← Full sheet index
FIELD NOTE · SUPPLEMENTAL SET

The money's in the big jobs: what contract value leaves out

ISSUEDJULY 4, 2026DRAWN BY THE NIGHTLY SWEEP
Architectural floor plans spread loose across a work table
The California Building Code keeps a list of the work you can do to a commercial suite without telling anybody about it. Paint, paper, tile, carpet, cabinets, countertops. Movable partitions, so long as they are not over 5 feet 9 inches in height. Almost everything else in that suite files.

Which makes a filed permit the cheapest qualifying step an integrator has, and it costs nothing, and most of the selling week still goes to jobs that haven't got one yet.

That list is section 105.2 of the 2025 California Building Code, Building exemption items 7 and 13. Section 105.1 covers the other side of the line: anybody intending to construct, enlarge, alter, repair, move, demolish or change the occupancy of a building "shall first make application to the building official and obtain the required permit."

So the code sorts the market at no charge, into work that leaves a public record and work that leaves nothing behind. What it won't do is rank what's left, and the ranking is where the money leaks out of a small shop.

The number on the proposal isn't the number that pays you

Rank by contract value and the ground-up job wins every time. It's the biggest figure on the sheet, it comes with a logo for the capabilities deck, and the pipeline review is built to reward exactly that. Contract value measures the invoice. What a shop actually runs out of is selling hours.

The review ranks by value anyway, partly because value is a number somebody typed into a field while hours are a number nobody typed anywhere, and partly because the largest job in the pipeline is the one everybody at the review has already agreed to want.

The arithmetic that respects the constraint is one line. Average quoted package value, times win rate, divided by hours of pursuit. Expected revenue per hour of selling, which is the only ranking that treats a week as finite.

Watch the value term. It is what you quoted, not what you won, because the win rate is already sitting in the formula and multiplying by it a second time buries the answer under itself.

Two of those three numbers are already in the CRM, quoted value and win rate both. The third one is where it falls apart.

Nobody fills in the hours honestly

Pursuit hours on a ground-up package never arrive as a block. They show up as a site meeting here, a revision there, a value engineering round in month nine that moves the head end because somebody else's budget moved first. (Add up the calendar entries on one of those pursuits, the drives, the two hours after every meeting spent redlining a set that is going to change again anyway, and the total is not a number that ever reaches a forecast.) Spread thin enough, hours read as overhead instead of cost.

On a small alteration they're countable, because there are so few of them. Which is exactly why the comparison feels rigged: one side of it is measured and the other side is an estimate made by the person who wants the big job.

Win rate by job size is not something this trade publishes. Bid-hit benchmarks exist, and they are cut by delivery method and by trade, not by deal size, so anybody quoting you a multiple for small work against large has built it out of numbers they made up.

Run it on your own records instead. The version that skips the multiplication is easier and harder to get wrong: total closed-won revenue over the last two years, divided by total pursuit hours over the same two years, sorted by job size. The wins are already the numerator, and two years is long enough to smooth out one lucky whale.

The math reads like an argument for doing smaller and duller work, which is how it usually gets waved off. Waving something away because it sounds like a pitch is a habit with a real cost, and it's a hard one to catch in yourself.

The code did the qualifying

Quantity ladder on one vertical scale in inches: the permit-exempt band for nonfixed and movable partitions stops at a hard line at 69 inches, a dashed line 3 inches above it marks Santa Clarita's amended 6-foot version of the same exemption, and permit-required work sits in a separate off-scale panel marked no height threshold applies

The only height the code actually fixes. Nonfixed and movable fixtures, cases, racks, counters and partitions are exempt when they are both movable and not over 5 feet 9 inches (1753 mm), item 13, and finish work is exempt at any size, item 7. Fix a partition to the structure and it files at any height, which is why the permit side of the drawing carries no scale: alteration, repair and demolition file under section 105.1 regardless of height. The dashed line is Santa Clarita's amended list, where the same exemption runs to 6 feet. Source: 2025 California Building Code sections 105.1 and 105.2; Santa Clarita Municipal Code 18.02.030.

Here's what 105.2 is worth to a low-voltage shop. A suite can be repainted, recarpeted, refitted with new cabinets and countertops, and divided up with partitions that are both movable and no taller than 69 inches, which is the 5 feet 9 inches the code writes as 1753 mm, and none of it files a thing. Both conditions, not either one. Fix that partition to the structure and it files at whatever height it happens to be.

None of that cosmetic work appears in the feed. It was never a job.

One qualifier, because it moves the line. Section 105.2 is the state floor and cities rewrite it. Santa Clarita picked up the 2025 edition on January 1 and amended the whole exemption list: finish work lands at item 9 there, cases and racks and partitions at item 8, the height runs to 6 feet, and the word nonfixed is gone. Read the state list for the shape of the rule. Read your own city's ordinance before you bet a week on exactly where it falls.

So an alteration permit is a qualified lead in the strict sense. The code decided the work was real construction before anybody spent an hour on it: walls, and the power and mechanical that come with walls, and somebody settling this month where the conduit runs.

That's a different starting position from a ground-up pursuit, where the first several conversations happen against a project that is still an intention, and where the low-voltage budget is a line three other trades are also trying to spend.

Most tenant improvement work follows a signed lease, which is a firmer commitment than a ground-up entry still waiting on financing. Not all of it does. Landlords build spec suites and owner-occupants file alterations on buildings they already sit in, so a permit proves construction is committed, not that a tenant exists.

The records themselves are public. The California Public Records Act (Government Code section 7920.000 and following) makes public records open to inspection during agency office hours, and the right to inspect any record sits at section 7922.525. Los Angeles publishes its building permit datasets on its open data portal. Nothing in this is privileged access. The data is open and the timing is the whole variable.

Demolition can be its own permit

Los Angeles carries demolition as its own permit type inside the same public permit file, so a demo filing arrives already sorted. The city's issued-permit dataset runs from January 2020 through mid-July 2026, and the commercial share of it breaks down like this: 35,277 alteration and repair permits, 1,778 demolition permits, 1,440 new building permits. Alterations outrun new commercial buildings by about 24 to one.

LOS ANGELES COMMERCIAL PERMITS ISSUED, JAN 2020 TO JUL 2026 ALTERATION AND REPAIR 35277 DEMOLITION 1778 NEW BUILDING 1440

Whether the demolition files ahead of the buildout is a choice, not a rule. Section 105.1 requires a permit to demolish and says nothing about when, and interior demo is routinely folded into the alteration scope, but in Los Angeles the applicant can split it out. Under the LADBS Early Start Program an interior non-structural demolition permit "may be issued for any building prior to the issuance of a tenant improvement permit", covering non-load-bearing partitions, ceilings, and the plumbing, electrical and mechanical behind them. Take that option and the demolition record posts while the buildout is still being drawn.

How far ahead it lands is not published anywhere, and the week counts quoted for it are guesses. The record is the part that holds, which is why demolition filings read as a redevelopment signal in the first place.

For the hour count, it matters in one specific way: a second record costs the same to read as the first. PermitFeed sorts the commercial permit record for AV integrators with demolition normalized as its own type, so a saved search on demo filings at commercial addresses runs nightly and takes no selling hours at all to keep open.

The second suite in the same building

Multi-tenant buildings turn over, and the denominator is the reason to care. The first job in a building costs full pursuit hours. The second one shouldn't. The property manager knows the crew, the riser is known, the head end lives in a closet one of your techs has already stood in, and the punch list from the first job is the reference set for the second.

Whether it converts is untested. Nothing published tracks how landlord vendor lists actually get used, so run it as a play rather than a finding. The question after a clean handoff is which list the leasing office hands to incoming tenants, and what it takes to get on it. If it works, it works by cutting the hours. Not by raising the price.

The filings post pre-sorted by a code section written for reasons that have nothing to do with low voltage. What survives 105.2 is walls, power, and a permit number. Somebody still has to pick which of them is worth a phone call, and that pick is the one hour in the week that prices all the others.