Customer acquisition cost: CPL, qualification and CAC

A cheap enquiry and a cheap customer are different outcomes. Dividing expenditure by completed forms is insufficient for a budget decision. Define an enquiry, a suitable customer and the expenses included before calculating. The same formula with different boundaries produces incomparable numbers. The model below is illustrative: US dollar amounts demonstrate arithmetic, are not agency prices and make no promise of payback.

1. Give each metric its own name

State boundaries in the report. If creative production is included, label the metric CPL including production rather than compare it with media-only CPL. A repeat purchase from an existing customer is not new acquisition. For full CAC, agree allocation of salaries, tools and shared work across channels. List excluded costs so the total does not appear more comprehensive than it is. Keep that method consistent when judging changes over time.

2. Build cohorts and allow time for outcomes

A cohort groups enquiries by acquisition time and consistent rules. August leads may pay in September: their sales belong to the August cohort, rather than September clicks. Display the last update and group age. Until the usual sales cycle has progressed, CAC is provisional. Dividing one week of spending by all sales that week mixes earlier enquiries and repeat customers with the new group. Track open opportunities instead of treating them as either guaranteed customers or permanent failures.

3. Check a simple worked example

Illustrative completed cohort: $2,400 in advertising, $600 of allocated creative costs and $600 for handling new enquiries. It generated 120 unique leads, 30 qualified enquiries and ten new paying customers. This model includes no other costs; a real full CAC must include the other agreed acquisition expenses. None of these numbers is a market benchmark for leads or a quotation for agency work.

If there are no new customers yet, CAC is not zero: division cannot produce the metric. Report expenditure, open deals and observation time. In a small cohort, one extra sale materially changes the average. Always display customer count and the cohort definition beside cost, so a precise-looking amount does not conceal fragile evidence.

4. Agree qualification before comparing channels

Qualification should reflect need, service area, offer suitability and a possible next step. Ask about budget only where it is needed for that decision. Assign an owner and consistent rejection reasons. When one manager qualifies after a conversation and another accepts every form automatically, qualified CPL differs because of reporting. Preserve status history so a later review does not silently rewrite an earlier report. Discuss borderline enquiries together before changing the rules for the next cohort.

5. Fix attribution and the record of truth

A customer may see an advertisement, return through search and finish a conversation in Telegram. The advertising platform, analytics and CRM can attribute the sale differently. Choose the model and attribution period for each report, record treatment of direct returns and do not add every platform claim as unique customers. Reconcile actual new paying customers with the agreed business system. Analyse channel influence separately from that customer total and explain missing source data rather than inventing it.

6. Compare CAC with contribution, not revenue

Subtract agreed variable fulfilment costs, refunds and other expenses affecting customer contribution from the payment. Revenue alone does not reveal what remains to cover acquisition. Include repeat purchases only from observed history for comparable customers, not a promise of future lifetime value. Assess time to payment separately: a channel with acceptable final cost may need cash before its opportunities generate receipts. A profitable-looking average does not answer that timing question.

7. Decide using comparable groups

• Compare the same service, geography and degree of cohort maturity.
• Show CPL, qualified CPL, customers and provisional or final CAC together.
• Separate traffic quality from unreachable contacts, response delays and sales terms.
• Adjust spending gradually and inspect the next cohort cost, not only the historical average.

Advertising planning: https://samsa.agency/en/services/paid-media. Enquiry journeys and technical barriers: https://samsa.agency/en/services/performance. Published projects: https://samsa.agency/en/cases. Metrics help choose the next testable action; increasing expenditure does not guarantee the previous customer cost will hold.

Official measurement sources

Google Ads qualification stages: https://support.google.com/google-ads/answer/11459091. Metrica attribution models: https://www.yandex.com/support/metrica/en/reports/attribution-model. The expense allocation and cohort boundaries above belong to the example. Check platform configuration separately against current documentation.