Skip to content
Worksheet

Lead-quality diagnostic worksheet (Meta Ads)

By: Lithe Agency · Companion to: Cheap leads, few customers · Version: 29 September 2026

This is a read-only template. Copy it into your own document or spreadsheet before filling it in. Use totals only: no names, phone numbers, emails, lead IDs or other personal data.

This worksheet is a template to copy. Empty cells are blanks for your own copy. This page has no input fields, does not autosave and sends nothing.

1. Cohort and reporting settings

Fill this in before looking at results.

FieldYour entry
Campaign / ad sets comparedBlank, to fill in
Lead created from (date)Blank, to fill in
Lead created to (date)Blank, to fill in
Time zone used for cohort datesBlank, to fill in
Ads Manager report: attribution setting and date basis as shown in the reportBlank, to fill in
Ad account time zoneBlank, to fill in
Source of captured records (Instant Form export / CRM / spreadsheet)Blank, to fill in
Maturity window (days after lead created, based on your own past time to purchase)Blank, to fill in
Results read on (date, no earlier than "lead created to" + maturity window)Blank, to fill in
Exclusion rules (what counts as duplicate, test, invalid contact data)Blank, to fill in
Follow-up rule (attempts, channels, time limit)Blank, to fill in
"Reached" means (two-way contact: answered, replied, spoke)Blank, to fill in
Qualified means (one or two sentences agreed by marketing and sales, including fit such as service area)Blank, to fill in
Customer means (paid? net of refunds or cancellations?)Blank, to fill in
Anything changed during the cohort (offer, form, price, team, qualification rule)Blank, to fill in

If the qualification rule changed during the cohort: re-score the whole cohort under one frozen rule if your records allow it. If they don't, split the cohort at the change date and treat the parts as not comparable.

2. Platform count and captured records

These two numbers may use different bases. Reconcile them before calculating any rate from the platform count.

ItemAd set 1Ad set 2Notes
Spend (cohort dates)Blank, to fill inBlank, to fill inBlank, to fill in
Platform-reported leads (Ads Manager)Blank, to fill inBlank, to fill inRecord the settings in section 1
Captured records (export / CRM)Blank, to fill inBlank, to fill inSame campaign, grouped by created date
DifferenceBlank, to fill inBlank, to fill inExplanation: date basis, attribution, duplicates, deleted records, other

If your ads click to WhatsApp: enter the messaging results Meta reports in the platform row and your team's captured enquiry records before exclusions in the captured row. Record duplicates in section 3 and subtract them only once when calculating workable enquiries. These are not interchangeable counts. Use the Notes column for the exact metric and denominator, the attribution setting and your deduplication rule, and explain the difference before calculating any rate. The labels below use lead-form wording, so adapt them to your enquiries; a cost per form lead is not the same as a cost per conversation.

3. Funnel buckets

Each row group adds up to the row above it. Every lead sits in exactly one bucket.

StageAd set 1Ad set 2Check
Captured recordsBlank, to fill inBlank, to fill inBlank, to fill in
Excluded: duplicatesBlank, to fill inBlank, to fill inBlank, to fill in
Excluded: test entriesBlank, to fill inBlank, to fill inBlank, to fill in
Excluded: invalid contact data (malformed or confirmed non-existent)Blank, to fill inBlank, to fill inNot "didn't answer"
Workable leadsBlank, to fill inBlank, to fill in= captured − all exclusions
Reached (two-way contact)Blank, to fill inBlank, to fill inBlank, to fill in
Attempted, not reached (agreed attempts used up)Blank, to fill inBlank, to fill inUnknown quality
Not attemptedBlank, to fill inBlank, to fill inUnknown quality. Reason, if known:
Attempts still in progress (not reached yet, attempts left)Blank, to fill inBlank, to fill inShould be 0 once the follow-up rule's time limit has passed
SubtotalBlank, to fill inBlank, to fill in= workable leads
QualifiedBlank, to fill inBlank, to fill inBlank, to fill in
Reached, not qualified: outside service areaBlank, to fill inBlank, to fill inFit decided here, not during cleaning
Reached, not qualified: other reasonsBlank, to fill inBlank, to fill inTop reasons:
Reached, qualification not yet decidedBlank, to fill inBlank, to fill inBlank, to fill in
SubtotalBlank, to fill inBlank, to fill in= reached
CustomersBlank, to fill inBlank, to fill inCount each lead once, even with several deals
Qualified, paid then refunded / cancelledBlank, to fill inBlank, to fill inOnly if customers are counted net; otherwise 0
Qualified, lostBlank, to fill inBlank, to fill inBlank, to fill in
Qualified, still openBlank, to fill inBlank, to fill inBlank, to fill in
SubtotalBlank, to fill inBlank, to fill in= qualified
Customer source recorded for all deals? (yes / no)Blank, to fill inBlank, to fill inIf no, mark customer counts missing

4. Calculated rows

How to fill empty or zero cells:

  • 0 / 0%: the numerator is 0 and the denominator is known and above 0. This is a real result.
  • n/a (denominator 0): for example, CPQL when there are 0 qualified leads. Write "n/a: 0 qualified, spend X." In a mature cohort, that is an actual outcome.
  • missing: the data was not recorded.
  • not yet mature: the maturity window hasn't ended. Write the date you will read it again.
MetricFormulaAd set 1Ad set 2
Cost per platform-reported leadSpend ÷ platform-reported leadsBlank, to fill inBlank, to fill in
Cost per workable leadSpend ÷ workable leadsBlank, to fill inBlank, to fill in
Attempt coverage(Reached + attempted, not reached + attempts in progress) ÷ workableBlank, to fill inBlank, to fill in
Reach rateReached ÷ workableBlank, to fill inBlank, to fill in
Qualified rate (reached)Qualified ÷ reachedBlank, to fill inBlank, to fill in
Qualified rate (workable)Qualified ÷ workableBlank, to fill inBlank, to fill in
Out-of-area shareOutside service area ÷ reachedBlank, to fill inBlank, to fill in
Close rateCustomers ÷ qualifiedBlank, to fill inBlank, to fill in
CPQLSpend ÷ qualifiedBlank, to fill inBlank, to fill in
Ad spend per customer (proxy, not full acquisition cost)Spend ÷ customersBlank, to fill inBlank, to fill in

Data check before diagnosing:

  • Not checked: Platform-reported leads and captured records are reconciled, or the difference is explained
  • Not checked: Every bucket subtotal matches the row above it
  • Not checked: All stages use the same cohort dates and time zone
  • Not checked: Customer counts are complete, or marked missing
  • Not checked: If one more or one fewer sale would reverse the conclusion, treat the result as low evidence and keep collecting

5. Where to investigate and what to change

The biggest drop shows where to look. It does not prove the cause.

QuestionYour entry
Stage with the biggest dropCaptured → workable / workable → reached / reached → qualified / qualified → customer
HypothesisBlank, to fill in
Evidence that supports itBlank, to fill in
Evidence that would rule it outBlank, to fill in
People to involve (roles only)Blank, to fill in
The one change for the next cohort of new leadsBlank, to fill in
Budget during the test (cap, no increase)Blank, to fill in
Start dateBlank, to fill in
Review date (after the maturity window)Blank, to fill in
Review metrice.g. qualified ÷ workable, reach rate, attempt coverage, ad spend per customer
Result that would make us reverse or stopBlank, to fill in
Other things that changed in the period (season, platform learning, audience mix)Blank, to fill in
Result at reviewBlank, to fill in

A before-and-after comparison is not a controlled experiment. Record anything else that changed so you can weigh it at review.


Questions about reading these numbers? You're welcome to email Lithe Agency at admin@litheagency.com, sharing totals only. If your records aren't ready, you don't need this worksheet to start: you can chat with us on WhatsApp instead, without sending your leads' personal data.

Meta Ads services

Allow Google Analytics to measure visits, contact clicks and CTA tests when enabled? Your choice and any test assignment last up to 7 days. You can use the site if you reject.

Privacy Policy · Change your choice anytime in the footer settings.