The signals worth acting on name one company and point to a new budget for pipeline: a sales or growth hire, a hiring surge, the company's own fundraise, or a large rise in total value locked (TVL). Of 6,069 signals we logged between July and October 2026, 96.5% were too imprecise to email about (our data, October 2026).
How many buying signals are worth acting on?
Very few. Our catalyst monitor logged 6,069 signals between 26 July and 7 October 2026. It watches a fixed list of target companies across their X accounts, public job boards, crypto news feeds and DefiLlama TVL data (our data, October 2026).
| Signal type | Count | Share of total |
|---|---|---|
| X posts | 5,170 | 85% |
| TVL moves | 566 | 9% |
| Go-to-market job posts | 180 | 3% |
| Hiring surges | 79 | 1% |
| New DeFi listings | 40 | 1% |
| Fundraise headlines | 34 | 1% |
| Total | 6,069 | 100% |
Of the 6,069, 5,857 (96.5%) were too imprecise to email about. That left 212. Then 91 of those 212 named a company that our contact database could not match to exactly one firm of 11 to 1,000 staff (our data, October 2026). At most 121 signals, about 2% of the total, could reach a named buyer at the right company.
We started emailing on these signals in October 2026, so we do not yet have reply rates to report. These counts measure precision: whether a signal names the right company and a real reason to write.
Which signals should you keep?
Keep the events that tell you a company has a new pipeline target and no pipeline yet. A new sales hire is the clearest case. The average SDR takes 3.0 months to ramp (The Bridge Group, February 2025). A company that posts a sales role today has a pipeline target and a seat that will not book meetings for a quarter.
These are the filters we run today (our rules, October 2026):
| Signal | Keep when | Drop when |
|---|---|---|
| Job post | The role is in sales, business development, partnerships, marketing or growth | The role is in engineering, design, product, support, operations or sales enablement |
| Hiring surge | Open roles rise by 5 or more | The rise is smaller, or the new roles are all technical |
| Fundraise | The company is the subject of the headline: "Company raises" or "Company closes" | The company appears as an investor, a partner or a passing mention |
| TVL move | TVL rises 25% or more and ends at $5M or more on DefiLlama | TVL falls, or the protocol is too small to fund outbound |
| X post | Never as a trigger; read it as context for another signal | Always, as a trigger |
| New listing | Never as a trigger; a listing says nothing about a sales budget | Always, as a trigger |
A TVL rise matters for DeFi protocols because more deposits usually bring more integration and allocator conversations. A TVL drop points the other way, so we drop it.
Why do so many signals fail?
Most signals describe activity, and activity rarely names a buyer. The 5,857 signals we dropped fell into four groups (our data, October 2026):
- Noise. Posts that mention a company with no event behind them, such as a meme, a reply or a price comment.
- The wrong company in the headline. A raise headline names the investor first, or a partner announcement names both firms.
- TVL drops. Falling deposits rarely come with a new budget for outbound.
- Roles outside sales and marketing. A company hiring five engineers is building product, and the post says nothing about a sales budget.
X posts are 85% of the volume, and we drop all of them as triggers. If you track only one source, track job posts. They name the company, the role and the date in one place.
How do you match a signal to the right company?
Match on the exact company name across all company sizes, and apply your size filter after the match. Namesake errors are the main data risk in signal-based outbound (our data, October 2026).
We caught two. A "Ramp" hiring signal (ramp.com) matched a different company called Ramp Network. An "Aura" signal (aura.com) matched AuraSOS. Both errors came from searching inside a staff-size band first. When the real company falls outside the band, the search returns the closest namesake inside it.
Use this procedure for every signal:
- Search your contact database for the exact company name, with no size filter.
- Take the top-ranked exact-name match.
- Check that its domain matches the domain in the signal, such as the careers page URL or the website in the press release.
- Apply your size filter to that one company. If it falls outside the range, skip the signal.
- If no exact match exists, skip the signal and log it. Do not take the nearest name.
A wrong match sends a confident email about a raise or a hire to a firm that had neither, so a skipped signal is the cheaper error.
How do you pick the right person at the matched company?
Rank buyer titles, cap the number of contacts, and check that each title belongs to the matched company. Our order is chief executive or founder first, then revenue leadership, then heads of growth, business development, partnerships or marketing, then the chief operating officer. We contact at most 2 people per company (our rules, October 2026).
Titles in contact databases can belong to another company. We found a co-founder title naming one firm on a different firm's record (our data, October 2026). We now drop any title that names a different company. We also drop investors, board members, advisors, recruiters and technical roles, who rarely own a pipeline budget.
Verify every address in the week before you send. We cover the cadence in how fast verified B2B email lists decay.
How fast should you act on a signal?
Quickly, and name the month once the signal is older than 14 days. We write "is hiring" for a job post seen in the last 14 days, and "was hiring in September" for an older one (our rules, October 2026). Both sentences stay true when the buyer checks them.
Once you contact a company, leave it alone for 90 days unless the buyer replies (our rules, October 2026). A second signal from the same firm a week later is a reason to follow up in the same thread, and a new cold email reads as a mistake.
Build every fact in the opening line from the signal itself: the role, the round, the TVL figure and the month. A model that writes the line from memory can add facts the signal never contained.
For the system that runs these filters end to end, read what GTM engineering is. To have signal filters and matching rules built for your market, start with The Diagnostic for $2,500.