Which channels produced paid repair orders?

Two channels can produce identical call counts at identical cost while one fills the bays with $60 oil changes and the other with $2,400 engine work. Cost-per-call rates them the same.

Ranking on repair-order revenue separates them immediately, and it is normally the difference between a channel that pays for itself and one that does not.

A first-time customer who returns twice a year for four years is worth many times the first ticket, and no click-based tool credits the original campaign for visit seven.

Because the match is on the customer, every later order credits the channel that produced them — which usually reorders the ranking entirely.

Every multi-point inspection produces a list of recommendations the customer declines that day. Brakes at four millimeters, a weeping water pump, tires that will not pass next year. Some of that work comes back through the door in three months and some of it never does, and when it returns it is booked as a fresh repair order with no memory of where the customer originally came from. It is the largest identified pipeline in the shop and almost no advertising report has ever counted a dollar of it.

Matching on the customer solves the memory problem without anybody building a report to do it. A repair order raised in November for work declined in August credits the channel that produced that customer, because it is the same phone number in the same file. What that changes is the apparent payback on the channels that bring genuinely new customers rather than repeat ones, which tend to be the expensive channels everyone is nervous about. Judged on the first ticket they look poor; judged across the customer they frequently do not.

It is worth exporting a full year for exactly this reason. A month of repair orders shows you first visits and nothing else; twelve months shows the shape of the relationship, including the customers who came once and were never seen again. The same two files answer how often a second visit happens by source, without any extra work, and in a trade where the second visit is the whole margin story that is the more useful of the two numbers on the page.

Shops schedule work and then lose it. The customer does not turn up, the estimate is declined at the counter, the part is on back order and the job is rebooked into a week that never arrives. The appointment book and the closed repair orders disagree by a meaningful margin in every shop, and it is not the same margin for every source of work. A campaign that fills the diary and a campaign that fills the bays can be two different campaigns, and only one of them pays for the building.

This is why the revenue file has to be closed, invoiced repair orders rather than appointments. The appointment list is still worth having — use it as the lead file and the report shows what a booking from each channel turns into — but the money side must be work that was done and paid for. Anything else is a forecast, and a forecast ranked as though it were revenue is precisely how a budget ends up funding the wrong half of the phone.

The gap itself is a finding. If one source produces appointments that convert to invoiced work at half the rate of another, that is a targeting or an expectation problem rather than a pricing one, and it shows up as soon as the appointment book is the lead file. A shop advertising a headline price for one specific service may find the gap sits exactly there: the promotion fills the diary with people who booked the promotion, declined everything else, and left.

A repair order's pay type changes both what it is worth to the business and what marketing had to do with it. Customer-pay work is retail revenue produced, at least partly, by whatever brought that person through the door. Extended-warranty and third-party administrator work arrives with a claim attached and a rate you did not set. Internal work is the shop billing itself. Averaged together they produce a figure that flatters the month and describes nothing, and they are averaged together constantly, because the export defaults to everything.

A pay-type column mapped as Product / service shows customer-pay, warranty and internal revenue as separate lines, with how much of each traces to a channel. To rank channels on retail work alone, export only customer-pay work; the ranking always covers everything in the file. The two are answers to genuinely different questions: the first tells you what the advertising produced, the second how busy the shop was. Confusing them is common and expensive, because a quarter that looks strong on total revenue can be one in which retail demand fell and warranty volume happened to cover the hole.

The same column is what makes a commercial account visible. A handful of local businesses running vans through the shop can be a large share of revenue at a rate nothing like retail, and if one of those accounts originally arrived through a campaign then that campaign has earned the credit. If it arrived through a relationship, it belongs in Direct / Unknown with everything else the advertising did not produce. Either answer is useful to have; a blended average of both is not.