Start from the reason for the outreach, not the size of the list
A number is not a brief. "Five thousand SaaS companies" tells a researcher almost nothing about which five thousand, and the gap gets filled with whatever a database returns. Start instead with the sentence you would say to the prospect: what problem you solve, for which kind of company, and who inside that company owns it. Every criterion downstream should be traceable back to that sentence. If a filter cannot be justified by it, it is decoration.
Write criteria that can actually be checked
There is a difference between a criterion and a wish. "Companies investing in growth" cannot be confirmed by anyone; "companies with a named head of demand generation" can. Useful criteria tend to be observable: industry and business model, geography, headcount band, whether they sell to businesses or consumers, whether a particular function exists in the company at all. Agree the unconfirmable ones out of the brief before research begins, rather than discovering at handover that they were approximated.
Qualify the company before you research anyone inside it
This is the step teams skip, and it is the expensive one to skip. A database category is a label someone applied once, not evidence of fit. Checking the company first — what it actually sells, who it sells to, whether the function you need exists there — costs a few minutes per account and removes the records that would otherwise waste an hour each in the CRM. Contact research on an unqualified account is work done twice: once to find the person, once to delete them.
Choose people by responsibility, not by title
Titles are inconsistent across companies and often across departments in the same company. A Head of Operations at a forty-person business may own the thing you sell; the same title at a four-thousand-person business almost certainly does not. Use department, seniority and the actual scope of the role together, and accept that the right contact will sometimes be two levels below the most senior person on the org chart. A list of impressive titles is not a list of relevant people.
Verify last, and keep the uncertainty visible
Verification belongs at the end, because verifying contacts at companies you have not qualified is paying to confirm addresses you will not use. When it does run, the result should reach you as a status rather than as a silent filter: what passed, what is uncertain and why, and what could not be confirmed at all. Standard valid deliveries typically hold a bounce rate below 1 percent under normal sending conditions. Records that could not be confirmed should not be counted as usable contact delivery at all.
Decide the structure before the file arrives
The last avoidable cost is reformatting. Agree the field names, the value formats, how segments are marked and how verification status is carried, before delivery rather than after. A file that imports is worth noticeably more than the same data in a shape someone has to reshape, and the reshaping is where records quietly get lost.
Related on this site