A lead score of 83 looks precise. If nobody can explain why it is 83, whether 82 changes the action, or whose assumptions shaped the weights, the precision is theatre.
A small gym can score leads using a few explainable signals: recency, explicit action, service and schedule fit, stated intent, and lawful contactability. Cap the model, document weights, exclude sensitive traits, keep a human override reason, and validate against attended visits and good-fit memberships.
This guide treats lead handling as an operating system, not a collection of persuasive tricks. The core owner question is: Can a staff member explain the score, the action threshold, and the evidence without using sensitive or proxy discrimination?
Place this decision inside the Cluster E system
The ten guides separate failures that are often blended into one conversion number:
Start with the lead-response SLA, then use the lead-response SLA, lost-lead reason codes, and pipeline forecast where their decisions apply. A strong result in one stage cannot repair bad data or a broken handoff in another.
Define the record before measuring it
| Term | Working definition |
|---|---|
| Signal | An observed fact tied to a sales action. |
| Weight | The transparent points assigned before looking at individual outcomes. |
| Threshold | The score band that changes priority or next action. |
| Override | A documented human decision with a controlled reason. |
| Validation | Comparison with downstream outcomes, errors, and segments over time. |
Write these definitions into the operating checklist. Otherwise one staff member will count a message as contact, another will require a call, and the monthly chart will compare different events under the same name.
Run the workflow with a named owner
| Step | Required action | Accountable role |
|---|---|---|
| Choose | Select five or fewer signals that staff can observe reliably. | Sales owner |
| Weight | Assign small understandable points and state why each changes attention. | Owner and team |
| Exclude | Ban health, body, caste, religion, disability, and other sensitive or proxy attributes. | Privacy owner |
| Route | Connect bands to capacity-aware actions, not labels. | Shift lead |
| Validate | Compare score bands with contact, attendance, joins, early churn, and false positives. | Owner |
The owner column cannot say “team.” Several people may contribute, but only one role should be accountable for the next due action. If responsibility changes at a shift handover, the new owner must accept it rather than inherit an invisible queue.
Start with an explainable score
| Signal | Evidence rule | Example points | Expiry | Risk check |
|---|---|---|---|---|
| Recent explicit enquiry | Timestamp in owned channel | +3 | Decays quickly | Channel bias |
| Booked or requested visit | Confirmed action | +4 | Until outcome | Duplicate booking |
| Schedule and service fit | Stated need matches verified delivery | +2 | Review if offer changes | Staff assumption |
| Contact permission and valid detail | Evidence of expected channel | +1 | Until withdrawn | Consent scope |
| Repeated non-response | Defined attempts in valid window | -2 | Reset on new action | Over-contact |
Blank cells are intentional. Fill them from the gym’s actual offer, hours, capacity, policies, and records. A copied industry number can be a question to investigate, but it is not a local operating standard.
Work through a fictional example
A fictional gym scores 120 open leads. The top band has a high contact rate but a weak attended-visit rate because a form-completion signal is overweighted. The owner lowers that weight and raises the value of a prospect-selected visit time. The model becomes less impressive on paper and more useful to the shift team.
The example is deliberately fictional and is not an industry benchmark. Its purpose is to show the calculation or control path. Replace every assumption with an observed record from your gym and preserve the source period.
Measure the whole outcome, not the easiest event
Track:
- Lead count and staff capacity in each score band
- Contact, booked, attended, paid, and early-retention rates by band
- False positives: high score with no meaningful action
- False negatives: low score that becomes a good-fit member
- Overrides by reason and by staff member
Use consistent denominators. A contact rate, an attended-visit rate, and a paid-join rate answer different questions. Keep source, staffed period, offer, and sample size visible before comparing teams or weeks.
Protect the prospect and the quality of the decision
- Never score protected or sensitive characteristics or inferred health status.
- Do not use budget assumptions based on neighbourhood, name, language, device, or appearance.
- Do not let a score trigger unlimited automated contact.
- Do not hide weights from staff who must act on them.
- Do not call correlation proof that a signal causes conversion.
Fitness enquiries can reveal sensitive context even when a person never writes a diagnosis. Collect the minimum needed for the current decision, restrict access, define retention, and move exercise or medical screening into a separate qualified process when it is genuinely required.
Implement this in one operating week
- Inventory reliable fields and delete speculative ones.
- Define a five-signal draft on paper.
- Back-test against prior attended visits and good-fit joins.
- Set action bands based on actual team capacity.
- Review errors monthly and retire useless signals.
At the weekly review, inspect a small sample of the underlying calls, messages, visit records, stage timestamps, and collections. A tidy dashboard cannot compensate for ambiguous events or invented reasons.
Frequently asked questions
What is gym lead scoring?
It is a transparent method for ordering follow-up using observed signals such as recency, explicit action, fit, intent, and contactability. It should guide attention, not judge a person’s worth.
What data should a gym not use for lead scoring?
Exclude health, body, injury, disability, religion, caste, ethnicity, gender, and other protected or sensitive traits, plus weak proxies that can create unfair treatment.
Does a higher lead score mean someone will join?
No. It means the current model assigns higher priority under stated assumptions. Validate against downstream outcomes and track false positives, false negatives, and changes over time.
