Inquiries are arriving. Some are obvious poor fits. Some turn into real opportunities or booked work. But when the owner asks which marketing activity produced the useful ones, the answer is usually a mix of platform reports, memory, and guesswork.
The missing piece is often not another dashboard. It is continuity: one inquiry enters the business, gets reviewed, moves forward or stops, and eventually has a result that can still be connected to what the business knows about its source.
This article starts after the inquiry has reached the business and someone or some process owns the response. When inquiries are not delivered, acknowledged, assigned, or owned yet, fix the website form handoff first.
For each inquiry, keep one record current enough to follow it from arrival to result. That is usually enough to explain which activity produces qualified interest, where good inquiries stop, and which sources appear to contribute to booked work—without pretending attribution is perfect.
Keep one practical lead record#
The record does not need every customer detail or every interaction. It needs enough continuity to connect the inquiry with a source and confidence level, what the person responded to, whether the work is a fit, where the inquiry stands, and how it ends.
| Field | Decision value |
|---|---|
| Received date | When the inquiry entered the process |
| Likely source | The channel, referral, partner, campaign, or other honest source label currently supported by the evidence |
| Source confidence | Whether that source assignment is known, likely, or unknown |
| Page or offer | What the person appears to have responded to |
| Service requested | What work the person believes they are requesting |
| Fit | Whether the inquiry is a reasonable fit, not a fit, spam, an existing customer, or not yet reviewed |
| Current stage | Where an open inquiry currently stands |
| Outcome | The final result once it is known |
| Outcome reason | Why the inquiry progressed or stopped |
| Last updated | Whether the record is current or abandoned |
Suppose an inquiry arrives through a service page and the person says they found the business on Google. The initial record might show a likely source of Google search with likely confidence, the service page as context, a reasonable-fit classification, and a current stage of received. The later entries should change as the business learns more.
Likely-source values might include Google Ads, organic search, referral, partner, existing customer, direct, or another label the business can defend. Use labels specific enough to support the current decision, but do not treat “direct” or another convenient bucket as certainty when the evidence does not establish the source.
Use the maintained tool the business already relies on: an existing CRM, spreadsheet, booking system, or another practical record. Do not create a second store of sensitive customer information merely for marketing reporting.
When a customer record already exists, the measurement record can use an internal identifier or link instead of duplicating names, contact details, messages, and other personal information.
Keep fit, current stage, and outcome separate#
These fields answer different questions. A lead can be a reasonable fit, waiting on a quote decision, and still have no final outcome.
Fit#
A small adaptable fit field may use:
- not reviewed;
- reasonable fit or qualified;
- not a fit;
- spam;
- existing customer;
- another business-specific classification when it improves the decision.
Fit determines whether the inquiry belongs in the new-business process. It does not describe how far the inquiry has progressed.
In the running example, the inquiry may become a reasonable fit after the owner confirms the requested service, location, timing, and basic commercial fit. That judgment belongs in the fit field even though the work has not been won.
Current stage#
A simple current-stage sequence is enough for many service businesses:
- Received
- Reviewed or contacted
- Consultation, estimate, quote, or booking
- Decision pending
The middle stages should match how the business actually works. A contractor may use “site visit” or “estimate sent.” A consultant may use “fit call” or “proposal sent.” A clinic may use “appointment booked.”
Do not force every business into one universal sales pipeline. Current stage only needs to show where an open inquiry stands. Our example might move from received to contacted to estimate sent while the fit classification remains reasonable.
Outcome#
Record the final outcome separately once it is known. A small honest set may include:
- won;
- lost;
- disqualified or not a fit;
- cancelled;
- another business-specific final result.
An open inquiry should have no final outcome yet, or an explicitly open state when the existing record system requires one. Do not treat “unresolved” as a final outcome equivalent to won or lost.
If the example inquiry later becomes booked work, record won as the outcome and add the reason that matters to the business. Until that happens, it remains an open inquiry at its current stage—not a win and not a loss.
Record the likely source and its confidence honestly#
Use three source-confidence states:
- Known: Supported by a reliable form value, booking record, tracked call route, referral record, or similarly concrete evidence.
- Likely: Context points strongly toward a source, but the connection is not definitive.
- Unknown: There is not enough reliable information to assign a source.
“Unknown” is a valid result. Invented precision is not.
A page, service, or offer mention can establish useful context. It does not by itself prove whether the person arrived through Google Ads, organic search, a referral, a partner, or another channel.
That distinction matters in the running example. “I found you on Google” plus a service-page visit may support Google search as the likely source. It does not establish Ads versus organic search. Keep the confidence at likely—or the source unknown—unless stronger evidence supports a more precise assignment.
A booking record tied to a tracked campaign may support a known source. A clear referral note may do the same. A phone inquiry with no reliable route, referral record, or other usable evidence may remain unknown.
Do not turn a weak clue into a confident channel assignment because the report looks cleaner that way. The unknown-source rate is useful precisely because it shows where source-path visibility is missing.
When the meaningful action itself is missing, duplicated, or being used incorrectly in advertising, return to the guide on what should count as a Google Ads conversion. That is an event and platform-use problem, not a reason to guess the source later.
Separate lead volume from lead quality#
A practical fit judgment asks whether the inquiry shows:
- a relevant requested service;
- an acceptable location or service area;
- basic eligibility;
- plausible timing;
- realistic commercial fit;
- genuine intent.
The useful progression is:
Total inquiries → qualified inquiries → active opportunities → won work
A source can produce a high number of inquiries and little useful work. Another can produce fewer inquiries but a stronger share of qualified opportunities. The raw lead count does not resolve that comparison.
Not every poor-fit inquiry is a marketing failure. Poor fit may expose broad targeting, unclear messaging, missing fit boundaries, weak intake, or ordinary market noise. The record helps distinguish a pattern from an isolated mismatch.
When paid traffic repeatedly produces irrelevant searches, use the Google Ads search-term triage guide. When the source is plausible but the search, offer, page, or resulting lead does not align, use the offer-alignment diagnosis.
Review only what supports a decision#
A weekly or monthly review should answer four questions:
- Which sources are producing qualified inquiries?
- Which sources produce activity but little useful work?
- Where do otherwise good leads stop progressing?
- How many records still have an unknown source or no final outcome?
A restrained set of calculations is usually enough:
- Qualification rate = qualified leads ÷ reviewed leads. The denominator is the leads that have actually received a fit judgment, not every unreviewed record.
- Win rate = won work ÷ qualified leads with a final outcome. Still-open opportunities do not belong in a completed-outcome win rate. When won and lost are the only relevant completed outcomes, the equivalent calculation is won ÷ (won + lost).
- Unknown-source rate = unknown-source leads ÷ reviewed leads. The denominator is the leads whose source evidence has been reviewed.
These are working ratios, not universal benchmarks. Their value is in comparing the same business over time or locating a specific break, not proving that one percentage is “good” for every service category.
Route the first observed failure correctly#
| What the record shows | First place to investigate |
|---|---|
| Inquiries are reported but never arrive | Form delivery or event implementation |
| Inquiries arrive but nobody owns them | Intake and follow-up process |
| Useful leads arrive but the source is usually unknown | Tracking and source-path visibility |
| One source produces mostly poor-fit leads | Targeting, search terms, offer, or qualification |
| Qualified leads repeatedly stop before an estimate, quote, consultation, or booking | Response or sales process |
| Quotes happen but work is rarely won | Offer, fit, price, trust, or sales process |
| Strong work is won but platforms cannot connect it back | Advanced offline measurement |
Not every downstream problem is a marketing or website problem. A clean source record cannot fix slow follow-up. Better tracking cannot repair an offer that rarely wins. A sales-process problem should not be disguised as another landing-page project.
Keep the first version small#
This is not a CRM migration or enterprise reporting project. Advanced attribution, offline imports, forecasting, and similar work may matter later, but they come after the business can consistently follow an inquiry through likely source, source confidence, fit, current stage, and outcome.
When the break crosses forms, event tracking, source data, lead delivery, and follow-up, Web Ops can trace the connected path and identify the first repair.
Take the next 10 inquiries—or the most recent 10—and record likely source, source confidence, fit, current stage, and outcome. Fix the first field that cannot be answered consistently before building a larger reporting system.
The better state is not perfect attribution. It is being able to explain which activity produces useful work, where good inquiries stop, and what needs attention next.