Buying signals versus purchase intent: what public evidence can tell you

Most articles on buying signals list fifteen to forty of them, often with reply or win rates attached. Those lists tend to merge two separate questions. Whether an event happened can usually be checked. Whether it means a company will buy from you is a guess, however well informed. Keeping the two apart is most of the skill.

What is the difference between a buying signal and purchase intent?

A buying signal is observable behaviour: the company paid for something, published something, hired someone, raised money. Purchase intent is a state of mind inside the company: someone with a budget wants to solve a problem your product solves, and plans to act on it. Signals are evidence you can collect. Intent is the conclusion you are trying to reach, and outsiders never see it directly.

The table below separates, for common signals, what you can check from what you are inferring.

Question one: is the evidence real?

Evidence comes in grades. From strongest to weakest:

• A receipt. A published amount or an explicit paid label shows that money changed hands. You can open the page and see it.
• Something installed or published. A tracking tag, a pricing page, an affiliate program. It exists, but it shows setup rather than current activity.
• An inference from a pattern. A pinned listing, a badge, a burst of hiring. Reasonable to read into, easy to misread.
• A score from someone else's data. Third-party intent data is built from browsing behaviour you usually cannot inspect. You are trusting the vendor's method, not checking a fact.

Whatever the grade, keep the source and the date. A signal you cannot point to is hard to use in a first message and impossible to audit later.

Question two: what does it predict?

Even a perfect receipt answers a narrower question than it seems to. Signals tend to fall into three kinds:

• General spending behaviour. Paying for placements, ads or affiliates shows a company will spend money to grow. It is the most common kind of evidence and the least specific to your product.
• Capacity. A high published price or an enterprise tier suggests a company sells larger deals and may be able to afford more. It does not show a budget for your category.
• Timing. A recent payment or a new hire suggests activity now. Timing evidence decays quickly, and the date you first saw something is not the date it happened.

None of these is category need. The step from “this company spends on growth” to “this company needs what I sell” is always yours to make, by reading its product, audience and pricing.

Why a score is not a probability

Scores combine signals so you can sort a long list. They are useful for deciding where to look first, and easy to overread. It helps to know how one is built. BuyerCue's buying-propensity score (model 1.7.0) adds up four weighted components:

• Distribution spend (40 points): directory placements that took money or real effort, weighted towards how many, how recent, and whether a payment was confirmed. The act of paying counts for more than the amount.
• Buying capacity (32 points): mostly the highest published price, plus signs like an enterprise tier.
• Marketing activity (20 points): affiliate or referral programs, ad tags, social presence, content and newsletters.
• Paid tools detected (8 points): a small component, because few of these businesses run several paid marketing tools.

Two design choices matter when you read the result. First, a component with no evidence scores zero and is flagged, rather than being filled in with an average. Second, the score has not yet been checked against outcomes: there is no reply, meeting or purchase data behind it. It ranks how much evidence a company shows, not how likely it is to buy. The cut-offs between bands have been redrawn more than once as the data grew, which is normal for a young model and a reason not to treat any band as a verdict.

The public sample profile shows this in practice. It scores 76.6, in the high band, because of a confirmed $3,555 placement, a published top tier of $499 a month and an enterprise plan. Those are real facts about the company. None of them says it wants to buy anything from you.

What does the evidence look like across 2,354 companies?

On September 24, 2026, BuyerCue's searchable index held 2,354 software businesses. Here is how they split across the bands.

The bottom band is called unclear, not low, for a reason. 87.5% of those companies are missing at least one component, against none in the high band, and for 84.1% of them the missing piece is buying capacity, usually because there is no public pricing. Every one of them has listing evidence, since that is how a company enters the index. “Unclear” mostly means “paid for a placement, and little else could be established”, not “unlikely to buy”.

It is also worth seeing which signals could be established at all from public pages.

Funding stage and headcount, two of the signals most guides lead with, were not established for a single company here, and hiring for only 13. For the businesses in this index, the usable evidence is what they pay for and what they publish: placements, prices, plans and programs. If your prospecting playbook depends on funding news, it will find little to work with among companies like these.

Method: counts cover BuyerCue's searchable companies on September 24, 2026, excluding suppressed, quarantined and unscored records. “Not established” is not the same as absent: a company with no detected ad tag may still advertise. The index is built mainly from one paid launch board, so it reflects that board's audience.

What can't any public signal tell you?

• Whether they need your category now. Only their product, customers and circumstances can suggest that, and only they know for sure.
• Who decides and with what budget. Public pages rarely show either, especially at small companies.
• Whether they already bought a competitor. A new tool may not show up on their site for months, if at all.
• Whether they want to hear from you. A public business inbox is a route, not an invitation. Follow the laws that apply to you and honour every opt-out.
• Your conversion rate. Be cautious with published reply or win rates for particular signals. Unless you can see how they were measured, treat them as marketing, and measure your own.

How should you use signals, then?

1. Choose signals that connect to your offer. A company that pays for placements is a natural prospect for distribution, conversion and analytics tools; less so for, say, payroll.
2. Verify before you rely on them. Open the source, note the date, and confirm the evidence belongs to the company you think it does.
3. Research fit separately. Read the product, audience and pricing. The signal earns a company a look; fit decides the rest.
4. Refer to evidence, not assumed intent. “I saw you launched on X” is checkable. “I know you're looking for Y” is a guess the reader will notice.
5. Keep your own results. Track which signals led to replies and customers for your product. Over time that is worth more than any general score.

For a step-by-step method of finding the payment evidence itself, see how to find SaaS companies paying for distribution.

Browse confirmed paid placements How BuyerCue scores evidence