Buying signals worth tracking, and the ones that waste a week

"Signal-based selling" has become a category with a price tag, and most of what gets sold under it predicts nothing. A signal earns its place when it passes four tests.

Public. You can verify it from outside without asking anyone. Dated. You know when it happened, so you know when it goes stale. Specific. It names a constraint rather than a mood. Connected. A straight line runs from the constraint to what you sell.

Drop any one of the four and you are guessing with extra steps. Here is how the common signals score.

Signals that carry weight

Repeated hiring for the same role

The strongest one in staffing, logistics and industrial services. A company that has advertised the same role four times since spring has a problem it has failed to solve, and the fourth attempt is a different conversation from the first.

It passes all four tests cleanly. The counts are public, the dates are on the postings, the role names the gap, and for an agency filling that role the connection is direct. Ripe Leads built a whole outbound method on it and publishes the underlying counts, including 11,761 open CE driver roles across 7,058 German companies in one monthly snapshot.

Weakness: it expires in weeks, so it demands a repeated crawl. More on that in what a job ad tells you.

Hiring bursts and multi-site spread

The same role appearing across twenty locations at once, or a dozen roles posted in a fortnight after a quiet quarter. Volume of this kind means a contract, a site opening or a policy change, and somebody inside is under real pressure.

Stronger than a steady drip of postings because it dates the change. You know roughly when it started, which tells you how long the window stays open.

Public tender awards and contract wins

A company that just won a three-year contract has committed to delivering something it cannot yet staff, supply or support. The award is published, dated and specific about scale.

Underused because reading tender portals is dull. That is also why the inbox is empty when you get there.

A new person in the seat that owns your problem

A new head of marketing rewrites the plan. A new operations director reopens supplier decisions the last one settled. The window runs roughly from month two to month six: too early and they are still learning, too late and they have chosen.

Weak on its own, strong when it lands on top of another signal. A new marketing lead at a company already publishing daily is a different prospect from a new marketing lead at a company that posts twice a year.

Sustained public behaviour

The one Social Intel reads. A company publishing on a schedule, in two languages, with a house voice rather than an agency template, is a company investing in the channel. That is a durable signal rather than a dated event, which makes it useful for defining the segment rather than for timing the send.

Our bias here is obvious. Weigh it accordingly.

Signals that mostly waste a week

Funding announcements

Sold hardest, delivers least. Every vendor in the market sees the same TechCrunch line on the same morning, so the prospect gets forty messages that week that all open the same way. The money is usually allocated before the announcement, and the announcement often trails the wire by months.

Useful only as a second filter. "Raised recently" plus "hiring twelve engineers across two cities" is a real signal, and the second half is doing the work.

Anonymous website visits

An unnamed company visited your pricing page. Perhaps a buyer. Perhaps a competitor, a candidate, or a bot that resolves to their ISP.

Reverse IP resolution guesses the company and cannot guess the person. Acting on it produces the awkward email that implies you were watching, sent to somebody who never visited. Keep it as an internal prioritisation nudge and never as an opening line.

Generic technographics

"Uses HubSpot" describes ten thousand companies and predicts nothing about this quarter. It becomes a signal only when the specific tool implies the specific gap you fill, and then you are really selling to the gap.

Anniversaries, awards and birthdays

A company turning ten is not a company with a problem. These get used because they are easy to automate, and they read as automated.

Generic engagement

Somebody liked a post. Under most feed algorithms that means the post appeared while they were scrolling. Treating it as intent produces the message everyone recognises, and the recognition costs you the reply.

Stack them

The practical rule: two mediocre signals stacked beat one strong signal used alone.

One signal alone puts you in a crowd. Everyone monitoring funding rounds emails the same companies in the same week. Everyone watching job boards for "warehouse operative" reaches the same twelve employers.

Stacked signals put you somewhere quieter. "Advertised the same role three times since June" plus "opened a second site this year" leaves a list nobody else is sending to, and the opening line writes itself from the intersection.

There is a floor, as always. Four filters stacked leaves eleven companies, which is a research project rather than a campaign. Two or three, then widen the weakest one when the list gets too thin.

Verify before you send

The failure that undoes all of it: a signal that was true last quarter, stated confidently in the first line.

"I saw you are hiring 23 drivers" sent to a company that filled them in July tells the reader that your research is automated and unchecked, and that everything else in the message is probably the same. One wrong detail costs more than the whole signal was worth.

Recrawl before the send, drop what you cannot confirm, and accept the shorter list. Precision is the entire value of working this way, and it survives exactly as long as the facts do.