Skip to main content
Almost every behaviour in the signal library can be given research, on top of the sources it already uses. The two do genuinely different jobs.

Sourcing finds companies, research covers them

Sourcing searches broadly. It looks across everything being published for anyone showing the behaviour, which is how you discover companies you weren’t watching. But it only ever sees what surfaces — a company whose news nobody posted is simply missed, and you have no way of knowing it was there. Research goes company by company. It takes your question to each of your tracked companies in turn and works through the sources for that company. Nothing about those accounts is missed for want of a post: if the information exists, research finds it. So they answer different halves of the same question. Use sourcing for reach across the market, and research for completeness on the accounts you have decided matter. That is the trade: sourcing has no limit on who it can find, research has no gaps on the companies you track.

The two kinds

Social research

What a company and its people say about themselves, checked daily. Counts as 1 signal.

Web research

What the rest of the world publishes about them, checked monthly. Counts as 5 signals.
The two find different things, and most teams run both. Social research catches what a company says about itself; web research catches what gets written about it — often the only place a mid-market company’s news appears at all.

What research needs

Tracked companies

Research runs against the companies in your Tracked Companies list. It doesn’t search the world at large — it investigates the accounts you’ve said you care about. To track a company, click Track on its row in the companies table or on the company page itself. Your plan sets how many companies you can track at once; the current count is shown next to the Tracked Companies list.
With no tracked-company allowance, research is unavailable and the option is locked when you create a signal. A company you haven’t tracked is never researched.

A question worth asking

Research is driven by a question, not a keyword. When you add research to a behaviour, Clearcue writes the question for you from the behaviour and whatever you configured — “has this company recently hired a VP of Sales”, “has this company announced a funding round”. You can edit the question. Keep it answerable from public evidence, and keep it about the company rather than about your product.

Add research to a signal

1

Pick a behaviour

Start creating a signal as usual, and configure what you’re watching for.
2

Open the Research tab

Next to Sourcing, the Research tab shows what’s available for this behaviour.
3

Choose what to include

Each research option is a card you can select or deselect. You’re not obliged to take both, and each one you keep counts against your signal allowance.
4

Save

Research runs on its own schedule from then on.

Which behaviours support research

Research is available on the behaviours where a company’s actions are the subject: Engagement behaviours don’t support research. “Who liked our post” is a fact about an interaction that already happened — there’s nothing to go and investigate.

Behaviours that are already research

The Research tab also holds four behaviours that answer their question from Clearcue’s own data rather than from a written source: They need tracked companies for the same reason, but they take filters rather than a question, so there’s no prompt to write. Nobody posts “we just made our first hire in Germany” — which is exactly why these exist.

What you get back

Findings arrive as evidence on the company, alongside everything else that behaviour has detected. Each finding states what was found and links the sources behind it, so you can check the claim before you use it.
A finding is the answer to your question about that company, whether or not anyone posted about it — which is the point of running research alongside sourcing rather than instead of it.

Signal Library

Every behaviour, its sources, and whether it supports research.

Manage signals

Signal limits, and how research counts against them.