01 · Decision first
A report needs a business decision.
Measurement starts by naming what someone must decide: keep or stop a channel, change a response workflow, expand a service area, revise an offer, protect capacity, or investigate a data gap. If the dashboard cannot change an action, it is probably reporting theater.
What allocation, workflow, offer, or operating decision needs better evidence?
What time period and sales cycle are long enough to observe a meaningful result?
Who can change the business process when the evidence points to a constraint?
What would support keeping, changing, expanding, or stopping the work?
02 · Shared definitions
Define the stages before counting them.
- Inquiry
- A submitted form, call, message, booking, or other received request. Not every inquiry is a lead.
- Qualified opportunity
- A real buyer whose service, territory, timing, and project fit the business’s stated rules.
- Estimate
- A scheduled or delivered proposal with a recorded status—not a vague conversation about price.
- Customer
- A won opportunity connected to a real job and, when available, collected revenue.
Invalid, duplicate, spam, wrong-service, out-of-area, unreachable, and economically poor-fit records should remain visible. Removing them without a rule can make a report look cleaner while making the decision worse.
03 · Handoff data
Record the events that explain conversion.
- 01
Source and received time.
Preserve the channel and campaign context available at inquiry creation, together with the timestamp and contact route.
- 02
Ownership and response.
Record acknowledgment, first human attempt, contact, responsible person, and next action.
- 03
Qualification and estimate.
Apply a consistent disposition, capture the reason, and connect valid opportunities to estimate status.
- 04
Sale and economics.
Connect the final sales state to job value, collected revenue, and available direct-cost or gross-margin context.
04 · Attribution limits
Use the best evidence available without pretending it is perfect.
Phone calls, repeat customers, referrals, multiple devices, offline estimates, privacy choices, unmatched CRM records, platform modeling, and long sales cycles can all weaken source attribution. A useful system names those breaks instead of forcing every customer into a false single-source story.
- Separate observed from modeled. A platform estimate is not the same as a matched business record.
- Preserve uncertainty. Unknown and multi-touch records are more honest than invented certainty.
- Audit the joins. Check whether source identifiers survive the form, phone, CRM, estimate, and job records.
- Protect customer data. Collect only what the decision requires and define access, retention, and export responsibilities.
05 · Owner scorecard
Show the few numbers that explain the constraint.
Demand quality
Spend, inquiries, valid opportunities, qualification rate, and cost per qualified opportunity.
Response
Acknowledgment, first human attempt, contact rate, open follow-up, and final-status coverage.
Sales
Estimate rate, win rate, loss reasons, sales-cycle timing, and sold work by source.
Economics
Collected revenue, acquisition cost, repeat or referral context, and gross margin where reliable data is available.
The scorecard should end with the decision, the owner, and the next check—not another pile of charts.
06 · Ownership and scope
The client keeps the records and the definitions.
Lerrow scopes measurement inside client-controlled accounts wherever the relevant platforms allow it. Third-party licenses, storage, retention, APIs, exports, and attribution models remain subject to vendor terms. Definitions, responsibilities, limitations, and handoff requirements belong in the written scope.
Measurement does not guarantee a favorable outcome. It is useful when it makes the commercial truth easier to see and the next decision harder to avoid.