Cost per usable prospect is the cost of researching a list, including review time, divided by the distinct businesses that pass your written acceptance rules. It answers a practical agency question: how much did it cost to produce a shortlist worth investigating?
This worksheet measures research output. A usable prospect is not a customer, a booked meeting or proof of buying intent. The numbers below are invented to demonstrate the calculation, not results from Margo or a comparison test of real vendors.
Define a usable prospect before comparing costs
Write down one niche, one territory and the evidence required for the next research action. For example: an active independent accountancy practice in the selected city, a confirmed public website or contact route, and a specific observation that makes your service worth investigating.
Accept a business only when it is distinct, within scope and supported by enough current evidence for that next action. A missing website alone does not establish demand. An email address alone does not establish permission to contact someone.
Keep this research label separate from your CRM stages. HubSpot's lifecycle-stage documentation distinguishes leads, sales-qualified leads and opportunities. Our worksheet uses its own earlier acceptance rule; it does not certify any of those stages.
If you are still deciding which kind of tool to test, start with our agency lead-tool comparison. This worksheet is the calculation to use once your test and acceptance rules are fixed.
The formula: include the work after the export
Use these two calculations:
Research cost = allocated tool cost + additional research charges + review hours multiplied by your chosen hourly cost.
Cost per usable prospect = research cost divided by accepted distinct businesses.
Use one currency and the same cost basis for both workflows. Allocate the relevant share of a subscription consistently. Do not count a subscription's included credits again as a separate cash purchase. Additional enrichment or verification charges belong in the calculation when they were actually incurred.
Keep cash spending and the value assigned to your own time visible as separate lines. An hourly cost is your planning assumption, not a promise that every saved hour becomes billable work. Count setup, cleanup, deduplication and evidence review when they belong to the test. If an integration will be reused, state how you allocated its setup cost rather than silently excluding it.
If no business passes, the result is undefined: zero usable prospects. Do not report a zero-dollar cost per usable prospect.
Worked example: the cheaper export can cost more to use
Imagine two workflows researching the same niche and territory, each returning 100 rows. These are hypothetical inputs in US dollars, with an assumed review-time cost of $25 per hour. The workflows are labelled A and B; neither represents a named product.
Apply exclusion reasons in order, recording the first applicable reason for each rejected row. This makes the buckets mutually exclusive.
| Sample disposition | Workflow A | Workflow B |
|---|---|---|
| Exported rows | 100 | 100 |
| Duplicate rows excluded | 10 | 5 |
| Outside the agreed niche or territory | 15 | 15 |
| Business confirmed inactive | 5 | 5 |
| Insufficient evidence for the next action | 20 | 20 |
| Accepted distinct businesses | 50 | 55 |
The accepted counts are 100 minus the four exclusion buckets: 50 for A and 55 for B. A business with several problems appears in one rejection bucket only. An unknown operating status is not evidence that a business has closed; record uncertainty and investigate or defer it.
| Cost input or result | Workflow A | Workflow B |
|---|---|---|
| Allocated tool cost | $40.00 | $65.00 |
| Additional research charges | $10.00 | $5.00 |
| Total data charges | $50.00 | $70.00 |
| Review time | 2 hours | 0.5 hours |
| Review-time cost at $25/hour | $50.00 | $12.50 |
| Total research cost | $100.00 | $82.50 |
| Data charges per exported row | $0.50 | $0.70 |
| Cost per usable prospect | $2.00 | $1.50 |
A looks cheaper if you divide data charges by exported rows. B costs less per accepted business in this illustration because the assumed review time and rejection counts differ. That is a consequence of these inputs, not evidence that a particular tool is faster or more accurate.
On a cash-only basis, A costs $50 divided by 50, or $1.00 per usable prospect; B costs $70 divided by 55, or about $1.27. The preference changes when you include time. Show both views so you can make a decision that matches your actual constraint.
Copy this prospect-review worksheet
Use one row per candidate business. Keep private contacts and client notes in your approved systems; the worksheet does not require publishing them.
| Field | What to record |
|---|---|
| Business identity | Public name, location and website or source link |
| Scope decision | Does it match the prewritten niche, territory and exclusions? |
| Relevant observation | A specific, dated fact; keep the service hypothesis separate |
| Evidence checked | Source URL, check date and any contradictory information |
| Disposition | Accept, defer or exclude, with one primary reason |
| Review minutes | Time spent checking and preparing the next action |
| Next research action | The question still to verify before progressing |
For this calculation, deferred records stay outside the accepted count. You can investigate them later and produce a new, dated result rather than quietly changing the original test.
Then record the run totals: exported rows, duplicates, other exclusions, deferred records, accepted distinct businesses, tool-cost allocation, additional charges, setup/review time and the hourly-cost assumption. Reconcile every exported row to one final disposition before calculating the denominator.
Use Margo as one input to the test
Margo's agency workflow supports local-business research by niche and geography. Inspect the available evidence against your own acceptance rules instead of treating returned rows as automatically usable prospects.
The Margo API reference documents that fields can be absent, email availability is separate from verification, and research credits depend on the source and current policy. Record the actual charges and limitations shown for your research; do not assume one requested result equals one complete record or one credit.
This worksheet is a method you can use in a spreadsheet alongside Margo. It is not a claim that Margo automatically calculates this metric. Read the product documentation for the supported workflow before using an integration.
What should you decide after the test?
Keep the acceptance rule unchanged while comparing results. Record the niche, territory, date and selection method so that another review can explain the result. A small pilot tells you about that sample; it does not establish a universal data-quality ranking.
If the lower cost comes from accepting weaker evidence, you have changed the output rather than improved the economics. If the cost is high because most records are out of territory, fix the targeting before buying more volume. If review time dominates, identify the repeated check that might be simplified while preserving the acceptance rule.
Finally, track replies, meetings, customers and collected revenue separately when they actually occur. This research metric cannot establish customer acquisition cost or return on investment by itself.
To try the method with a focused local-business search, start with Margo, choose one niche and territory, and review the first bounded sample against your written rules.
Sources and method
Product and terminology sources checked on 1 November 2026: Margo API reference, Margo agency workflow, and HubSpot lifecycle stages. The formulas, worksheet and hypothetical A/B example are Margo editorial work, prepared with AI assistance and independently reviewed before publication. No customer dataset or paid research call was used.