If your plumbing business got 50% more ad clicks last month but booked the same number of jobs, what would you change? Before raising the budget or rewriting the ad, trace what happened between the click and the booking. A monthly review should give you one thing to investigate or test.
1. Compare groups of inquiries that had the same time to progress
Use the date each request first came in to assign it to a group. Compare the same number of days or similar service periods, and give both groups the same time to reach each stage. If you use 14 days to count bookings, wait until the newest request has had 14 days, or label that group as still open. Log any holiday closures, changes in service area or staffing gaps alongside the numbers.
Google Ads explains that conversions can arrive after the original ad click, so recent cost per conversion can look higher before those conversions are recorded. Your own booking delay may differ. Use completed periods for decisions and show recent activity separately as provisional. [1]
2. Set one definition for each step
Pull your call and form records into the same review. Count one inquiry per distinct job request: a form and follow-up call about the same leaking pipe belong together. Keep spam and internal tests separate. Agree what makes a request qualified, such as a job you actually handle at an address you serve. Keep requests you couldn't assess in a pending row; a missed call with no voicemail doesn't establish fit.
| Measure | What to record |
|---|---|
| Spend and traffic | Record media spend and ad clicks from the ad account. Keep agency fees and staff time separate unless you include them consistently. |
| Distinct inquiries and qualification | Count actual requests after deduplication; mark qualified, out of scope or pending review. |
| Response coverage | Divide qualified requests that received a human response within your chosen window by qualified requests. Include missed targets. |
| Bookings and new customers | Count confirmed bookings and new paying customers separately. Count a customer once and document how cancellations are handled. |
| Unknown source | Count these requests and outcomes in overall totals; leave the source unknown until you have evidence. |
Google Ads distinguishes a qualified lead from a lead that reached a later step, such as a sale. You can apply the same distinction in a spreadsheet or in the workflow your team uses with OG JobSite. Record one row per inquiry with first-contact date, source evidence, qualification, first response and next step. [2]
Google Analytics documents how campaign tags can populate traffic-source fields, including session source and medium. A tagged visit alone doesn't establish who booked a job. If you can't reliably connect a request to a channel, leave it unknown instead of attributing it to the campaign with the most traffic. [3]
3. Find the stage that deserves a closer look
| Measure | Earlier period | Later period |
|---|---|---|
| Media spend and clicks | $400 / 200 clicks | $480 / 300 clicks |
| Distinct inquiries tied to paid search | 20 | 22 |
| Qualified inquiries | 12 | 10 |
| Qualified requests answered within target | 9 of 12 (75%) | 6 of 10 (60%) |
| Bookings / new paying customers | 6 / 4 | 6 / 4 |
| Media spend per known qualified inquiry | $33.33 ($400 ÷ 12) | $48.00 ($480 ÷ 10) |
In this fictional example, clicks rose 50%, but known qualified inquiries fell from 12 to 10 and bookings stayed at six. Suppose there were also three requests from an unknown source in the later period. Keep them outside the paid search row. With just ten qualified requests tied to the campaign and a few unknowns, you can't tell whether the ad attracted worse prospects or the difference came from normal variation, staffing or incomplete tracking.
The response row gives you something concrete to check: four qualified requests in the later group missed your response target. Check their timestamps, whether they landed outside staffed hours and who owned the callback. If the gap clusters around lunch, test assigning a backup owner at lunch for the next comparable period. Check response coverage first, then bookings as the group matures.
Your next action: trace one request that stalled
Pick one recent, complete period. Match a small sample of calls and forms to your booking records. Choose one stalled request and write down the last step you can verify, who owned the next step and what you'll change. Set a date to review the same measures again. If you can't trace a request from the public website to an owner, start with the plumbing Request Path Map below.
Sources & research notes
Sources and their access dates are listed below. Survey findings describe their own samples; examples and calculations in this guide are illustrative.
- About conversion lag reporting
Google Ads Help · Primary guidance checked November 9, 2026
Accessed .
Explains delays between ad clicks and conversions and how recent cost per acquisition may appear inflated.
- About qualified leads and converted leads
Google Ads Help · Primary guidance checked November 9, 2026
Accessed .
Distinguishes lead qualification from later conversion milestones defined by the business.
