[drop_cap]E[/drop_cap]very lever in this playbook has been pulled once by the time you reach this stage. The phone rings faster. Estimates get booked. Move day runs clean. Reviews come in. What none of that guarantees is that you know which pull is doing the work and which one is dead weight you’ve been hauling out of habit.
This is Stage 10, the last stop in the Refine band, and it is not a new lever. It is the discipline that tells you which of the nine you already have is worth tightening next, and which one to leave alone. A moving company that never looks at its own numbers is running a good truck with the fuel gauge covered by tape. It’ll go until it doesn’t, and you won’t see the empty coming.
The Friday Numbers Ritual
Pick a fixed hour every Friday — before the weekend’s booked moves start, while this week’s leads are still fresh. Sit down with whoever runs your funnel (you, your office manager, your sales lead) and pull the same handful of numbers, in the same order, every week. Not a deep audit. A gauge check.
The ritual has three parts:
- Read the whole funnel at once. Every stage, side by side — awareness, lead, estimate, booked, move day, advocate. Not just the number you’re worried about. A weak spot upstream disguises itself as a weak spot downstream, and you only catch that by reading the whole run, not the one gauge that’s flashing.
- Name the weakest lever. Of everything you looked at, which single number moved the least, or moved the wrong way, this week?
- Decide what gets tested next week. One lever, one change, one week. Not five ideas at once — you’ll learn nothing from five things changing together.
Do this fifty-two times a year and you have a business that corrects itself weekly instead of drifting for a quarter before anyone notices the phone stopped ringing.
The Nine-KPI Dashboard
This is your fill-in dashboard. Set your own targets — what follows are working defaults, not laws of the trade. Track these nine every Friday.
1. Lead response time. Minutes between a lead landing and your first reply. This is the single biggest drop-off lever in the whole business — a mover who answers in five minutes beats one who answers in five hours, because the customer books whoever calls back first. Working default: under 15 minutes.
2. Lead → estimate rate. The share of leads who book an in-home or video estimate. This is where “just gathering info” turns into “on the calendar.” Working default: around 70% — set yours, and treat anything meaningfully below it as a sign your intake conversation isn’t earning the next step.
3. Estimate → book rate. Of the estimates you run, how many turn into a signed, dated move. Working default: around 70%. When this one sags, the fix is rarely a lower price — it’s usually a trust gap. Lean on your white-glove value story before you touch the number on the quote.
4. End-to-end close rate. Lead all the way through to booked move — the composite of the two gates above. This one line tells you whether your whole front-of-house is healthy. Set a working default and watch the trend more than the absolute figure.
5. CSAT score. Post-move satisfaction, scored simply (1–5). This guards the asset your whole local reputation sits on. Working default: 4.0 and above. Below that, hold the review ask — fix the experience first.
6. Happy close → review rate. Of the customers who scored well on CSAT, what share leave you a public review. Working default: around 40%. This is the compounding lever — every review lifts every city page you’ve built, which is why it belongs on the same dashboard as your revenue numbers, not off in a separate “marketing” folder.
7. Referral rate. Of reviewed customers, how many name a friend, neighbor, or relative as a referral. This is your lowest-cost lead source by a wide margin. Working default: high single digits — small numbers here still matter because the acquisition cost is close to zero.
8. Customer acquisition cost. Blended cost per booked move, across every source. Referral and your Google Business listing run cheap; paid search runs several times more expensive. Set a working ceiling and pause any channel that sits above it for several weeks running.
9. Pipeline velocity. New leads per week landing in your intake. This one is seasonal by nature — a peak-season week and a slow-season week are not the same test window, so track it against the season you’re in, not a flat annual average.
Tracing the Cascade
A weak number downstream is a diagnosis waiting to happen, and the diagnosis is not always where the number lives.
Say your estimate → book rate drops this Friday. The obvious read: something is wrong at the estimate. Maybe the quote felt high, maybe the in-home walkthrough felt rushed. Worth checking. But before you touch your pricing or your estimator’s script, look one stage upstream: what did the lead → estimate rate do the same week?
If lead → estimate also dropped, or if your pipeline velocity spiked with a batch of lower-intent leads from a new source, the real story isn’t “our estimate is weak” — it’s “we’re feeding the estimate stage worse raw material than usual.” An estimator can’t close a lead who was never a serious mover, no matter how good the walkthrough is. Fix the intake filter, and the estimate → book number recovers on its own — no one touches the pricing conversation.
The rule: before you fix a stage, check the stage that feeds it. The honest answer to “why is this number weak” is either “this stage has a real problem” or “the stage before it is starving it of good material” — and those two diagnoses call for completely different fixes.
Designing the A/B Test
Once you’ve named the weakest lever, don’t guess at five fixes — design one test.
- One variable. Change exactly one thing: the response-time SLA, the estimate script’s opening line, the review-ask timing, the referral incentive wording. Nothing else moves during the test window.
- One metric. Pick the single number the change is meant to move, and watch that one.
- A defined window. Two weeks is a reasonable default for most of these — long enough to smooth out a slow Tuesday or a lucky Friday, short enough that you get an answer before the next season shifts the baseline.
- A pass/fail threshold set in advance. Decide before the test starts what counts as a win — for example, “the estimate → book rate needs to beat the trailing four-week average by at least three points to call this a keep.” Setting the bar after you’ve seen the number is how businesses talk themselves into keeping changes that didn’t work.
The One-Page Friday Report
Keep it to one page, every week, in the same format:
- This week’s lead → estimate rate, against your target.
- This week’s estimate → book rate, against your target.
- This week’s review rate, against your target.
- The weakest lever this week, and the one-sentence reason (its own stage, or the stage upstream starving it).
- The single test running next week: what changed, what metric it’s measured against, and the date it ends.
That’s the whole report. It should take five minutes to read and less than thirty to build.
AI Earns Its Place Here
This is exactly the kind of week-in, week-out pattern-matching a model is good at and a busy owner has no time for. Feed it your business profile — service area, seasonal pattern, current targets — alongside this week’s nine numbers, and ask it for two things: a draft of the one-page Friday report, and a ranked list of which lever looks weakest given the cascade logic above (checking upstream before blaming the stage that’s merely being starved).
Treat the output as a first draft, not a verdict. The model can spot a number that moved; it can’t tell you whether your estimator had a bad week or a competitor undercut you on price. Read its diagnosis, check it against what actually happened on the ground this week, and then pick the one test that goes live Monday. The report is generated. The judgment about what to test next is still yours.