Start with the decision the review must support
A list of lost leads can contain several different problems: a roofing enquiry outside the service area, an unanswered callback, a postponed renovation and a quotation declined by the customer. Combining them into one loss rate makes it difficult to decide whether to change targeting, qualification, follow-up or estimating.
Begin with one question, such as whether the business is attracting work it can deliver or where qualified projects stop progressing. The workflow below is a proposed operating method for home-service teams. It does not establish industry benchmarks or explain an individual customer’s motives without evidence.
Define outcomes before counting them
Keep the project outcome separate from contact activity and the reason recorded. HubSpot’s lifecycle documentation illustrates this distinction: it identifies opportunities through a deal association and maintains separate lead-status options for attempted contact, connection, qualification and timing. A spreadsheet can preserve the same separation without adopting a particular CRM.
| Category | Evidence needed | Treatment |
|---|---|---|
| Disqualified | A confirmed mismatch with service, area, scope or another written acceptance rule. | Record the rule; review targeting and screening separately. |
| Unreachable | Logged contact attempts completed under the agreed contact policy, with no two-way exchange. | Use for enquiries that never became accepted opportunities. If an accepted opportunity is administratively closed, record closed lost with the customer reason unknown. |
| Closed lost | An accepted opportunity has ended without a sale, with closure evidence or a documented administrative closure rule. | Separate customer-confirmed reasons from inferred or unknown reasons. |
| Pending or deferred | An active next step, unresolved decision or agreed future review date. | Keep outside closed-outcome calculations. |
| Incomplete | Missing information prevents reliable classification. | Flag data quality and assign a record-repair action. |
| Closed won | The business’s written sale criterion has been met. | Use a consistent criterion, such as accepted contract and required deposit. |
Build a reviewable project record
Use one record per distinct project opportunity. Link repeated calls, forms and messages to it; otherwise a persistent homeowner can look like several failed leads. Preserve separate projects for an existing customer instead of deduplicating solely by phone number. Remove spam, tests and accidental duplicates from the enquiry population using explicit rules, while retaining an exclusion count.
Capture project ID, enquiry date, recorded source, requested service, service location, scope, acceptance decision and owner. Add first-response timestamp, contact attempts with channel and outcome, two-way contact date, survey date, quotation date and version, quoted value, next action, closure date and outcome. For each reason, retain the supporting message or note, its date, and whether it came from the customer or staff interpretation. Keep source attribution unknown when it cannot be verified.
Completeness does not guarantee accuracy: the Government Data Quality Framework explicitly distinguishes them. A filled-in “price” field is weak evidence when nobody recorded what the customer actually said.
Run the review in a consistent sequence
Save the review date, selection rules and an unchanged export so later corrections can be explained. Restrict access to supporting records to the people who need them; use project IDs in the shared review sheet.
- Choose the population. For operational closure trends, review opportunities closed during a stated period. For enquiry-to-sale conversion, group enquiries by arrival period and report outcomes as of a stated date. Do not mix the two denominators.
- Reconcile coverage. Compare the export with enquiry channels and scheduling records. State which channels or dates are missing. Reconcile the selected population to mutually exclusive outcomes plus pending and incomplete records.
- Check classification. Read the timeline before accepting the status. A quote with no reply is not evidence that another contractor won. An enquiry still awaiting its first action is pending work, not a completed unreachable sequence.
- Code reasons conservatively. Use a short list such as service mismatch, outside area, timing, capacity, price, scope disagreement, competitor chosen and unknown. Allow secondary factors, but select one primary reported reason for a reason-share calculation. Preserve the customer’s wording where available.
- Identify an action. Assign an owner, due date, eligible project group and verification method. Separate repairing missing notes from changing the sales process. Review previous actions before opening more.
Calculate rates with visible exclusions
Always show counts alongside percentages and return “not calculable” for a zero denominator. These are descriptive definitions for this workflow, not universal CRM conventions.
- Closed-opportunity win rate = closed-won opportunities / (closed-won + closed-lost opportunities) × 100. Exclude disqualified enquiries, pending projects and unreachable leads that never became opportunities. Show those counts separately.
- Customer-confirmed reason coverage = closed-lost opportunities with a customer-confirmed reason / all closed-lost opportunities × 100. Low coverage limits any explanation of the loss mix.
- Recorded reason share = closed-lost opportunities assigned a particular primary reason / all closed-lost opportunities × 100. Include an unknown category so missing reasons remain visible.
- Enquiry-cohort won share to date = enquiries meeting the sale criterion by the review date / valid unique enquiries in the arrival cohort × 100. Report pending and incomplete shares alongside it; a recent cohort’s figure is provisional.
Blank loss-review template
Copy this into a sheet or CRM note. Complete one entry per project; leave unsupported conclusions unknown.
Review date: ____ Project ID / owner: ____ Population and reporting period: ____ Enquiry date / source / service / area: ____ Qualification rule and evidence: ____ Contact attempts and two-way contact evidence: ____ Survey / quotation dates and version: ____ Current outcome / closure date / closure rule: ____ Primary reason / secondary factors: ____ Reason evidence, date and source: ____ Customer-confirmed, staff-inferred or unknown: ____ Missing information / repair owner / due date: ____ Pending next action / review date: ____ Proposed process change / owner / due date: ____ Measure, eligible population and comparison period: ____ Result, limitations and next decision: ____
Use patterns to choose a test, not declare a cause
Compare similar services, areas, project sizes and decision windows before interpreting a difference. Source groups can attract different work; urgent repairs and planned renovations need different observation periods. Small groups can change sharply after one outcome.
If unclear quotations repeatedly appear in loss notes, test a clearer scope-and-exclusions section and track its use, questions received and closed outcomes. If service-area mismatches recur, review the advertised coverage and intake questions. These are hypotheses suggested by records. NIST’s statistical handbook cautions that association does not establish cause and effect. A before-and-after improvement can also reflect seasonality, staffing, pricing or a changed project mix.
Record what changed and when, retain comparable groups where practical, and describe results as observed differences unless the evaluation supports a stronger claim. A useful review ends with a specific next decision and evidence plan, including a decision to collect better data before changing the business.
