Almost every outbound list is built the same way: pick an industry, pick a headcount band, pick five job titles. It is the fastest filter available in every tool, which is exactly why it carries so little information — everyone targeting your market is sending to the same people.
Across the programmes we run, these five signals consistently outperform title as predictors of a reply.
1. A recently opened role in the relevant function
A company hiring for a problem is a company that has admitted it has one, has secured budget, and has told the world publicly. It is the closest thing to a declared intent signal available for free.
The nuance: seniority of the role matters more than volume. One open leadership role in the function usually means a mandate. Five junior roles often just means churn.
2. A leadership change in the last ninety days
New leaders re-evaluate. The first two quarters of a tenure are when vendor decisions get reopened, tooling gets reviewed and budgets get reallocated. Reaching someone in month two is a fundamentally different conversation from reaching them in year three.
This signal is also self-limiting, which is good — the window closes, which forces your engine to act on it rather than letting it sit in a list.
3. Tooling adoption or removal
A company that just adopted an adjacent tool has budget and appetite in that category. A company that just removed one has a gap. Both are actionable; the second is rarer and stronger.
Be careful with detection quality here. Tool-detection data has wide accuracy variance between providers, and a confidently wrong claim about someone's stack in a first email is worse than no personalisation at all.
A confidently wrong claim about someone's stack is worse than no personalisation at all.
4. Funding, but only the right kind
Funding is over-used because it is easy to buy. Raw "raised a round" is weak — every vendor in the market hits the same list within a week. What works better is funding combined with a second signal: raised and now hiring in your function, or raised and adopted an adjacent tool. Compound signals cut through because almost nobody bothers to build them.
5. Prior engagement your own systems already know about
The strongest signal available to you is one you already own and probably discard: an account that opened your last campaign three times, or visited pricing twice, or downloaded something last quarter.
These accounts convert at multiples of cold. They are also usually sitting in a system nobody has connected to the sequencer. Before buying another intent product, check whether your own engagement data is being used at all.
How to combine them
Do not build an elaborate weighted model. Count signals. Accounts with three or more get high-touch treatment, personalised research and priority in paid. Accounts with one get a standard sequence. Accounts with none wait.
A simple count beats a sophisticated score for the same reason simple ICPs beat elaborate ones: everyone on the team understands it, so everyone actually uses it.