Job Costing in Jobber: Getting Per-Job Gross Margin When Jobber Has the Jobs and QuickBooks Has the Costs

Jobber tracks revenue per job down to the line item, and it lets you enter costs and expenses against a job by hand. Per-job gross margin is the harder half. The margin math needs every cost to reach the job record, and in most 5 to 25 person shops the costs live in four other systems and never get there.

We build reporting systems on Jobber’s API for a living, including per-job margin reports that pull from more than one system. We are independent builders and not affiliated with Jobber or Intuit. This post maps where the costs actually live, what Jobber covers on its own, and the three ways shops actually get the margin number.

The five-system sprawl

Here is the pattern in a typical 5 to 25 person shop on Jobber. A shop owner wrote it up almost word for word in a public hiring post this month. Five working tools that never talk about costs.

Each system is good at its own job. The sprawl happens because costs are born in different places. The crew lead grabs extra material on install day, and it surfaces on a supplier statement three weeks later. A sub sends one invoice covering three jobs. The punch list return trip burns a crew day and never gets costed anywhere.

So the owner builds a sheet. Every week it gets fed by hand: revenue from Jobber, bills from QuickBooks, statements from the supplier portal. The sheet works. It also only works while the owner keeps feeding it, and nobody else can read it.

The definition inside that sheet is usually deliberate. One version keeps showing up: gross margin equals job revenue, minus equipment and material costs, minus subs hired for that specific job, with labor left out on purpose. More on that choice below. The point here is that the owner already knows the definition. What is missing is a system that computes it without the owner in the loop.

What Jobber gives you out of the box

Jobber’s revenue side is genuinely complete. Quotes, invoices, and payments all hang off the job, so “what did we bill on this job” is always one click away.

On the cost side, Jobber has two built-in tools.

• Line item costs. You can record what a line costs you next to what you charge, so a quote can show its expected margin before you send it.
• Expenses on jobs. You can log a cost and tie it to a specific job, which covers receipts your team enters as they go.

Some of this varies by plan, so check your plan’s current features rather than taking any blog’s word for it, ours included.

Both tools share one dependency: a person has to type the cost in, at the time it happens, on the right job. That is a discipline system. Jobber cannot see a bill in QuickBooks, a charge on a supplier statement, or a sub’s invoice in your inbox. Whatever nobody types stays invisible, and the job’s margin reads higher than it was.

The QuickBooks sync covers the other side

Most shops assume the Jobber and QuickBooks sync handles this. It handles a different problem. The sync exists so your books reconcile: customers, invoices, and payments flow from Jobber into QuickBooks Online. That is the revenue side again.

Your costs enter QuickBooks from the other direction. Sub bills, card charges, supplier payments. They arrive coded to a vendor and an account, and in most setups nothing on them reliably names the Jobber job. Sync configurations vary, so check how yours is set up, but this is the wall shops hit.

The two systems agree on what you earned. Neither one can tell you what job 4127 cost. And when the sync itself is broken, taxes off, invoices missing, that is a different problem with its own checks: we wrote the diagnosis guide for it.

Some shops add a fourth seam: a separate CRM that tracks where each job came from. That system holds the marketing spend behind the job, which matters if you want margin net of what it cost to win the work. Same problem in another spot. The record exists, and nothing keys it to the job.

Three ways to get per-job margin

There are three approaches that exist in the real world. Everything else is a variation of one of these.

Discipline only

Enter every cost into Jobber, on the job, when it happens. This works in a two or three person shop where the owner touches every purchase.

At 5 to 25 people it decays fast. Costs get bought by four different people on six different accounts, some at 7am at the supply counter. The busiest weeks, when margin matters most, are exactly when entry slips. The number you get is honest on slow weeks and fiction on big ones.

It can be shored up. A rule that the crew lead photographs every receipt before leaving the counter. A Friday sweep where the office enters what the week bought. Shops do make this hold, and the ones that do treat cost entry like the punch list: part of the job, checked weekly, never optional.

The spreadsheet bridge

This is the sheet most owners already run. Export jobs and invoices from Jobber to CSV. Export cost detail from QuickBooks. Paste both into a sheet and match costs to jobs by lookup and by memory. We covered the export routes and their limits in how to export data from Jobber.

The sheet gives a true number once a week, for the price of an evening, call it 2 to 4 hours. Its failure modes are human: job names that never match exactly, a moved column, a week off. And one person maintains it, because one person knows its rules. When that person is also running the company, the sheet is the first thing to slip.

The built data bridge

A small system pulls jobs from Jobber and costs from QuickBooks into one database, matches them by job, and serves one report. Every job, revenue, material costs, sub costs, margin dollars, margin percent. Nobody assembles it. It is there every morning.

That is the short version. The next section is the honest version.

What a built bridge looks like in practice

Nightly, not live. Jobber’s one API is GraphQL, gated to the Plus plan, and it meters every query against a cost budget. There is no reporting API and no bulk export, so a margin report gets rebuilt from raw records through paginated queries. Query Jobber live on page load and you hit the throttle: we measured a single dashboard load at 13,456 to 20,762 query points against a 10,000 point budget. The full numbers are in our production measurements of the Jobber API. So the bridge syncs on a schedule into its own database, and the report reads local.

Jobs keyed across systems. The unglamorous core of the build is matching. A QuickBooks bill has to land on the right Jobber job, so the bridge needs a key. That can be a job number in the bill memo, vendor rules, or whatever your bookkeeper can keep consistent. On the cost side the bridge reads bills, sub payments, and card charges, and every bookkeeper codes those differently. The first week of any build is mapping how yours does it today. Expect some costs to miss every rule, call it 1 in 10, and land in a review queue for a human to assign. Any honest builder will tell you that queue exists. The version without it quietly drops those costs, and the margin report reads high.

“Our biggest problem was that our office team was constantly moving information between systems. Danilo’s team built the integration for us and automated a big part of the workflow. It’s one of those projects where you wonder why we didn’t do it sooner.”

Owner, home services business. Client name shared on a call.

Your definition, computed the same way every night. One table, every job. If your definition is revenue minus material and equipment costs minus job-specific sub costs, that is what it computes, without anyone’s Sunday attached. This is work we run in production for Jobber shops. Our Jobber reports and dashboards page covers the margin reports and sync pipelines we build.

Labor in or out

The definition above leaves labor out on purpose, and shops running steady crews keep arriving at that same choice. Their reasoning holds up. The crew costs the same this week whether job 4127 ran long or short. Leaving labor out gives a clean contribution number: did this job cover its hard costs and put money toward payroll and overhead. Nobody argues about splitting a crew day across two jobs, because nobody has to.

Include labor, at its burdened rate, when labor cost changes per job. That means piece rate or per-job pay, jobs that swing from one crew-day to five, or comparing install work against service work. Burdened means wage plus payroll taxes, comp, and benefits, call it 1.3 to 1.4 times the hourly wage.

Either definition works if it stays put. Write yours down in one sentence, then make the sheet, the report, and every Monday meeting use that sentence.

Where margin fits on the scoreboard

Margin per job is one number on a longer scoreboard. We mapped 12 field service KPIs across Jobber, Housecall Pro, and ServiceTitan. Our field service KPI guide shows which of those numbers each one can display out of the box. The pattern there matches the pattern here. The software is strong wherever money touches it, and weak on any number that joins two record types.

Per-job margin joins three: jobs, invoices, and costs that live in another company’s database. That is why it is the number owners want most and see last.

One thing to do this week

Before you buy or build anything, run the test by hand once. Take last month’s five biggest jobs. Pull revenue from Jobber, costs from QuickBooks and your supplier statements, and compute margin using your written definition.

If that takes 20 minutes, your discipline system is holding. Keep it. If it takes a Saturday, or two of the five numbers surprise you, the sheet has already told you what it needs to become.

Building that is what we do. Our custom Jobber reports page covers per-job margin reports that pull Jobber and QuickBooks into one place. Or book a discovery call and bring your sheet. If a cleaner export and a better sheet gets you there, that is the advice you will get on the call.