Deal sourcing data providers, and which we pay for.

We publish our origination benchmarks including the parts that do not flatter us, so it seems fair to publish the layer underneath them too. These are the deal sourcing data providers we actually pay for, what each one is genuinely good at, and the single test that decides between them.

No affiliate links, no vendor paid for a slot, and the numbers below are from our own systems rather than from anyone's marketing page.

The test that actually separates providers

Almost every comparison of deal sourcing data providers argues about record counts. Record count is close to meaningless for this job.

Here is the test that matters, and it is narrow: can the provider cover one to twenty person firms, at partner level, in the geography you actually sell into?

That is the shape of this market. If you sell origination services, your buyers are small private equity firms, boutique M&A advisories and independent sponsors, most of them under twenty people with a one-page website. If you are buying companies, your targets are owner-operated businesses that have never filled in a form. Both are the long tail. Both are exactly what large-company-biased databases handle worst.

A database indexed for enterprise SaaS selling will return the firms that need you least, quickly and in volume, and it will feel like it is working. We have covered why that produces a quiet quarter in why deal origination stalls.

What we use, and for what

GetLeads, for partner-level coverage and depth of filter

GetLeads is what we reach for when the segment is defined by title and size band rather than by industry, and when the geography is outside the United States.

Since 9 July 2026 we have pulled 93,427 contacts across 24,497 companies through it, in 34 separate lists, spanning retail, solar and renewables, wealth management, asset and investment management, private equity and M&A, and specialist medical practices. The breadth is the point: the same tool handled all of those without us changing vendor per vertical.

The pull that made the case was European. 14,188 net-new partner-level contacts at European private equity, M&A and investment banking firms, in a single pull, none already in our system. European coverage at partner level in the one to two hundred employee band is precisely where most US-built databases thin out, and it is the gap that had been limiting us.

On quality, the number we care about is bounce. Two campaigns built on those lists have been sending long enough to judge: 5,994 emails into US and Canadian PE and M&A firms at 0.97% bounce, and 1,248 into Australian firms at 0.40%. At partner level in small firms, under one percent is a good result, because that segment is where bounce rates normally go bad.

It also has an MCP integration, which matters more than it sounds: the list gets built inside the system that will send it, rather than exported to a spreadsheet where it starts aging immediately.

AI Ark, for cheap breadth and lookalike expansion

AI Ark is what we use when we need a lot of companies fast and cost per contact is the binding constraint, particularly for lookalike expansion off an existing customer. It is meaningfully cheaper per contact than anything else we run, which changes what is affordable when you are building a list of several thousand companies to then filter hard.

We use it differently from GetLeads: AI Ark for company-level breadth, GetLeads where the person and the title are the hard part.

Public filings, where the jurisdiction allows it

In the United Kingdom, Companies House gives you officers, persons with significant control, charges and accounts, free and public. That beats any commercial estimate of ownership, and it changes how a UK list should be built. We set that out in UK deal sourcing providers compared. In the United States there is no equivalent, which is why US private company data is inference sold back to you.

What we do not pay for

We do not pay for a platform whose value is a deal marketplace, because that is on-market flow and prices accordingly. The distinction is in off-market versus auction pricing and in deal sourcing marketplace versus proprietary origination.

We also do not treat any provider's email as trustworthy on arrival, including the ones above.

Verification is your job, not the vendor's

This is the part most comparisons skip, and it costs people their sending infrastructure.

No data provider's emails should be sent to unverified. Of the 93,427 contacts we pulled through GetLeads, 236 came back invalid and 411 undeliverable when we ran them through our own verification, and every one of those was excluded before a single send. That is a normal rate at that volume for any provider. Catching it is the sender's job.

Roughly a third of everything we pull gets excluded before sending, across ICP fit, do-not-contact rules and verification. If that sounds wasteful, consider the alternative: a 3% bounce rate takes a mailbox pool down, and rebuilding one costs more than the data did. The metrics worth watching are in deal origination metrics, and the ones that mislead are in deal origination vanity metrics.

How to choose, in order

Define the universe before you shop. If you cannot describe the target in terms of size band, title and geography, no provider will save you. Most list problems are thesis problems wearing a costume.

Test coverage on the hard part, not the easy part. Ask for a sample in your narrowest segment and smallest size band, in your least-covered geography. Every provider looks good on US companies with 200 employees.

Check the person, not the company. Company coverage is commoditised. Partner-level and owner-level contact coverage is not.

Assume you will verify everything. Budget for it, build it into the pipeline, and never let a provider's own status field decide what you send to.

Then ignore record count entirely. The useful question is how many rows survive your filtering, not how many arrived. Ours is about two thirds, and we would rather it were lower.

Frequently asked questions