How to Find Buying Signals: A Manual Playbook (No Tools Required)

You don't need a subscription to find buying signals. Across 279 classified signals in AvairAI campaigns, about 72% came from two free sources, company newsrooms and trade press, and another 15% came from job boards. Investor filings, the source most teams build alerts on first, produced about 1%. This is a playbook for working those sources by hand: what to monitor, what to read once you're there, and how to verify a signal before a rep spends time on it.

Where signals actually come from

Here's the measured distribution across our classified set.

The ratio is close to the inverse of where most sales teams point their attention.

Filings feel authoritative, so people wire up alerts on them first. But filings are quarterly, lagging and only exist for public companies, which for most B2B sellers working the mid-market is a small slice of the addressable list. Meanwhile a regional trade publication will report a warehouse acquisition or a systems migration months before it shows up in anything you can subscribe to.

If you want the framework behind which of these events matter for your product, the guide to B2B buying signals covers the derivation. What follows is the mechanical part.

Trade press: the underrated first stop

Trade press is 32.3% of what we find and it's where I'd start if I were building a watchlist from scratch today.

Every industry has two or three publications that cover it obsessively and rank well for company names. Logistics has them. Higher-ed IT has them. Offshore energy has them. They report the operational detail that mainstream business press ignores: a TMS migration, a new distribution facility, a practice launch, an equipment order.

How to work it: identify the two or three publications that cover your buyers' industry, not yours. Subscribe to their newsletters with a dedicated inbox folder. Set up alerts for the operational vocabulary your buyers use when the pain you solve is active, not your product category. A vendor selling integration software should be watching for "migration," "ERP implementation," "new distribution center," "EDI," not "integration platform."

That distinction is the whole trick. Search for the symptom in the buyer's language, not the cure in yours.

Company newsrooms: highest volume, easiest to automate

At 39.4%, company press releases are the single largest source. They're also the most straightforward: the company publishes, on its own site, exactly what it wants known.

The limitation is that press releases are written to sound good, so you're reading past the adjectives for the operational fact underneath. "Strategic expansion into the Southwest" means an office with a target and no customers. "Investment in digital transformation" means a project with a budget and a deadline.

How to work it: for a defined account-based target list, the newsroom or press page of each company, checked monthly, is more reliable than any aggregator. For discovery beyond your list, alerts on the phrases that describe the change, things like "opens new office," "appoints," "expands facility" and "launches practice," combined with an industry qualifier.

Job boards: free, daily and barely used for this

15.1% of our signals come from job postings, and I'd argue this is the most under-exploited free source in B2B sales. The supply is structural: with median job tenure at 3.9 years per the Bureau of Labor Statistics, the postings never stop coming.

A job description is a company documenting its own operational pain in public, in detail, with the budget already approved and a hiring manager who owns the outcome. Nothing else you can read for free is that explicit.

What to read: skip the title and the boilerplate. The requirements and responsibilities sections are where the current state leaks out. "Experience with manual reconciliation across multiple systems" tells you what's broken. "Will own the migration from [legacy system]" tells you the project, the timeline and the incumbent you'd be replacing.

What the volume tells you: three postings for similar roles in one quarter is a different signal from one. It means the team is scaling against a problem faster than it can hire, which is the moment software becomes an easier sell than headcount.

How to work it: the company's own careers page for named accounts, and the big boards filtered by role keyword plus geography for discovery. Free, updated daily, no vendor required.

Government and procurement portals: low volume, near-certainty

Government portals and procurement notices together are about 3% of what we find, and they're among the most decisive signals that exist.

An RFP is a company or institution announcing, publicly and with a deadline, that it intends to buy something in a defined category. There is no inference involved.

Regulatory registers are the other half of this. Safety regulators, environmental agencies and financial supervisors publish enforcement actions, meaning fines, notices and findings, with dates and case numbers. For anyone selling compliance, safety or risk software, an enforcement notice is a public record that the exact pain you solve just cost the company money and probably a headline.

How to work it: these are jurisdiction-specific and worth an hour of setup. In the US, SAM.gov is the federal portal where contract opportunities are published, and most states and large agencies run an equivalent. Find the register that covers your buyers' industry and geography, and check it monthly. Low volume means it stays manageable by hand forever.

Public review sites: the pain described by your buyer's customers

Only 2.5% of our signals come from public review data, and it's the category I find most interesting, because it inverts where you look.

Every other source watches the buyer. Review data watches the buyer's customers, and their customers describe the pain without any of the polish a company applies to its own communications.

The clearest example from our campaigns: a hotel where guests kept posting reviews saying the air conditioning didn't work. For an HVAC services company that's a live, specific, evidenced problem, described in public, by the people experiencing it. The hotel never researched HVAC vendors. No behavioral data would ever surface it.

This works anywhere the pain you solve eventually reaches an end customer: hospitality, healthcare providers, logistics, consumer services, field service, retail operations.

How to work it: review platforms, app-store reviews and industry-specific complaint boards, searched for the symptom vocabulary rather than the company name. Look for repetition. One complaint is a bad day, six across three months is a system.

Verify before anyone spends time on it

Two minutes here saves a bad call, and the cost of skipping it compounds: Harvard Business Review pegged bad data at more than $3 trillion a year in the US.

Check the entity. Company names collide constantly, and mid-market firms share names with unrelated businesses more often than you'd expect. Match on domain, not name.

Check the date. Aggregators re-publish old news with fresh timestamps. Find the original.

Check it's still true. Announcements get reversed. Offices close. The executive who was appointed in March may have left in June. B2B data goes stale at a documented pace, more than 20% a year by HubSpot's database-decay research, and the facts inside a months-old announcement decay the same way.

Check the pain link. This is the one people skip. Ask out loud: what does this event mean for the specific problem we solve? If the honest answer is "it shows they're growing," that's not a signal. That's a company doing well, which is not a reason to call.

The routine, and where it stops working

Set up properly, this is about 30 minutes a day: fifteen minutes on inbound alerts and newsletters, ten on job boards for your priority accounts, five on registers and review sources on a weekly rotation.

At around 40 accounts that's genuinely sustainable, and it will outperform a purchased list by a wide margin, because everything you find is verified, sourced and connected to a pain you can articulate.

Past 40 it stops scaling, not because the work gets harder, but because it never compounds. Every account you add is monitoring load forever, and the first thing that slips when you get busy closing is the monitoring. Which is exactly when your pipeline needs it most. That's the trade-off we built Pain-Signal Targeting to remove: derive the pain from your website, work out which events reveal it, then search continuously so nobody has to choose between prospecting and selling.

Either way, the sequence is the same, and knowing which event families deserve your attention is what makes the 30 minutes pay.