Buying signals: the 12 types and the heat score
A buying signal is a sign that a person or company may be ready to buy, such as a job change, a funding round or a comment on a competitor's post. Ragworks AI SDR has 12 signal types, 10 ready to set up, and scores each person from 0 to 100 on how strong, fresh and varied their signals are.
By Kishore Kannan, Co-Founder & CEO. Last updated .
What is a buying signal?
A buying signal is something a person or company does that suggests they may be in the market: starting a new job, raising money, hiring in a department, or commenting on a competitor's post. Using signals like these to decide who to contact, and when, is often called signal-based selling.
In Ragworks AI SDR a signal is a standing instruction: watch something, and once a day report the people who fit. You choose what to watch and which job titles count. People you leave out, and people whose headline matches none of your titles, are dropped. The rest are scored against your description of a good fit, and those who clear the bar join a saved list with a work email, where one can be found.
Signals sit under Signals in the left menu of the outreach app. They run on their own prepaid balance, signal credits, kept apart from your plan's credits, and everything before a person qualifies is free.
What are the 12 types of buying signal?
Ragworks AI SDR groups its signals by where the data comes from: your own LinkedIn account, public posts about your market, company events, and your own tools. Ten of the twelve can be set up today. The names are the ones the app shows.
| Signal | What it picks up | Strength | In the app today |
|---|---|---|---|
| People engaging with your posts | Everyone who reacts or comments on your recent LinkedIn posts. | A reaction is light. A comment is medium. A comment that shows buying intent is strong. | Ready. Reads through your connected LinkedIn account. |
| New company-page followers | People who start following a company page you manage. | Light. | Ready. Reads through your connected LinkedIn account, which must be an admin of the page. |
| Job changes in your ICP | People in your saved Sales Navigator search who just changed jobs. | Medium. | Ready. Reads through your connected LinkedIn account, which needs a Sales Navigator seat. |
| Competitor engagement | People reacting or commenting on your competitors' posts. You choose which competitors, up to 10. | A reaction is light. A comment is medium. A comment that shows buying intent is strong. | Ready. No LinkedIn account needed. |
| Keyword posts | People who post, or engage with posts, about topics you care about. You choose the keywords, up to 10. | A post's author is medium. A reaction is light. A comment is medium, or strong when it shows buying intent. | Ready. No LinkedIn account needed. |
| Experts your buyers follow | People engaging with posts by experts you name, up to 20. | A reaction is light. A comment is medium. A comment that shows buying intent is strong. | Ready. No LinkedIn account needed. |
| Funding rounds | People at companies that recently raised money. | Medium. | Ready. No LinkedIn account needed. |
| Hiring by department | People at companies hiring in the departments you pick. | Medium. | Ready. No LinkedIn account needed. |
| New executive | People at companies that just brought on a senior hire. | Medium. | Not available yet. It is being verified. |
| Headcount growth | People at companies that are growing quickly. | Medium. | Ready. No LinkedIn account needed. |
| Leadership changes at public companies | People at public companies that just reported an executive or board change to the SEC. | Medium. | Ready. No LinkedIn account needed. |
| Events from your tools | People your own tools tell us about: a demo booked, a form filled, a trial started. | You set it: light, medium or strong. Medium if you do not. | Coming soon. |
Strength is how much a signal weighs in the heat score. A small model reads each comment and labels it. A comment counts as buying intent when it shows the writer is evaluating a product, already uses a competitor's, or has switched away from one.
How is the heat score worked out?
Heat says how warm one person is right now. It is a fixed formula in five steps, and it uses only what the signals recorded: their type, their strength and when they happened. It is not the fit score, which says how well a person matches the ideal customer you described.
- Each signal gets a weight from its strength. Strong is 1, medium is 0.6 and light is 0.3.
- The weight halves every 14 days. A signal is worth half after two weeks and a quarter after four. One that is more than 30 days old no longer counts.
- Repeats of one type add little. The best signal of a type counts in full, and each other signal of that type adds a quarter of its own weight. Together those extras never add more than the best one again. Ten reactions are not ten times one reaction.
- Different types add in full. Someone who comments on a competitor's post and then follows your company page is hotter than someone who does only one of the two.
- The total becomes the score. Heat is 100 × (1 − e^(−total)), rounded and never above 99. One fresh strong signal scores 63, and the curve flattens after that, so there is always room for one more signal.
| One signal | Today | After 7 days | After 14 days | After 28 days | 31 days or more |
|---|---|---|---|---|---|
| Strong | 63 | 51 | 39 | 22 | 0 |
| Medium | 45 | 35 | 26 | 14 | 0 |
| Light | 26 | 19 | 14 | 7 | 0 |
Each cell is the heat of a person who has only that one signal.
Ten reactions of one type score 45. One reaction scores 26.
Two strong signals of the same type score 71. Two strong signals of different types score 86.
An example: one head of sales, day by day
A head of sales at a company that fits your titles. Each row is their heat on that day, counting the signals recorded so far.
| Day | What happens | Heat | Label |
|---|---|---|---|
| 0 | They comment on a competitor's post, and the comment says they are evaluating options. A strong signal. | 63 | Warm |
| 2 | They follow your company page. A light signal, and a different type, so it adds in full. | 70 | Hot |
| 6 | Their company announces a funding round. A medium signal, and a third type. | 80 | Hot |
| 20 | Nothing new. Every signal has faded. | 55 | Warm |
| 30 | Nothing new. | 38 | Cool |
| 37 | Nothing new. Their newest signal is now 31 days old, so none of them counts. | 0 | None |
On day 2 the Why now line reads: Engaged with a competitor 2 days ago, and followed your company page today.
Where does the heat score show up?
- Hot right now. The Signals dashboard lists the people hit by more than one signal, hottest first.
- Hot accounts. The same idea by company. An account's heat combines the heat of its people, so a second warm person raises it, and it never reaches 100. Two people with heat 63 and 45 make an account with heat 80.
- Why now. The lead list and campaigns show a one-line reason built from the two heaviest signals, in fixed phrases with the age in days, never a name, a headline or a comment.
- Labels. A heat of 70 or more is labelled hot, 40 or more warm, and anything lower cool. A single signal never reaches hot: even a fresh strong one scores 63.
What does the heat score not tell you?
- Whether someone will buy. Heat measures how strong and how recent the signals are. It does not use replies, meetings or revenue, and it does not predict a purchase.
- Anything about people your rules left out. Only people who clear your leave-out list and your job-title check are recorded, so heat exists only for people who already fit your titles.
- Anything older than 30 days. The events behind the score expire after 30 days, so a lead that has gone quiet has no heat, however hot it once was.
- Signals you have not set up. Heat counts only the signals you have added. Someone who did something that none of your signals watches has none of that heat.
What does Ragworks keep behind the heat score?
The events behind the score hold an identifier for the person, the signal type, the strength and the time, and no text: no comment, headline or name from a post. They expire after 30 days.
A lead that qualifies keeps its evidence, such as the words of a comment, so you can see why it surfaced. Comment text is read, and shown to you, but it is never pasted into a message. A campaign can use the fact that someone is evaluating a competitor. It cannot quote them.
Common questions about buying signals
What is a buying signal?
A buying signal is an action or event that suggests a person or company may be ready to buy: starting a new job, a company raising money or hiring, or someone commenting on a competitor's post. Ragworks AI SDR has twelve types of signal, ten of them ready to set up today, and combines the signals it finds for each person into a heat score.
What is signal-based selling?
Signal-based selling means choosing who to contact, and when, from what people and companies are doing, instead of working down a cold list. In Ragworks AI SDR you set a signal up once. It runs once a day, drops the people you leave out and the people whose headline matches none of your job titles, scores the rest against your description of a good fit, and adds those who clear the bar to a saved list, with a work email where one can be found.
What buying signals does Ragworks AI SDR track?
It has twelve types: People engaging with your posts, New company-page followers, Job changes in your ICP, Competitor engagement, Keyword posts, Experts your buyers follow, Funding rounds, Hiring by department, New executive, Headcount growth, Leadership changes at public companies and Events from your tools. Ten can be set up today. New executive is not available yet, and Events from your tools is coming soon.
How is the heat score calculated?
Each signal weighs 1 if it is strong, 0.6 if medium and 0.3 if light. The weight halves every 14 days, and a signal more than 30 days old no longer counts. Within one type, the best signal counts in full and each other adds a quarter of its weight, together at most the best again. Different types add in full. The total becomes 100 × (1 − e^(−total)), rounded and never above 99.
How long does a buying signal count?
A signal loses half its weight every 14 days, and it counts for nothing once it is more than 30 days old. One strong signal on its own gives a heat of 63 today, 39 after 14 days and 22 after 28 days.
What is a good heat score?
A heat of 70 or more is labelled hot, 40 or more warm, and anything lower cool. A single signal never reaches hot: even a fresh strong one scores 63. It takes two or more signals, and different types add the most.
Does Ragworks keep what people write?
The events behind the heat score hold no text: only an identifier for the person, the signal type, the strength and the time, and they expire after 30 days. A lead that qualifies keeps its evidence, such as the words of a comment, so you can see why it surfaced. Comment text is read, and shown to you, but it is never pasted into a message.
Sources
The Ragworks docs describe how signals are set up and read.
Ready to put your pipeline on autopilot?
Talk to our team. We'll help you pick the right plan and build your outreach playbook.
Book a Demo