Most marketing reports for home service businesses are long, colorful, and carefully arranged so the reader cannot tell whether the marketing is working.

The structure is always the same. Impressions up. Reach up. Engagement up. Click-through rate up. Then a chart of website sessions. Nowhere in fourteen pages does anyone state how many jobs it produced or what each one cost.

Those top-line numbers are not fake. They are just upstream of everything you care about, which makes them useless as a decision input. Here are the five that are not.

1. Cost per booked job

Not cost per click. Not cost per lead. Cost per job actually on the calendar.

Total marketing spend in a period, divided by jobs booked from marketing in that period. That is the number that tells you whether to spend more or less.

The reason agencies report cost per lead instead is that cost per lead is flattering and mostly within their control, while cost per booked job depends on lead quality and on what happens after handoff. It is also the only one connected to your bank account.

Track it per channel, not just in aggregate. A blended figure hides the channel that is quietly consuming a third of the budget and producing nothing.

2. Lead-to-booking rate

Of the leads marketing produced, what percentage became scheduled work?

This is the diagnostic that tells you which problem you have. If leads are plentiful and this number is low, you do not have a marketing problem. You have a follow-up problem, a pricing problem, or a lead quality problem, and buying more leads will make it worse rather than better.

Segment it by source. Google Search leads and Facebook lead-form leads convert at very different rates, and averaging them produces a number that describes neither.

The most common finding when a business measures this for the first time is that a channel everyone assumed was working generates volume that never books.

3. Speed to first contact

How many minutes between a lead arriving and a human or automated system reaching them.

Measure the actual distribution, not the average. The average is dragged down by the fast ones and hides the tail. What you want to know is what percentage are contacted within five minutes, and what happens to leads that arrive after 5pm or on a weekend.

The classic academic work here is the Lead Response Management study by Oldroyd and colleagues, which found dramatically higher qualification odds for contact inside five minutes versus thirty. The direction of that finding has held up across many replications even where exact multiples vary.

Most businesses have never measured this and are surprised by the answer. The gap is almost never during business hours, it is the 6pm Friday lead sitting until Monday morning.

4. Revenue per booked job, by channel

Not all jobs are the same size, and averaging across channels hides the thing that actually determines whether a channel is profitable.

A channel producing jobs at a low cost per booking can still be your worst channel if those jobs are small. A channel with a high cost per booking can be your best if it produces replacements rather than repairs.

You need this to make any sensible budget decision, and it requires your CRM to carry the source through to the invoice. That is usually the missing plumbing rather than a hard analytical problem, most businesses capture source at lead creation and lose it by the time revenue is recorded.

5. Repeat and referral rate

What share of this period's revenue came from customers you already had, or from someone they sent?

This is the metric that separates a business from a lead-buying operation. If every dollar of revenue requires a new dollar of acquisition spend, you are renting demand rather than building an asset. Marketing that raises this number (reactivation, review generation, service reminders, genuinely good work followed up properly) compounds in a way that paid acquisition does not.

It is also usually the cheapest revenue available and the most neglected. Most home service businesses have a database of past customers that nobody has contacted in two years.

The reporting standard to hold an agency to

Ask for one page, monthly, containing:

- Spend, by channel

- Leads, by channel

- Booked jobs, by channel

- Cost per booked job, by channel

- Revenue from those jobs, by channel

Five columns. If an agency cannot produce it, there are two possible reasons, and it is worth knowing which. Either the tracking was never built (fixable, and they should tell you the timeline) or it was built and the numbers are bad. Both are common. Only one is a reason to leave.

Be reasonable about attribution. Someone who sees a Facebook ad, searches your name a week later, and calls is not purely an organic lead, and any attribution model will misassign some of these. The goal is not perfection. It is a consistent method applied every month so the trend is real, plus enough honesty about the model's limits that nobody mistakes it for truth.

What to do with the numbers

The point of measuring is to change something.

If cost per booked job is rising in a channel while others hold, cut it and move the budget. If lead-to-booking is low and speed to first contact is slow, fix the follow-up before buying more leads. You already have the demand. If repeat rate is near zero, the cheapest revenue available is a reactivation campaign to people who already paid you once.

None of that is possible from a report about impressions.