Every database for sales leads quotes a price per record, and a record is not what you are buying. You want a person who fits your market, still works where the file says, and has an address that takes mail. A file of 1,000 records can hold far fewer of those, and the price per record never tells you how many.
There are three kinds you can actually buy: a list file from a broker, a subscription to one company's own index, and a lookup run across data sources when you add each person. Each prices the gap between records and usable leads differently. This page works through all three with one worked example, then gives you a 100-record test to run before you pay for any of them.
If you would rather start from a search than a file, the people search in Find leads builds the list from your filters and checks each work email as the person is added. You can search your market and count the matches for free before you spend a credit.
What is a database for sales leads?
A database for sales leads is any source you can filter down to companies and the people to contact at each one, with a way to reach them: a work email, a phone number, a LinkedIn profile. The term covers a spreadsheet a broker emails you and a search box with hundreds of filters. That range is part of why buyers end up comparing the wrong numbers.
Your CRM holds the people you already know. New names have to come from somewhere else, and that somewhere is what gets sold as a lead database.
What turns a record into a usable lead comes down to three checks. The rest of this page leans on them:
Fit. The company and the role match the market you sell to.
Still there. The person still works at the company the record names.
Delivers. The email address accepts mail, which a deliverability check can tell you before you send.
A record that fails any one of them was paid for and goes nowhere.
Business contact database, list file or live lookup
Here is how the three kinds compare on what you pay for and what arrives.
List file | Subscription to an index | Lookup when you add a person | |
|---|---|---|---|
What you buy | A fixed set of records | Search and export from one company's index | A search, then each detail looked up as the person is added |
How you pay | Per file or per record, up front | A recurring fee for a set allowance of exports or credits | Per result found |
Checked as of | The seller's last update | The seller's last refresh, then your export date | The day you add the person |
A record with no email | Paid for anyway | Depends on the seller's export rules: ask | In Oneprofile, a work email lookup that finds nothing costs nothing |
Fit to your market | Industry codes and size bands | The seller's filters | Your filters, with the matches counted before you pay |
Best for | A rough first map of a new market | High monthly volume from one index | Small teams that want each lead checked as it arrives |
The list file is the oldest kind and the simplest. You pick an industry, a region and a headcount band, pay, and download. So when should a team of one or two running cold email still buy one?
"When you're sizing a market you've never sold into. A file is a fine first map of who's out there. Just don't put it straight into a sequence."
Utku Zihnioglu, CEO of Oneprofile
The trouble with a file in a sequence is that it was last checked on the seller's schedule, not yours.
A business contact database sold by subscription fixes part of that. The seller keeps refreshing its index, and you can search it whenever you like. But every export you take is a snapshot, and from the day it lands in your CRM it ages like a bought file. You are also living inside one index's coverage.
The third kind hands you no stored copy. You set filters, see how many people match, and only then pay for the details you need. In Oneprofile, running the people search and counting the matches are both free. Find work email then checks several data sources in turn for each person you add, in an order set by accuracy, not price, and stops at the first answer.
One opinion that sells nothing: index size is the least useful number on any data pricing page. A small team sells into a few thousand accounts, and a count in the hundreds of millions says nothing about whether yours are in it. Your own 100-record test does.
Sales leads for sale: the cost per usable lead
Sales leads for sale are priced per record. The number that matters is the cost per usable lead: what you paid, divided by the leads that pass all three checks.
Here is a scenario. Say you buy 1,000 records and test 100 of them:
Check | Passed | Still usable |
|---|---|---|
Fit: right company size and role | 85 of 100 | 85 |
Still there: same employer as the record | 65 of 85 | 65 |
Delivers: the address verifies as valid | 52 of 65 | 52 |
Fifty-two usable leads out of 100 means each usable lead costs about 1.9 times the price per record. A file that looked cheap per record can come out dearer than a subscription that looked expensive, and the other way round. The figures above are an illustration, not a benchmark, so run the division on your own sample.
A lookup changes which side of the division you pay on. In Oneprofile, Find work email costs up to 10 credits for an address it finds, and a work email lookup that finds nothing costs nothing. Verify email costs up to 0.8 credits a check. A found and verified address is up to 10.8 credits, which on the Team plan, at $20 per 1,000 credits, is up to $0.216 a lead.
Fit is settled before you spend, since the filters decide who is counted. Two checks still apply after the lookup. An address that is found but fails verification was still paid for, and a role is worth spot-checking on a sample here too.
Free leads for sales: what you actually get
Free leads for sales come from three places:
Downloadable free lists. In practice these are the files that have been passed around the most, so expect the delivers check to hurt.
Free tiers of lead tools. Oneprofile's Free plan is 200 credits a month, 1 seat, up to 3 integrations, 3 tables and 100 rows per table. At up to 10.8 credits for a found and verified work email, that is at least 18 such leads a month. Here is what 200 free credits buy on one worked list.
Your own CRM. Closed-lost deals and trial sign-ups that never bought already know your name. Pulling them out is covered in building a lead list from a CRM export.
The third is the one we would start with, and it costs nothing.
How to test a database for sales leads before buying
Ask every seller for 100 records pulled with your own filters, not their showcase sample. Then put each kind through the same checks:
Write the filters down first. Industries, headcount band, region and the job titles you sell to. Send the identical filters to every seller you compare.
Check fit. Read each record against those filters. To score them instead, the ICP scorer rates each company against the ideal customer profile in your business context and gives its reasons, at 3 credits a lead.
Check still there. Open the LinkedIn profile for 20 of the people who passed fit and compare the employer. Read a LinkedIn profile does the same for every person, at up to 19.6 credits each.
Check delivers. Import the 100 as a CSV and run Verify email. It returns valid, risky, catch-all, invalid, disposable or unknown, at up to 0.8 credits a check, so 100 addresses fit inside the Free plan. Count only valid as a pass, and keep catch-all apart as a soft yes.
Divide. Scale the usable share of your sample to the full list, then divide the list's price by it.
Which of the three checks fails most often on a seller's sample?
"Still there fails most. People change jobs faster than files get refreshed, so on a bought sample I'd expect a fifth or more to name an employer the person already left."
Utku Zihnioglu, CEO of Oneprofile
Test a lookup the same way: run your filters, look at the count, then add 100 people with Find work email ticked and run Verify email on the addresses it finds. That is up to 1,080 credits for the 100, because each lead is charged only for the data source that answered, and nothing when none does.
Anyway. The test takes an afternoon, and it is the one number no seller can print for you.
Keeping a database for sales leads current in your CRM
Whichever kind you buy, the clock starts at export. HubSpot's database decay simulation, built on MarketingSherpa research, puts contact decay at 2.1% a month, about 22.5% over a year. On 1,000 contacts, about 21 go wrong in the first month alone. That is the same aging the subscription section mentioned, and it applies to a lookup too, from the day the lead lands in your CRM.
"I look at where the list came from and how old it is. A sequence can't fix a list someone exported three months ago for a different campaign."
Utku Zihnioglu, CEO of Oneprofile
A routine that keeps a list honest:
Record where each contact came from, and when. A source property and an import date in the CRM answer the first question anyone asks when replies drop.
Verify again before each campaign. In Oneprofile, a column can refresh on a schedule under Run conditions, and the editor shows what a row costs a month before you save.
Let new matches arrive on a schedule. A search attached to a table can run daily, weekly, monthly or quarterly. Each run prices the people the table does not hold yet and waits for you to click Approve.
Write back without duplicates. Export with Create or update records, matched on email, so a second push updates the CRM record instead of creating another. Run the export again after a refresh: it pushes when you run the column or save changes to it, never because a cell changed.
The longer view is in how CRM data decay builds up.
So which kind should you pick? A file maps a new market once. A subscription suits a team pulling thousands of contacts a month from one index. For a small team sending cold email, the lookup is the one we would choose: you count before you pay, each work email is checked as the person is added, and the same list goes to your CRM. Make every seller pass the same 100-record test, us included. To run it on a lookup, count the people in your market and check your first leads free.

