyright 2020, so the document is dated and worth reading against your own province or state's rules).
Two clauses in that form matter more than the fee itself.
The first is design ownership. The form offers sample language stating that all drawings produced under the agreement remain the property of the shop and cannot be used to bid the job elsewhere, with an option to sell the plans for competitive bidding at a stated additional price (NKBA BMF15). If you have ever watched your drawings walk to a cheaper shop, that clause is the answer, and it only exists if somebody signs something.
The second is the credit. The form provides for the design fee to be credited, in whole or in part, against the purchase if the client goes ahead within a stated period, six months in the sample wording (NKBA BMF15). That is what turns the fee from a barrier into a deposit. The client is not paying extra. They are paying early.
What the fee does to the funnel, honestly
Here is the part where most articles overclaim. Whether charging for drawings raises or lowers the number of jobs you sign has not been measured in any published study of this trade. Anyone who tells you it lifts close rates by a specific percentage is quoting nobody.
What can be shown is the arithmetic under chosen inputs. Take the same thirty enquiries. Suppose half of them decline to pay a design fee and walk. Fifteen pay. Fifteen times twelve hours is 180 hours, all of them paid for. You have halved the drawing load and been paid for every hour of what is left.
The risk is on the other side of that trade. If the fifteen who pay close at the same rate as the thirty did, you sign three instead of six. If paying filters for people who were actually going to buy, you sign more than three. Nobody has published which way it goes in custom cabinetry. You will find out from your own numbers within about two quarters, which is why it is worth recording lost reasons from the day you start.
What is not in doubt is the scale of the purchase you are asking people to be serious about. Houzz put the median US renovation spend at $20,000 in 2025, with the ninetieth percentile at $150,000 (2026 U.S. Houzz & Home Study, 22 April 2026). And professionally installed work made up 84.1 percent of all owner improvement spending in 2023 (Joint Center for Housing Studies, Improving America's Housing 2025, released 2025 on 2023 data, so those figures are dated). These are not casual purchases and they are not going to DIY.
What this means for a two-person shop
Count the hours before you change anything. For one month, write down how long each set of drawings and each take-off takes and whether that project signed. Three months of that is better data about your shop than any benchmark you can buy.
Then make the smallest possible change. Put a written design agreement in front of people before you draw, with three things in it: what they get, what it costs, and that the fee comes off the job if they proceed within a set period. Keep the fee modest enough that a serious client does not blink. Say out loud that the drawings stay yours until the job is signed.
Expect to be uncomfortable the first three times you ask. Expect one person to walk who would have bought. Weigh that against seven working weeks.
Casework is built around that gate: the estimate, the design agreement and the paid step are one record, so the drawings only start after the client has agreed to the terms in writing.
The free design problem is not solved by working faster. It is solved by drawing less often, for people who have already said yes to something.