How to test a lead generation company before you sign anything

Part 4 of the lead generation companies buyer's guide. A one-week, low-cost test that tells you whether a vendor's data and process actually work before you commit.

7 min read · Published September 19, 2026

The short answer

The short version.

Part 4 of the lead generation companies series: the final check before you sign.
Take a free or small sample, dial 20 records yourself in one sitting, and record how many reach the named person.
Then run the smallest paid month at a narrow profile and judge it on connects per dial and meetings per connect, not on how the spreadsheet looks.
Any vendor that cannot survive a one-week test cannot survive a twelve-month contract.

This is the last part of the buyer's guide series. Part 1 defined the market, part 2 gave you the comparison questions, part 3 the pricing. Part 4 is the test itself, because everything before this point is the vendor talking and everything after is your own data.

Step one: the 20-dial sample test

Take the vendor's free sample, dial 20 records yourself in one sitting, and write down how many reach the named person. Eight or more means a fresh, mobile-weighted source.

Twenty calls takes under an hour and no specification sheet survives contact with it. Log three outcomes per dial: reached the named person, reached someone else, or dead number. A strong sample hits the named person on eight or more of twenty. Two or three means a warehouse export, whatever the invoice calls it.

Do this yourself even if someone else will do the real calling. You are calibrating what good data feels like, and that calibration is worth more than any clause in the contract.

Step two: the smallest paid month

Buy the minimum monthly volume against one narrow profile, pace delivery daily, and have the real caller work it for four weeks while logging every dial.

The test month has one job: produce your two numbers, connects per dial and meetings per connect, on the vendor's data with your real caller and your real script. Five hundred contacts at $75 is enough. Keep the profile narrow so a bad result is diagnosable, and insist on daily delivery so no one faces the whole month at once.

Change nothing else during the test. Same script, same caller, same offer. You are measuring the data, and every other change contaminates the result.

Step	Effort	What it proves
Dial 20 sample records	Under an hour	Data freshness and phone coverage
Smallest paid month, narrow profile	4 weeks of normal calling	Real connect and meeting rates
Flag wrong numbers as you go	Minutes	How replacements actually work
Review the two numbers	An hour	Whether to scale, change or walk

The one-week vendor test at a glance

Step three: test the service, not just the data

Flag wrong numbers during the test month and watch what happens. Replacements that arrive without a fight are worth as much as the data itself.

Every vendor has wrong numbers. The question is what happens next. Flag them as you go and see whether replacements arrive promptly and free within fair-use limits, or whether the process involves a ticket, a review and an argument. You are buying a monthly relationship, and the replacement process is the part of it you will touch most often.

Deciding: scale, change or walk

Scale if meetings arrive at an affordable cost, change the profile if connects are fine but meetings are not, and walk if the data fails twice.

A good test month ends with arithmetic, not feelings. Meetings at a cost your customer lifetime value supports means scale the volume. Healthy connects but no meetings means the problem is the conversation or the profile, so change one and retest. Poor connects on a second sample, from a vendor that argued about the first replacement, means the market has other options and so do you.

Scale: cost per meeting works, raise the volume.
Change: connects fine, meetings missing, fix one variable.
Walk: the data failed twice. It will not improve on a bigger contract.

Key takeaways

Dial the sample yourself. An hour of your own calls beats every reference check.
Run one clean test month, changing nothing but the data source.
Judge the replacement process as hard as the data. You will use it every month.

See the data behind the advice

Five ICP matched prospects with mobile numbers, direct dials and work emails, free. Same pipeline that fills a paid account, no credit card.

Get 5 free sample leads See pricing

Frequently asked questions

How long does it take to test a lead generation company?

The sample test takes under an hour. A full test month takes four weeks of normal calling. Anything shorter judges noise, anything longer is just using the service.

What is a good connect rate on bought data?

On a fresh, mobile-weighted B2B list, roughly 8 to 15 percent of dials reach the named person. Under 5 percent on a tested sample points to the data.

Should I sign an annual contract to get a better rate?

Only after a successful test month. Annual billing earns its discount, typically 20 percent, once you know the data works for your profile. Before that it is a gamble, not a saving.

How do I start the LeadNumbers test?

Take five free sample prospects at /sample, dial them, then start at 500 contacts for $75 if the sample holds up. No contract, and volume changes month to month.

Keep reading

Economics