The same project is $5,276 underwater or $400 ahead, and the difference is a guess

The moment

Fernbank Studio is one person. Brand and web work, Athens, Georgia, fourth year, working out of a rented desk in a shared studio off Prince Avenue. Subcontracts development when a project needs it.

It is a Tuesday in late September. A regional credit union has asked for a price on a brand refresh and a new site: new mark, type system, eleven page templates, content migration. Fernbank has scoped it at 180 hours of their own time plus about 40 hours of subcontracted development.

The number on the draft proposal is $18,500. It has been $18,500 for two days because that is roughly what the last one went for, rounded to something that sounds deliberate.

What they did instead

The usual method, which is three methods stacked on top of each other and none of them written down.

First, anchoring on the last job. The previous site build went out at $17,000, so this one is a bit bigger, so $18,500.

Second, asking a friend. Two other freelancers in the same city, both of whom answered honestly with what *they* charge, which is information about their cost base and not about Fernbank's.

Third, fear. The credit union has an in-house marketing person who mentioned they were "also talking to an agency in Atlanta." So the number went down rather than up.

What that costs is invisible, which is why it keeps happening. On this project it is roughly $5,276 — and the same method produced roughly the same shortfall on the last four projects, none of which felt like losses, because every one of them put money in the bank on the day the invoice cleared.

A job that pays you today and costs you this year does not announce itself.

The questions a CFO asks first

1. What do you need to take home in a year? Not what you would like — what the year actually requires.
2. What does the business spend in a year regardless of which projects come in?
3. How many hours a year do you actually bill? Not work. Bill.
4. What does this specific job cost you in materials and paid outside labour?
5. How many of your own hours does it take?
6. What are you charging for it now?

Question 3 is the whole thing. It sits in the denominator, so getting it wrong does not nudge the answer, it scales it.

The arithmetic, shown

Fernbank's figures. Take-home target $95,000. Fixed costs $34,000 a year — health insurance, the studio desk, software, accountant, liability cover, equipment. This job: 180 of their own hours, $5,200 of direct cost ($3,800 of subcontracted development, $1,400 of type licences, stock and hosting setup). Current price $18,500. They believe the market pays about $20,000.

Scenario one — 1,250 billable hours a year.

This is Fernbank's honest number: about 24 billable hours in a working week, allowing for quoting, invoicing, chasing, selling, admin and holiday.

Scenario two — 1,800 billable hours a year.

This is the number the same owner gives on a confident day, because it is closer to hours worked.

Same owner. Same job. Same costs. One guess apart, and the project goes from $5,276 underwater to $400 ahead.

Note which direction the error runs. Overstating your billable hours makes the required rate come out lower, so the optimistic answer is the one that says your price is fine. That is the dangerous direction, and it is the direction almost everybody errs in.

At 1,800 the tool says so out loud rather than correcting it — that is a high share of a working year to bill, and if any of it goes on quoting or invoicing the real rate is above the one shown. It uses the number you gave. It will not substitute one of its own.

Scenario three — how much scope would have to come out.

At $18,500 and a required rate of $103.20, the project works at 128.9 hours instead of 180. About 51 hours would have to leave the scope. That is the whole content migration and most of the template build.

Scenario four — bringing the development in-house.

The instinctive fix: drop the $3,800 subcontractor and do it yourself. Direct cost falls to $1,400, own hours rise to 220. Floor becomes $24,104, and the gap gets *worse* by $328.

Your own time is not free. It is the most expensive input on the job, because it is the one that is capped.

And the market figure. Fernbank believes the market pays about $20,000 — below the floor of $23,776. The tool does not treat that as a reason to discount. It treats it as a finding about the business: at this cost base and these billable hours, this work does not pay at what they believe the market is. Something has to move — the costs, the scope, the hours, or the kind of work. Pricing under the floor moves none of them and hides all four.

What changed

The proposal did not go out at $23,776. That is not what happened and it is not what the tool is for.

What happened is smaller. Fernbank sent the same $18,500 proposal with the content migration removed and priced separately, which took the scope close to the hours the price supports. And they started logging billable hours properly for the first time, because the entire answer turned out to hang on a number they had been estimating from memory.

Eleven weeks of real logging will settle whether the floor is $23,776 or something else. Until then the floor is a number with a known uncertainty in it, which is a considerable improvement on a number with an unknown one.

What the tool will not do

It will not tell you what the market pays. It asks what you believe the market pays, labels that as a belief, and uses it only to frame the gap. It never estimates it.

It will not correct your billable hours. If you tell it 2,200 it will use 2,200, warn you loudly that that is a full working year with nothing taken out, and show you a required rate that is too low as a result.

It will not tell you to raise your price. The floor is the level below which the job takes money out of the business. Holding your price is a real option, and the tool prices it as a subsidy you can name rather than a mistake.

It is built around one person's time. A firm with several billers can use it by treating everyone else's salary as a fixed cost and pricing the owner's hours, but the hours warning is calibrated for one person and a multi-person total will trip it.

It does not handle retainers, value-based pricing, or work where your hours are not the constraint.

It does not know whether you want this client. That is frequently the right reason to take a job below the floor, and the tool's job is to make sure you know what that costs before you decide, not to decide it.