Inside a sourcing campaign: from brief to signed companies

AI labs that need company data often ask the same question: what actually happens between sending a brief and receiving data? The short answer is that a sourcing campaign is outbound sales work pointed at an unusual product. This post follows one hypothetical campaign from start to finish so the moving parts are easy to see.

Imagine a lab building agents for customer operations. It sends a sourcing brief asking for companies in B2B software, logistics and professional services, with 50 to 300 employees, at least four years of history, and a tool stack that includes a help desk, a CRM, email and either Slack or Teams. It wants companies in the US, Canada and Europe, and it is open to both operating companies and companies winding down.

Each of those fields does work later in the campaign. Sector and headcount shape the target list. Tool stack and years of history decide which companies are worth calling. Regions decide which reps make the calls. The operating or winding-down choice changes both who signs and how quickly the conversation moves. Our guide on how to write a good sourcing brief covers how to make each field precise.

The brief is turned into a target list using company and technographic data. Firmographic filters handle sector, headcount and region. Technographic data, meaning signals about which software a company uses, narrows the list to companies that ran on the tools the lab cares about. That step matters because tool stack is what determines connection depth: a company with a help desk, CRM and chat tool all running for the same years can produce connected records that follow one customer issue across systems.

The list is then ordered, not just filtered. Some companies show signs that make a conversation more timely, such as a wind-down, a restructuring or a pause in hiring. Those signals help decide who gets called first. We go into this in building a target list from technographic data and intent signals.

A good list is only useful if someone reaches the right person at each company. This is where Martal Group's background matters. It has run B2B outbound for 17 years, with more than 200 onshore sales reps across the US, Canada, the EU and LATAM, and experience prospecting across 24 industries.

For this campaign, reps would work the list through several channels at once:

• Cold calls using a power dialer, with signal-driven prospecting to prioritize the most relevant companies
• Email, written for founders, CFOs and operators who have never considered licensing their data
• LinkedIn, often to reach a director or officer at a company that has already wound down
• Referrals through advisors, such as lawyers, liquidators and investors, when they are involved

The person who can sign depends on the company's state. At an operating company it is usually an executive. At a closing company it may be a director, officer, liquidator or assignee with authority over company assets, and an authority document is requested up front. Finding that person is often the hardest part of the campaign, which is why reaching the person who can sign has its own post.

When a rep reaches the right person and they are interested, the next step is a 15-minute intro call. It explains what a licence is, what the lab will and will not be allowed to do, and how privacy is protected. The licence covers training and evaluation only, and buyers may never market to, profile, contact or re-identify anyone in the data.

If the company wants to proceed, it sets its scope and exclusions, such as a sensitive client, a channel or a folder. Those are annexed to the contract before any data moves. The evaluation is read-only and takes under an hour of the seller's time. It counts what is actually there: years of history, headcount, systems and how well they connect. That is also when the value is set. Typical licences range from $20K to $400K+, and there is no firm number before counting.

Some companies drop out at this stage. An archive might be thinner than expected, or a key system might have been cancelled. That is normal, and it is one reason a campaign targets more companies than the brief strictly needs.

Signed companies move into anonymization, done in-house. Every person, company, address, email, phone and account becomes a consistent token across every tool. Credentials, API keys, card numbers and government IDs are removed outright. Faces, logos and signatures are blacked out in images and PDFs. HR, payroll, medical and legally privileged folders are excluded by default, and direct messages stay out unless the company includes them and has the right to.

Each seller reviews an anonymized sample before anything is delivered, and payment goes to the company entity in 2 to 4 weeks. The raw archive is deleted within 30 days of sign-off, with a deletion certificate.

Delivery to the lab is rolling. Rather than waiting for every company to sign, the lab receives each dataset as it clears review. That gives the lab something to work with early, and it lets both sides adjust. If logistics companies are signing faster than professional services firms, the brief can be rebalanced, more calls can go to the slower segment, or the lab can decide the mix is good enough.

Looking across the whole process, a few things decide how well a campaign goes. A precise brief makes the list sharper. Accurate technographic data keeps reps from calling companies whose archive will not fit. Reaching the signer quickly matters most for companies winding down, where tools may be cancelled within weeks. And a light process for the seller, about an hour of their time in total, keeps interested companies from losing momentum.

None of this depends on companies finding the lab on their own. It depends on going to them, which is the argument made in why outbound beats waiting for data to come to you.

If your team has a sector, size or tool stack in mind and wants to see what a campaign would look like for it, send us a sourcing brief.