Fractional CFO for Software Development Agencies: Bench, Blended Rates, and Fixed-Bid Risk

A software agency billing a $150 blended rate against a $95 loaded cost rate isn't making $55 an hour. It makes $55 on every billed hour and loses $95 on every hour a developer sits on the bench or works past a fixed-bid budget. That asymmetry is why dev shops with healthy rate cards can still post thin margins.

A fractional CFO for software development companies manages the four numbers that decide the outcome: blended bill rate, loaded cost rate, utilization, and how fixed-price projects finish against their estimates. Get those four right and growth funds itself. Get one wrong and growth just makes the hole bigger.

The four numbers that run a dev shop

Delivery revenue = delivery headcount × available hours × utilization × realized bill rate
Delivery cost = delivery headcount × available hours × loaded cost rate
Delivery gross margin = 1 − (loaded cost rate ÷ (utilization × realized bill rate))

Available hours means paid hours net of holidays and PTO. Loaded cost rate includes salary, payroll taxes, benefits, contractor fees, and the software and equipment each developer needs. Realized bill rate is what you actually collect per billed hour after write-downs, discounts, and fixed-bid overruns, not the rate card.

A hypothetical 40-person delivery team with a $95 loaded cost rate:

Read across a row and you see the price lever. Read down a column and you see the utilization lever. At a $150 realized rate, going from 70% to 80% utilization more than doubles delivery margin. Going from $150 to $160 at 70% adds less. In this table, the faster money is in utilization and realization, not the rate card.

Every cell sits far below Bennett's 60% gross margin standard, and that gap is the useful part. With a $95 loaded cost, a pure time-and-materials shop would need a realized rate near $297 at 80% utilization to reach 60%. That's why the 60-15-15 conversation in software services turns to rate structure, such as value-based or productized pricing, reusable components, and senior leverage, rather than trimming overhead.

Bench: the cost nobody invoices

Bench is time a developer is paid but not billable: between projects, waiting for a client decision, or ramping onto a new stack. Some bench is the price of being able to start new work quickly. Too much of it is a margin leak that looks like "capacity."

Measure bench in hours by skill, not headcount. A firm can report 78% average utilization while its two senior React developers sit idle and its backend team works overtime. Those are different problems. The consulting bench cost model shows how to put a weekly dollar cost on unassigned capacity and set a decision date before it hits cash.

The Bennett rule for agencies: name the bench budget. If you intend to carry 8% of delivery hours unbilled as a hiring and readiness buffer, write that down. Anything above it gets a decision within two weeks, such as reassigning, discounting a short engagement to fill the gap, or slowing hiring.

Fixed-bid risk: estimate at completion every two weeks

Fixed-price projects are where agency margin can disappear without anyone noticing, because it goes slowly. Nobody misses a budget by 40% in one sprint. They miss it by 6% a sprint for seven sprints.

The control is an estimate at completion (EAC) refreshed every two weeks by the delivery lead and reviewed by finance:

EAC hours = hours spent to date + remaining hours estimated by the team
Projected margin = (fixed fee − EAC hours × loaded cost rate) ÷ fixed fee

A hypothetical $240,000 fixed-bid project, estimated at 1,600 hours:

At sprint 4, the overrun is 100 hours, small enough to fix with a scope conversation. By sprint 12 it's 460 hours and the client has already seen the deliverables they think the fee covers. The value of EAC is that it moves the change-order conversation forward while it's still a minor adjustment.

Revenue recognition should follow the same estimate. If you recognize revenue by percentage of completion, an overrun means you recognized too much revenue early. The guide to revenue recognition for project-based firms covers the mechanics, and the fixed-price AI consulting contract review covers pricing terms that protect margin.

If your rate card looks right but margin doesn't, the gap is usually utilization, realization, or fixed-bid overruns, and often all three. A Leak Check on delivery margin can usually tell which one in 20 minutes.

Domestic and offshore developers: the tax line in the budget

Many dev agencies mix US-based developers with offshore or nearshore teams. Since 2025, the tax treatment of research and experimental spending has depended on where the research happens.

The One Big Beautiful Bill added Section 174A, which reinstates a current-year deduction for domestic research or experimental expenditures paid or incurred in tax years beginning after 2024. The same IRS instructions say expenditures attributable to foreign research must still be capitalized and amortized over 15 years. Rev. Proc. 2025-28 sets out the accounting method changes and elections, including transition options for small businesses.

Whether an agency's client work counts as its own research spending is a separate question. It can depend on contract terms, such as who bears the financial risk if the work fails and who keeps rights to the result. Get that answered by a tax adviser before you build it into the plan.

Where it does apply, the cost of an offshore developer isn't just the lower rate. It's the rate plus the cash cost of deducting that spend over 15 years instead of now. The Section 174A domestic vs offshore budget guide works through that comparison in cash terms.

Staffing mix: the senior squeeze

The other margin lever is who does the work. Replacing senior hours with junior hours lowers cost per hour, but only improves project margin if the juniors' rework and the seniors' supervision time are counted. A fixed bid staffed with too many juniors can show EAC rising in exactly the way the table above shows. The project staffing-mix model gives a work-package method for testing a mix before the project starts.

Dedicated-team contracts, where a client rents a named squad by the month, change the risk but not the math. Utilization is close to 100% while the contract runs, so margin depends almost entirely on the monthly rate against loaded cost. The exposure moves to the end date. When a dedicated team rolls off, five or six people can land on the bench in the same week. Track every dedicated-team contract's notice period and renewal date in the cash forecast, and treat a renewal that hasn't been confirmed 45 days out as a bench event you need to plan for.

What a fractional CFO does in the first 90 days

Days 1–30: is the margin real?

Rebuild the chart of accounts so delivery salaries, contractors, and tooling sit in cost of delivery, not G&A. Reconcile time-tracking hours to payroll hours. Produce the first clean close and delivery gross margin by client.

Days 31–60: where is it leaking?

Calculate utilization and realization by team and skill. Put every active fixed bid on a two-week EAC. Name the bench budget. Start the tax plan, including Section 174A questions for any offshore spend.

Days 61–90: what changes?

Set rate and staffing rules for new proposals from the actual numbers. Build a 13-week cash forecast that reflects milestone billing and client payment terms. Begin monthly reviews of utilization, realization, EAC drift, and 60-15-15 at the company level.

That's how Bennett's fractional CFO services start with an agency, from $5,000 a month with bookkeeping and tax included. If your immediate problem is a late or unreliable close rather than pricing, fractional controller services may be the right first step.

Run one check before your next proposal: realized rate for the last 90 days, by client, against the rate card. If any client sits more than 10% below card, that client is either a pricing problem or a scoping problem, and a Leak Check on your realization and fixed-bid exposure will help you see which.