How to Test a New Lead Source: A Pilot Plan You Can Actually Review
November 5, 2026 Comment off
To test a new lead source, define the decision before buying the first record. Write down the product being tested, the conditions your team will hold steady, the spending limit and the evidence that would justify another step. A pilot should help you decide whether to stop, fix a problem, gather more evidence or expand cautiously.
This is a test-design worksheet for a source that has already passed initial review. Use our lead vendor evaluation guide for the earlier due-diligence conversation. A seller’s promises and a successful delivery demonstration do not establish how the product will perform in your business.
Choose one primary question
“Is this a good vendor?” is too broad. “Can this source deliver our agreed inquiry type during staffed hours, and do those inquiries progress under our normal sales process?” is testable. If you also change pricing, offer, territory, routing and the sales team, the result will be difficult to interpret.
Write the proposed explanation in advance. For example: the new source may reach customers whose needs better match your service. Then identify what you would expect to observe and what else could explain it. More appointments might reflect a stronger team assignment rather than a different audience.
Complete a one-page pilot charter
| Decision field | What to record | Why it matters |
|---|---|---|
| Product version | Format, filters, sharing terms and source identifiers | Keeps a later product change from becoming part of the original test |
| Operating conditions | Receiving hours, team, routing and approved follow-up process | Makes differences in handling visible |
| Exposure limit | Maximum spend and volume, with a named person able to pause | Prevents an open-ended purchase |
| Primary outcome | A defined customer milestone and its observation window | Prevents choosing a flattering metric afterward |
| Diagnostic measures | Delivery, acceptance, contact and progression counts | Locates where a problem occurs |
| Review decisions | Stop, repair, extend or expand, with the evidence required | Turns the report into an action |
Use your own costs and capacity to set limits. There is no universal number of leads that makes every pilot reliable. The outcome frequency, variability and size of the difference you care about affect how much evidence you need. Treat a small pilot as exploratory; use appropriate analytical support before making strong statistical claims.
Run a delivery rehearsal before the paid pilot
Use clearly marked synthetic records to test acceptance, rejection, duplicate handling, routing and reporting. Confirm that tests cannot trigger real customer outreach or enter paid performance reports. Reconcile a test record through the supplier’s system and your CRM before opening the live flow.
Separate operational stop conditions from performance review conditions. Missing records, unexpected charges or a broken receiving route need prompt investigation. A slower-than-expected sales cycle requires a different decision. Do not keep purchasing through a broken process merely to reach a target sample size.
Keep a change log while the pilot runs
Record a dated change whenever the supplier changes the product or your team changes handling. Include staffing gaps, revised filters and routing outages. A test with a material change may need separate before-and-after groups; combining them can hide why results moved.
Where a fair comparison is feasible, use similar operating conditions and observation time. Document remaining differences rather than presenting a commercial trial as a controlled experiment. A new source delivered on a holiday week is not directly comparable to an older source worked by a full team in a different period.
Use a decision log, not a winner announcement
Consider a hypothetical pilot with 80 accepted leads. Forty arrive in week one and forty in week two. At the first review, the earlier group has had twice as long to progress. A combined sales rate may be useful as a current snapshot, but it is not a finished comparison of the groups.
A useful note would say: “Delivery passed. Seven records missed the assigned follow-up process. Customer outcomes are still developing. Fix the assignment issue and review both groups at the same age before expanding.” That note is more actionable than “The conversion rate looks low.” These figures illustrate the review process, not a benchmark or recommended sample size.
Decide what the evidence permits
- Stop: the product does not meet the agreed requirement or the economics cannot support another test.
- Repair: a specific delivery or handling failure prevents a fair review.
- Extend: the process works but important outcomes remain immature or uncertain; approve any additional spend explicitly.
- Expand: observed outcomes support a limited next step, with new caps and a review date.
Recheck capacity before increasing volume. The process that handled a small pilot may not handle a larger order. Keep your customer economics and actual handling costs in the expansion decision.
Bring a specific test question to LGW
Compare approaches with people who work on buying and delivering leads. Review eligibility and apply for LGW’s $395 Lead Buyer Pass. Application and approval are required; the pass is for qualifying end buyers purchasing for their own use, not lead sellers, publishers, agencies or call center providers.