Cold calling as a data-sourcing channel

The companies with the richest operational data are almost never looking for someone to license it to. They are busy shipping product, serving customers or closing down in an orderly way. If an AI lab wants that data, someone has to go and ask, and for this kind of request the phone is still one of the most effective ways to do it.

Most data acquisition starts with inbound: a marketplace listing, a web form, a partner network. Those channels work for companies that already know their data has value and are actively trying to sell it. That is a small and unusual slice of the market.

The typical company with five years of Slack, Jira, GitHub and email history has never considered licensing it. The founder of a company that is winding down is thinking about payroll, creditors and customer notices, not about the archive that will vanish when the subscriptions lapse. Neither will find a marketplace on their own. We look at the gap in more detail in why inbound data marketplaces have coverage gaps.

The result is that a lab relying on inbound gets whatever happens to show up, not what its sourcing brief asked for.

Email and LinkedIn are useful, and we use both. But licensing company data is an unfamiliar idea, and unfamiliar ideas raise questions. Is this safe for our employees? What about our customers? Who actually sees the data? How much of my time does it take?

A short email cannot answer all of that without becoming a long email nobody reads. A conversation can. In a few minutes a rep can explain the basics, hear the specific concern and answer it directly:

• that every person, company, email and account is replaced with a consistent token, and credentials, card numbers and government IDs are removed
• that HR, payroll, medical and privileged folders are excluded by default, and the company sets its own exclusions before any data moves
• that the licence is for training and evaluation only, with no marketing, profiling or re-identification allowed
• that the raw archive is deleted within 30 days of sign-off, with a deletion certificate
• that the seller's total time is about an hour

A call also tells the rep things that are hard to learn any other way. Whether the person on the line has authority to sign. Whether the company actually ran on the tools the target list suggested. Whether there is real interest or polite deflection.

Cold calling only works if the list is right. Before anyone dials, we build a target list from company and technographic data against the lab's sourcing brief: sector, headcount, tool stack, years of history, region and whether the company is operating or winding down.

Then the rep needs to reach the person who can actually make the decision. In an operating company that is usually the founder, CEO, COO or CFO. In a company that is closing it might be a director, officer, liquidator or assignee with authority over company assets. Calling the wrong person, however friendly, mostly produces a forwarded email and a long wait.

Timing helps too. Signal-driven prospecting means paying attention to events that make a conversation more relevant, such as a company announcing a wind-down, a change in leadership or a shift in its tool stack. A call that arrives when the topic is already on someone's mind lands very differently from one that arrives at random.

Martal Group has spent 17 years doing B2B outbound, and that experience is what we bring to data sourcing. More than 200 onshore sales reps work across the US, Canada, the EU and LATAM, with experience prospecting in 24 industries. They use a power dialer, signal-driven prospecting, and coordinated email and LinkedIn outreach so that the call is one touch in a sequence rather than a lone interruption.

Onshore matters. A finance leader in Toronto or an operator in Madrid is more likely to stay on the line with someone who works in their market, understands the local business context and can speak plainly about what is being proposed.

The goal of the first call is modest. It is not to close a licence. It is to book a 15-minute intro call where the process, the scope and the safeguards can be explained properly. From there the company sets its scope and exclusions, a read-only evaluation takes under an hour of its time, anonymization happens in-house, and the seller reviews a sample before anything is delivered.

Cold calling has a reputation, some of it earned. Done badly it is pushy and repetitive. Done well it is short, respectful and useful to the person on the other end, even when the answer is no.

In practice that means a rep who can explain the idea in under a minute, who stops when someone is not interested, who does not overstate what the data is worth (there is no firm number before counting, and typical licences range widely, from $20K to $400K+), and who follows up only when invited. Many good conversations start with "not now" and come back months later when a company's circumstances change.

For a lab, the benefit is coverage. Instead of waiting for the right companies to appear, the brief becomes a list, the list becomes conversations and the conversations become signed companies, with data evaluated, anonymized and delivered on a rolling basis.

If you are an AI lab with a specific data need, send us a sourcing brief and we will show you how a campaign against it would run.