Intent signals: your first 30 days

In your first 30 days with intent signals, start small: pick two or three templates that match your offer, send the first messages in approval mode, read every reply, and review the results at the end of each week. Keep the signal that earns real replies and drop the rest. Sellinger has 26 templates in 8 categories, included with every bundle.

Intent signals are an easy thing to switch on and a hard thing to use well. The feed fills quickly, the temptation is to message everybody, and a month later you cannot say which signal worked. This plan keeps the first month small and measurable. It follows how we started on our own LinkedIn outbound, and it uses only the signal types that ship today.

What do you need before week one?

• A clear idea of the behaviour that shows interest in what you sell. Write it as a sentence: "People who comment on posts about [topic]."
• A segment or campaign where new leads should land.
• An agent set up with your knowledge base, persona and rules. See set up an agent.
• Approval mode on, so you read before anything sends.

If you have no idea what shows interest, start with the people who engage with your own content and your own founders' posts. That is the safest signal, because they already know you.

Which signals to pick?

The 26 templates sit in 8 categories.

For a first month, choose three from different categories, for example one from My Brand, one from Sales Opportunities and one from Company Activity. Different categories give you different kinds of reason to write, and you learn which kind your market answers. Our post on which intent signals are worth acting on ranks them by how closely each ties to a need.

Week 1: set up and read

1. Create the three signals from the template library. Fill in what to watch and the segment.
2. Switch on lead creation only if you are ready to see leads arrive. Matching people can be created as leads automatically.
3. Watch the feed for a few days without messaging. Read what comes in. Is it your market? Are competitors' staff in there? Exclude them.
4. Attach one signal's leads to a campaign, with approval on.

The aim of week one is not volume. It is to learn what the feed looks like and to fix obvious noise. If half the matches are students or competitors, narrow the signal before sending a word.

Week 2: send the first messages

Write a message for each signal that refers to the reason. For a person who commented on a post about [topic]: refer to the topic, ask one question, no link. For a company that just raised funding: congratulate them briefly and ask about the next hire or priority. The principle is the same in each case: the signal gives you the reason to write, and the message shows you used it.

Send the first 20 to 30 per signal in approval mode. Read each draft. Check:

• Does it use the signal truthfully, without pretending a relationship?
• Is it short, with one question?
• Would you be glad to receive it?

Our post on competitor engagers as intent signals covers one category in depth, including what not to say.

Week 3: read the replies

By now the first replies are in. Do three things.

1. Read all of them. Note which phrases from your message the person echoed.
2. Sort them. Positive, neutral, refusal, wrong person.
3. Compare the signals. Which of the three produced replies that say something?

Resist the urge to total up one blended number. The point is the comparison between signals. In our own work, the difference was large: in October 2025 signal lists replied at 55% and booked four times as often as untagged lists. Your numbers will differ, but if one signal is clearly ahead and another is clearly behind, you have your answer.

Week 4: decide

At the end of the month, make three decisions.

Then add one new signal to replace the one you dropped. Keep the total at three while you learn.

What did our own data teach us?

Two cautions come out of our case study. First, signal lists beat untagged lists at the start, but results are not constant. Replies held at about 40% all year, while meetings per person messaged peaked at 2.4% in November 2025 and drifted below 1% from June 2026. Audiences saturate. Second, cutting reply time from 19 to 2 minutes did not lift meetings: the list was the bigger lever. A good signal is a better list, and that is where the work belongs. See the full case study.

How should messages differ by signal?

The signal is the reason to write, so the first line should change with it.

Notice that none of them says "I saw you were looking at us". The line shows you read the signal without making a show of watching.

What are the common mistakes?

• Switching on ten signals at once. You will not know which worked.
• Messaging everyone in the feed. The feed is a source, not a to do list.
• Ignoring recency. Write to signals while they are fresh.
• Writing the same message to every signal. Each signal gives a different reason.
• Forgetting the rules. LinkedIn limits apply, and pacing protects your account. No tool can guarantee that LinkedIn will not restrict an account.

Can an assistant help?

If you use Claude or ChatGPT with Sellinger's MCP connector, there is a recipe that turns intent signals into leads. See turn intent signals into leads. It is the same process, driven from a chat.

What if the feed is too quiet or too loud?

A quiet feed usually means the signal is too narrow: a single profile with few posts, or a keyword nobody uses. Widen it by adding two or three related profiles or phrasings, or switch to a template in a broader category. A loud feed means the opposite: too many people, mostly the wrong ones. Add role filters, exclude competitors and agencies, and shorten the window you act on. Aim for a number of matching people per week that you can read and message with care, which for most small teams is tens, not thousands.

A 30 day checklist

• [ ] Week 1: three signals from different categories, noise excluded.
• [ ] Week 2: first 20 to 30 messages per signal, read in approval mode.
• [ ] Week 3: every reply read and sorted, signals compared.
• [ ] Week 4: keep, narrow or drop each signal.
• [ ] One new signal added, total still three.