What Average Ticket Actually Measures, and What It Hides

Why this matters

Average ticket is the number owners quote to each other at supply-house counters, and it is the one most likely to be compared between two shops that are not measuring the same thing. It is not the average value of a job. It is the average value of an INVOICE, and that difference is large enough that two shops doing identical work at identical prices can report figures three times apart. Before you act on a move in it - raise prices, push add-ons, question a technician's selling - you need to know which of its inputs moved, because the largest driver is not the work.

Read the figure back with its parts named

• Numerator: the sum of invoice totals, tax included.
• Denominator: the count of invoices. Not jobs, not visits, not customers.
• Time anchor: the date the invoice was issued, not the date it was keyed in and not the date the work was performed.
• Silent exclusion: prepaid deposit invoices, so that a deposit and its final invoice count as one sale rather than two.

Each of those four produces its own predictable wrong conclusion, and the denominator produces the biggest one.

Billing cadence moves it further than pricing ever will

Two shops run planned-maintenance routes. Over one quarter each serves 20 customers with 3 visits apiece, so each performs 60 visits and bills the same total. Index one visit's price at 1.0 unit, so each shop bills 60 units in the quarter.

• Shop A invoices at the truck, one invoice per visit. 60 invoices, 60 units, average ticket 1.0 unit.
• Shop B invoices once a quarter per customer. 20 invoices, 60 units, average ticket 3.0 units.

Shop B's average ticket is exactly three times Shop A's, and the multiple is just the visits-per-invoice ratio of 3 to 1. Same work, same prices, same customers, same technicians, same revenue. Compute revenue per completed visit instead and both shops read 1.0 unit, which is the correct answer and the one that survives the comparison.

Now run it inside one shop. Shop A moves to quarterly billing to cut the office's invoicing load. Its average ticket goes from 1.0 to 3.0 units in a single quarter, a 200 percent rise, and its revenue does not change by one unit. An owner reading the average alone concludes the add-on training landed. The tell that it did not is sitting right beside the figure: the invoice count fell from 60 to 20, down by two thirds.

That is the rule to keep: never read a move in average ticket without the invoice count beside it. The table in the next section reads that pair off for the six things that move the figure. Rule the sales tax rate out first, since it takes one lookup and moves the figure without touching anything you control.

What else moves it with no extra work sold

The last two rows are the ones the owner is actually asking about. They are not the only rows where the invoice count holds still, because a tax-rate change does that too, so a still count rules out cadence and bundling rather than everything.

Tax. Because the figure is tax-inclusive, a rate change moves it. Going from a 6 percent rate to an 8 percent rate multiplies every invoice total by 1.08 over 1.06, which is about 1.9 percent higher, with not one price in the book changed. Small, but it is the difference between a quarter that looks flat and one that looks up.

Deposits. A deposit is not a second sale. The final invoice bills the full scope and credits the deposit as a payment, so the two documents are one sale between them. On books prepared under US GAAP that treatment is a requirement rather than a preference: the revenue-recognition standard, ASC 606, recognises revenue as the work is performed, so a deposit sits as a liability until then, and it binds only an entity that actually prepares GAAP financial statements, usually because a lender, a bonding company or an owner asked for them. A shop on cash books is bound by none of it, and it changes nothing here either way, because average ticket counts invoices whatever the books do. If your figure excludes deposit invoices, that works. If it does not, and every job takes a deposit of 30 percent of scope, then each sale produces two invoices totalling 130 percent of the sale, so the mean lands at 65 percent of the sale value - 35 percent below the same shop's correctly computed figure, on identical work. A shop that introduces a deposit policy mid-year, without the exclusion, watches its average ticket collapse and goes looking for a pricing problem that does not exist.

The issue-date anchor. The office falls two months behind and clears the backlog in one week. Every one of those invoices carries an issue date in that week, so two months of work lands inside one period's figure and the two periods behind it are understated. The average itself may barely move; the invoice count for that period roughly triples. Any period comparison you make across a billing catch-up is comparing a period that was current against one that was not.

Mix moves it too, and mix is not paperwork

Cadence and bundling are billing artifacts. There is a second driver that is entirely real and still tells you nothing about pricing: the proportions of the populations being blended.

Take a shop whose service invoices average 1.0 unit and whose project invoices average 9.0 units. In a period where projects are 5 percent of the invoice count, the blended average ticket is 0.95 times 1.0 plus 0.05 times 9.0, or 1.40 units. Let projects grow to 12 percent of the invoice count and the blend becomes 0.88 times 1.0 plus 0.12 times 9.0, or 1.96 units. That is a 40 percent rise in average ticket with neither population's own average moving by anything at all.

This one is worth separating from the cadence cases because the invoice count does NOT give it away. The count can hold perfectly still while the mix underneath it shifts. The tell is different: split the period's invoices into their two or three natural populations, compute each population's own average, and check whether any of them moved. If none did, you are looking at mix, and the honest headline is that you sold a different blend of work, not that you sold better.

Mix is also the reason a rising average ticket can sit on top of a falling margin. Project work is frequently the thinner work. See the margin cards below rather than assuming the two move together.

What the number does answer, and how to ask it cleanly

The owner's real question is almost always "are tickets getting bigger." Average ticket answers that only when the unit of billing is held constant. Three ways to hold it:

• Revenue per completed job, which anchors the denominator to the work rather than to the paperwork. This is a different number from average ticket and is worth carrying beside it.
• The median invoice next to the mean. One large project invoice in a month of service calls moves the mean substantially and the median barely at all. If the mean rises and the median does not, one invoice did it.
• Segment before comparing. Service invoices, maintenance-agreement invoices and project invoices are three different populations, and a shift in their proportions moves the blended figure with none of the three changing.

One more boundary worth stating once: average ticket is billed value, not collected value. An unpaid invoice counts in full the day it is issued.

The decisions it should and should not change

Average ticket is a billing-side number built from billing-side inputs, so the decisions it legitimately informs are all on that side:

• Whether the minimum charge is doing its job. Look at the bottom of the invoice distribution, not the mean. If a meaningful share of invoices land at or just above the minimum, the minimum is setting the price for that share of your work, which is a decision you should be making on purpose.
• Whether presentation of options is landing. This is legitimate only when the invoice count and the population mix both held still across the comparison, which is the condition the checks below establish.
• Whether to split pricing by population. If your service and project averages are far apart and their proportions move seasonally, a single blended target for the whole shop will be wrong in both seasons.

It should not be an input to any cost or margin decision, because it carries tax, it is denominated in invoices rather than jobs, and it says nothing whatever about what the work cost to deliver. An owner who sets a margin expectation off a target average ticket has connected two numbers with different denominators and a tax difference between them.

Check your own figure

Four checks, each of which takes a few minutes against your own records for a single past period:

1. Count invoices and count completed jobs for the same period and divide. If jobs per invoice is materially above 1.0, your average ticket is partly a billing-cadence figure, and how far above 1.0 tells you how much.
2. Open the largest invoice in the period. Ask whether it represents one sale or a month of sales. If it is a rollup, the mean is being pulled by the cadence.
3. Confirm deposits are out. Scan the period for any invoice whose total is a round percentage of a later invoice to the same customer. If those are in your count, the figure is understated.
4. Re-run the last four periods on issue date and plot the invoice count beside the average. Any period where the count jumps without the customer count jumping is a billing catch-up, and it is not comparable to its neighbours.

The failure mode this prevents is specific and common: an owner sees the average ticket fall for two quarters, runs a sales push, pays for training, and finds nothing improves - because the fall came from a deposit policy introduced in the first of those quarters, and the sales were never the problem. The four checks above separate that case from a real one in under half an hour, and you only have to do it once to know which of your figures needs the caveat attached permanently.

References

• See related: Gross Margin Percent Is an Average of Averages - the weighted-versus-unweighted question that governs every average in this set
• See related: Trip Yield: The Number That Prices a Truck Roll - the same revenue read against completed visits rather than invoices
• See related: Deposits and Prepaids: Money You Haven't Earned Yet
• Generally Accepted Accounting Principles (GAAP), revenue-recognition standard ASC 606 - why an entity that prepares GAAP financial statements treats a customer deposit as money held rather than as a sale, and why a shop on cash books is not applying it