On this page
Most lead reports answer the easiest questions: How many inquiries arrived? How quickly did someone reply? How much did the channel cost?
Those numbers matter, but they do not tell an owner whether the lead-response process is working. A fast automated greeting can still lead to a stalled conversation. A high lead count can hide spam, wrong-fit jobs, weak handoffs, or follow-up that never happened.
After years of looking at lead generation for local service businesses, I prefer a short weekly scorecard that follows each inquiry from arrival to a useful business outcome. It gives the owner, dispatcher, marketing team, and agency one shared view of where leads are moving and where they are getting stuck.
The goal is not a giant dashboard. It is a repeatable 20-minute review that produces one or two specific fixes for the next week.
Start with one clear lead funnel
Before calculating rates, define the stages your team will use. Keep them simple enough that dispatchers and managers apply them consistently.
A practical funnel is:
- New lead: A unique inquiry entered from a supported source.
- Engaged: The customer replied or a live call connected.
- Qualified: The inquiry is real, in the service area, and suitable for a next step.
- Handed off: A person or connected system received the qualified lead with context.
- Booked: The customer accepted an appointment, estimate, or other defined commitment.
- Completed or closed: The work was completed, lost, declined, duplicated, or otherwise resolved.
Calls and message leads behave differently. What should stay consistent is the outcome language.
Google's Local Services Ads reporting guidance reflects this distinction. Its reports include charged leads, lead type, spend, and booking information, and U.S. and Canadian advertisers can mark a lead as booked with details such as job type and customer name. That makes the platform useful for source reporting, but your CRM or operating system should still be the final record for revenue and completed work.
Clean the denominator before judging performance
Every rate depends on what you count as a lead. If one platform sends a duplicate, a vendor solicitation, and a real homeowner inquiry, those should not be treated as three equal sales opportunities.
Use a short exclusion list:
- Exact duplicates
- Obvious spam or solicitations
- Test inquiries
- Existing-customer service conversations that were not new opportunities
- Records created by an integration retry rather than a new customer action
Keep wrong-fit inquiries visible. An out-of-area request or unsupported job type is still useful information about targeting. Mark it as real but unqualified instead of deleting it from the report.
This prevents two common mistakes: inflating lead volume with noise and hiding targeting problems by removing every inquiry the team did not want.
Track seven measures each week
A useful scorecard connects responsiveness, conversation quality, team execution, and business outcomes. These seven measures are enough for most local service teams.
1. New and valid leads
Record total inquiries, then separate valid inquiries from excluded noise.
The gap is a diagnostic signal. A sudden rise in duplicates may point to an integration problem. More out-of-area requests may suggest that targeting or service-area settings need attention.
Show the numbers by source, location, and job type when volume supports it. Do not bury a weak channel inside a healthy company-wide total.
2. Time to first meaningful response
Measure from lead arrival to the first response that helps the customer move forward.
A receipt such as “We got your message” may reassure the customer, but it should not automatically count as meaningful. The first useful reply normally acknowledges the need, asks a relevant question, offers an approved next step, or connects the customer with the right person.
Track both the median and the slowest responses. The median describes the normal experience. The slowest group exposes coverage gaps around lunch, evenings, weekends, shift changes, or particular sources.
InstantResponse.AI's current reports and analytics show first-response time by source and time of day. That is the right starting point, but the weekly review should still sample conversations to confirm that the quick reply was relevant.
3. Customer engagement rate
Use this formula for message leads:
Engaged message leads ÷ valid message leads that received a meaningful response
Engagement means the customer replied after the initial business response. For calls, use a separate connected-call or answered-call measure.
If response time improves but engagement falls, review the opening message. It may be too generic, too long, repetitive, or asking several questions at once.
Segment engagement by source and service type. An emergency plumbing request should not be compared directly with a remodeling prospect gathering estimates for next season.
4. Qualification rate
Use:
Qualified leads ÷ valid leads
Define “qualified” before the meeting. For many contractors, it means the requested service is offered, the property is inside the service area, and the customer has provided enough information for a real next step. It does not have to mean the job is booked.
A falling rate can indicate broader targeting, seasonal demand, outdated service rules, or weak questions. Review a sample of qualified and unqualified conversations before changing the channel.
List the leading unqualified reasons: outside service area, unsupported job type, timing mismatch, price-only request, duplicate, or no response after follow-up. Reasons turn a percentage into an action.
5. Handoff acknowledgment
Automation can qualify a lead perfectly and still fail if nobody owns the next step.
Track:
Handoffs acknowledged within the team target ÷ total handoffs
The acknowledgment can be a dispatcher claiming the conversation, a salesperson calling, or a CRM status change—whatever reliably proves ownership. Set different targets for emergencies, routine service, and estimate requests.
InstantResponse.AI's centralized dashboard keeps the conversation, source, tags, status, and activity together and supports one-click human takeover. Use that context to review whether handoffs are accepted, not merely sent.
If acknowledgment is weak, changing the customer-facing reply will not solve the problem. Fix ownership, schedules, escalation, or notification rules.
6. Follow-up completion
Some customers do not answer the first question. Others provide details and pause before choosing a time. The scorecard should show whether the agreed follow-up process actually ran.
Use:
Follow-up actions completed ÷ follow-up actions due
Define “due” using your approved playbook. Stop when the customer opts out, the job is closed, the request is no longer relevant, or a person takes over.
Also sample the wording. Repeating the same message three times is technically complete and operationally weak. Each follow-up should add clarity, offer a simple next action, or make it easy to close the conversation.
7. Qualified-lead cost and booked outcome
Marketing cost should follow the same funnel definitions.
Calculate:
Cost per qualified lead = channel spend ÷ qualified leads
If booked-job data is complete and reliably joined to the source, also calculate:
Cost per booked job = channel spend ÷ booked jobs
Do not estimate booking or revenue when records are incomplete. InstantResponse.AI reports lead volume, response times, qualified outcomes, and cost per qualified lead; revenue remains in the CRM. Google LSA can also track leads that advertisers mark as booked. Reconcile those systems instead of presenting an unsupported blended number.
For channels with small weekly volume, use a four-week or monthly view before making budget decisions. One large job or one quiet week can distort a short period.
Read the scorecard as a sequence
The value is not any single percentage. It is the pattern across stages.
| Pattern | Likely area to inspect |
|---|---|
| Slow response, healthy qualification | Coverage, connection, or staffing |
| Fast response, low engagement | Opening message relevance or tone |
| Healthy engagement, low qualification | Targeting, service rules, or intake questions |
| Strong qualification, weak handoff | Ownership, alerts, or escalation |
| Handoffs accepted, few bookings | Sales follow-through, availability, pricing, or tracking |
| Bookings present, missing revenue | CRM process or source attribution |
These are starting hypotheses, not automatic conclusions. Read several real conversations before changing the workflow.
A five-conversation sample makes the dashboard concrete: one fast win, one slow response, one unqualified lead, one failed handoff, and one lead that stopped replying. The numbers identify where to look; the conversations explain what happened.
Run a 20-minute weekly review
Use the same agenda every week.
First five minutes: confirm data quality
Check duplicates, missing sources, stale statuses, and unassigned records. Confirm that the team is using the funnel definitions consistently.
Next ten minutes: find the largest drop
Compare the current week with the prior four-week range. Focus on one meaningful gap rather than discussing every metric.
Ask:
- Which source, location, service, or time window changed?
- Is the change large enough to matter?
- Do sampled conversations support the same explanation?
- Is the problem customer-facing, operational, or measurement-related?
Final five minutes: assign one change
Choose an owner, deadline, and expected signal.
Examples include adjusting an intake question, changing an after-hours owner, fixing a broken source tag, reviewing a slow handoff queue each morning, or training the team on one repeated objection.
Write the change into the next scorecard. Without an owner and follow-up date, the review becomes reporting theater.
Avoid scorecard traps
Do not rank employees from a tiny sample or treat one difficult lead as a trend. Do not compare branches without accounting for service mix, hours, and volume. Do not reward speed if the message is inaccurate. Do not reward qualification volume if the team labels weak inquiries as qualified to improve a dashboard.
Most importantly, do not let the tool define the process. The business must decide what qualified, handed off, booked, and closed mean. Software should make those definitions easier to apply and measure.
Build the first version in 30 days
In week one, agree on funnel definitions and exclusion reasons. In week two, confirm source, timestamp, owner, qualification, and outcome fields are captured. In week three, run the scorecard without changing anything and audit the data. In week four, choose the first operational improvement and measure it against the same definitions.
Keep the first version simple. Seven reliable measures beat 30 inconsistent ones.
A weekly lead-response scorecard should connect marketing spend to real conversations, clean handoffs, and business outcomes. When every team uses the same stages, the report stops being a pile of channel metrics and becomes an operating tool: where the process slowed down, why it happened, who owns the fix, and what should improve next week.