Fixed-Price vs Time and Materials for AV Integration

Here's the answer most guys don't want to hear: it's not one or the other. The AV shops that consistently hit 30%-plus margins bid the physical installation fixed-price and keep programming, integration tweaks, and service work on time and materials. Do it the other way around — T&M on the install, fixed on the programming — and you've built a machine for losing money.

Let me walk through why, with real numbers.

Start with what each model actually does to your risk. A fixed-price contract means you named a number, and every hour past your estimate comes out of your pocket. T&M means the client pays for every hour, so they carry the overrun risk — but they also get to question every line on every invoice, and your upside is capped at your hourly rate. There's no efficiency bonus on T&M. If your crew finishes a rack build in six hours instead of ten because you prefabbed it in the shop, you just gave the client four hours back.

That last point is the whole argument for fixed-price on installation work. Industry benchmarks put integrator labor margins around 55% and equipment margins around 43% — and that labor margin only exists because good shops beat their own estimates. Your loaded cost on a tech is probably $35 to $55 an hour once you count wages, burden, truck, and insurance. Your sell rate on install labor should be running $85 to $125 in most markets, more in a major metro. On T&M, that spread is your entire margin. On fixed-price, the spread plus every hour you shave off through prefab, better staging, and repeat-design efficiency is yours to keep.

There's also a market reality: on commercial projects, you often don't get a choice. General contractors and most corporate clients want a number they can plug into a budget. Show up to a conference room buildout bid with "we'll bill hourly and see where it lands" and you're not getting the job. Fixed-price is the price of admission for construction-adjacent AV work, so the real question isn't whether to use it — it's what you refuse to include in it.

What belongs in the fixed price — and what doesn't

Put inside the fixed number everything you can count and everything you've done twenty times: displays mounted, speakers flown, cable pulled and terminated, racks built, standard commissioning. If you can walk the drawings and say "that's a 6-hour room, we've built forty of them," fix the price and defend it. This is exactly what labor standards like NSCA's installation data exist for — published labor units per task, with multipliers for normal, difficult, and challenging conditions. If you're not using a labor unit reference or at least your own historical job-cost data to build fixed bids, you're guessing, and fixed-price punishes guessing brutally.

Keep outside the fixed number anything where the hours depend on someone else's decisions. Programming is the big one. Crestron, Q-SYS, Biamp, Extron control work sells for $100 to $175 an hour for a reason: the hours aren't driven by the hardware, they're driven by how many times the client changes their mind about what the touch panel should do. Same goes for integration with the client's network and IT stack — if their IT department takes three days to open firewall ports, that's not a cost you should have baked into a fixed bid six weeks earlier. And service after substantial completion should always be T&M or a service agreement, never a vague "warranty support" line inside the project price.

The data on this is ugly and worth knowing. Studies of T&M engagements show final invoices routinely landing 30% to 80% above the original estimate. Clients know this, which is why they push for fixed pricing — and why fixed-price bids typically come in 10% to 25% higher than the equivalent T&M estimate at signing. That premium is real and you should charge it. You're selling insurance against overrun, and insurance isn't free. When a client asks why your fixed number is higher than a competitor's "estimate," say exactly that: my number is the number. Theirs is the opening bid.

The hybrid structure that actually works

Here's how I'd paper a typical $80K conference room and huddle space package:

1. Fixed price for equipment, installation, and standard commissioning, tied to a signed scope, drawings, and a finalized equipment list.
2. A defined programming allowance — say 60 hours at your programming rate — with anything beyond it billed T&M at $125 to $150 an hour, authorized in writing before the meter runs.
3. T&M with a not-to-exceed cap for owner-driven extras discovered mid-job: added drops, relocated displays, the fourth revision of the touch panel layout.
4. A separate service agreement or published T&M service rate card that starts the day of substantial completion.

The not-to-exceed cap on item three matters more than it looks. Straight T&M makes clients nervous, and nervous clients micromanage. A cap gives them a ceiling to budget against and gives you room to work without a phone call over every hour. Most clients who resist open-ended T&M will sign a capped version without blinking.

One more piece: your contract language has to make the boundary between fixed and T&M enforceable. "Programming includes two rounds of user interface revisions" is enforceable. "Programming as required" is a blank check written against your own account. Spell out revision counts, response-time assumptions for client decisions, and what happens when their IT team goes dark for a week.

The mistake that eats shops alive

The most expensive mistake I see — and I've watched it sink real companies — is signing a fixed price before the scope actually exists. A shop I know bid the control programming for a six-room corporate job at a flat number built on 40 hours. Reasonable, based on the room count. But the client hadn't decided how the rooms would be used. First the divisible room combining logic changed. Then facilities wanted scheduling panels tied into their calendar system, which wasn't in anyone's drawing. Then the CEO saw the touch panel and wanted the whole UI redone. Ninety-plus hours later, the programmer was working for roughly half of loaded cost, and every one of those change requests had felt too small individually to stop the job and fight over.

The fix isn't heroic estimating. It's refusing to fix a price on undefined work. If the client can't tell you exactly how a room will function — sources, use cases, who touches the panel, what integrates with what — then that portion of the job is T&M or it's a fixed price contingent on a signed functional spec, with changes after signature billed hourly. Put it in the proposal that way. Clients respect it more than you'd expect, because the alternative you're offering is a padded number that punishes them for your uncertainty.

A few pricing details that'll save you grief regardless of model. Bill T&M in half-day minimums for truck rolls; a two-hour minimum on a site 45 minutes away is a losing trade. Quote after-hours and weekend T&M at 1.5x — a lot of commercial AV work happens nights and weekends because the client won't give up the boardroom during business hours, and if that's not priced, it's a gift. And on fixed bids, carry a contingency line of 8% to 10% on labor for anything in an occupied building or a renovation, because ceiling conditions never match the drawings.

The pattern to internalize: fixed-price is how you get paid for being good at things you've done before. T&M is how you avoid paying for other people's indecision. Every line item in your proposal should sit on one side or the other of that fence deliberately — not because that's how the last proposal was formatted. Get the fence in the right place and both models make you money on the same job.