The numbers a home services owner should watch weekly

I spent sixteen years running multi-unit restaurants with more than $54M in annual P&L, and the thing that changed how those businesses ran was not a better report. It was a shorter one, looked at on a fixed day.

The same is true here. The failure is not that contractors lack data. It is that the dashboard has forty numbers and nobody opens it.

What are the five and where do they come from?

Calls answered versus missed. From your phone system's call detail, split by hour. This moves fastest and gets watched least. CallRail's benchmark of 1.1 million de-identified conversations put the missed call rate for home services at 14%, against 32% in health care, 28% in legal and 9% in real estate. The spread is the finding. Your own number is the one that matters, and measuring your actual missed call rate before you do anything about it is the first step.

Leads by source. How many inquiries arrived and from where. Phone, form, Local Services Ads, referral. If you cannot split them, that is your project this month.

Booking rate. Leads that became scheduled jobs, as a percentage. This tells you whether a lead problem is actually a sales problem.

Average ticket. Revenue divided by completed jobs. Weekly, it tells you whether your mix is drifting toward cheap work.

Revenue per truck day. Revenue divided by truck days run. This is the operating number and most contractors have never calculated it.

Five numbers, one page, fifteen minutes. If it takes longer, the report is too big and you will stop opening it.

What does the weekly page look like with real numbers?

Why not track rankings and traffic weekly?

Because they do not move weekly, and watching them at that frequency teaches you nothing except how noisy they are.

Local ranking is not a single number anyway. It is different at every point in your service area and it fluctuates day to day for reasons unconnected to anything you did. An owner checking daily will find a drop, panic, and change something that was working.

There is a harder version of this. No platform emits a local pack position. Google Business Profile performance reports publish views, calls, direction requests, website clicks, messages and bookings. No rank. No position. Search Console has no local pack dimension either. Any weekly ranking number you receive came from a third party tool sampling a location, not from Google, and it should be read as a sample.

Website traffic is similarly noisy at contractor volumes. A few hundred visits a month moves twenty percent on weather. Check both monthly, against the same month last year rather than last month, because seasonality dominates everything in this business.

Weekly numbers should be things you can act on before Friday. Everything else is monthly.

Should I compare my profile against an industry average?

No, and the study people quote for this has aged out of usefulness.

The benchmark in circulation traces to BrightLocal's Google My Business Insights study, published July 2019, covering 45,264 businesses across 36 industries. It reported medians of 1,260 views and 59 actions per listing per month, 33 of them website clicks, and a view to action conversion rate of 4.68%. It is a real study with a disclosed method, and it is still the only large sample one that exists.

Three things disqualify it as your benchmark. It has never been refreshed, so the data window is September 2017 to December 2018. The sample came from BrightLocal's own audit tool users, meaning businesses that already cared about local SEO, which skews high. And several metrics it was built on no longer exist: Google removed photo insights and the direct versus discovery breakdown, and Views is now a deduplicated unique visitor count, in Google's words counting a person only once a day. That is a different quantity from 2018 views.

So a 2026 profile cannot reproduce a 2018 conversion rate, and an agency showing you one is comparing two things that are not the same measurement. Benchmark your profile against its own trailing ninety days and against the named competitors in your map pack. There is no current published average worth the risk.

What do you look at monthly instead?

Six things, and one of them is a competitor check.

Ranking position sampled from three points across your service area. Website traffic and form completions. New reviews collected in the last thirty days, against your top three competitors. Cost per booked job by channel, not cost per lead. Total marketing spend as a share of revenue. And your paid search terms report.

Cost per booked job is the one that changes budgets. A lead is worth what it closes at, and two channels with the same cost per lead can differ by double once you divide by jobs. If you run both paid products, the honest comparison between Local Services Ads and standard Google Ads only appears at that level.

The search terms report catches a specific and expensive problem. Competitors bidding on your company name show up there, and what to do when a competitor bids on your brand name has a real answer that is not always to bid back.

What does each number tell you to do?

That is the test for whether a metric belongs on the list at all.

Missed calls rising means a staffing or coverage problem this week. It moves when a dispatcher takes vacation and nobody notices.

Leads by source dropping in one channel means something specific broke. A campaign ran out of budget, a form is failing, a listing changed. Aggregate lead count would have hidden it.

Booking rate falling with steady lead volume means the leads changed quality or the sales process slipped. If homeowners are balking at price, that is a financing conversation rather than a marketing one, and where financing goes on the site changes who is even willing to ask.

Average ticket falling means your mix is shifting toward cheap work, usually a symptom of what you rank for rather than what you sell.

Revenue per truck day falling with a full schedule means drive time, callbacks or job length. That is an operations problem and no marketing fixes it.

Where does this break down?

Three places.

Small numbers are noisy. At 118 calls a week, a swing from 37% to 33% booking is four jobs and could be one bad Tuesday. Read weekly numbers as a trend across four weeks before you act on any single one.

Attribution is imperfect and always will be. Calls arriving from AI Overviews and AI Mode land in your analytics as ordinary Google organic, with no separate channel, so part of your source split is an estimate no matter what you buy.

A number you cannot act on does not belong. If nobody owns the fix, drop it from the page. A review with no decision attached is a habit that dies in a month.

What to do this week

Fix the day and shorten the list. Monday morning, same time, one page. Not a dashboard you log into and configure. A page.

Build it from the six steps above using last week's actual figures, including revenue per truck day, which almost nobody has.

Have somebody else produce it if you can. An owner who has to build the report will skip it during a busy week, which is exactly the week the numbers change.

Write down what you decided, not just what you saw.

And keep the emergency dimension in front of you. If you take urgent work, your missed call number during the busiest hours is a different number than your weekly average, and emergency searches are won by whoever answers rather than by whoever built the better report.

Questions about which numbers to watch?

Email me at eric@seod.com with the systems you use for phone, scheduling and invoicing. I will send you back a one page weekly review built from those specific tools, showing which report each number comes from, so nobody has to build anything. Takes me about fifteen minutes and there is nothing attached to it.

More on running the marketing side of a contracting business sits in the home services library.