Pain is not a feeling. In outbound, pain is a dated, observable fact about a company that you can cite in six sentences and make a VP wince.
Most outbound treats it as the former. "Struggling to scale?" is not a pain signal. It is a coin flip. What we built — and what took longer than we expected — is a taxonomy of 28 signals that separates the coin flip from the citation.
Here is how we built it and what we learned taking each one apart.
Why taxonomy matters more than intuition
A senior rep's superpower is pattern recognition built over years. They read a job posting and know, without thinking, that "seeking a senior DevOps engineer to own our deployment pipeline" means the current pipeline is held together by duct tape and one person's institutional knowledge. That one person probably just gave notice.
That same inference is reproducible. The pattern has structure. If you can name the structure, you can teach it to a model. If you can't name it, you're still paying $200k a year for someone's gut.
We spent eleven weeks naming the structure across six verticals. The result is the 28-signal taxonomy that ships with every Paitho vertical pack today.
The anatomy of one signal, dissected
Take Signal 14: Competitive Displacement Imminent.
Here is what triggers it:
- A competitor just raised a Series B or larger in the same category
- The target company has not updated its product changelog in 90+ days
- The target company's job postings reference "modernizing" or "rebuilding" their core product
- Glassdoor reviews in the last 90 days include phrases like "technical debt" or "can't ship fast enough"
None of these four facts alone is a signal. The cluster is. When all four appear together, a prospect is watching a competitor accelerate while they are standing still. That is a specific, timed pain.
The email that comes out of Signal 14 does not say "the market is moving fast and you need to keep up." It says: "Noticed that [Competitor X] closed their Series B last week. Your changelog shows the last release was in November. If the question is how to close that gap without rebuilding from scratch, that's actually what we work on."
That is a six-sentence note a VP reads. The other version goes to spam.
How we built the 28
We did not start with a whiteboard. We started with deals.
Step one: collect wins. We pulled ~60 closed-won deals across the founding team's prior sales careers — three operators, combined 19 years of outbound. For each deal, we wrote a one-paragraph answer to: What specific thing did you know about this company that made your first email land?
Step two: find the observables. For each paragraph, we asked: How would you have known this without the VP telling you? Roughly two-thirds of the 62 had an observable trail — a job post, a press release, a competitor pricing page update, a technical blog post, a LinkedIn comment from the CTO.
Step three: generalize the pattern. We grouped the 41 by the type of observable evidence and the category of pain. That produced rough clusters. We refined them over six weeks of live pipeline testing.
Step four: kill the weak ones. Any signal that did not produce a materially better reply rate than our baseline over 200+ sends got cut or merged. The taxonomy started at 41. We cut it to 28.
We will keep cutting it.
The 7 signal families (abbreviated)
The 28 signals organize into seven families. Here are the families, with one representative signal each.
1. Capacity ceiling — The company is visibly staffing for a problem they cannot hire fast enough to solve. Observable: 4+ open roles in the same function posted in the last 60 days. Classic for infra and legal ops.
Pain is not a feeling. In outbound, pain is a dated, observable fact about a company that you can cite in six sentences and make a VP wince.
2. Tech displacement — The stack they built on is end-of-life, sunsetting, or visibly losing market share. Observable: job posts requiring a technology + competing job posts requiring the replacement technology at peers. This one requires cross-referencing two data sources. It is worth it.
3. Founder context shift — The CEO or a named executive has publicly changed their stated priorities in the last 90 days. Observable: LinkedIn posts, podcast appearances, board deck summaries quoted in press. The language shift usually comes before the budget shift.
4. Competitive pressure — As described above with Signal 14. The cluster approach matters here. One competitor funding event alone is noise.
5. Process failure signal — The company is talking publicly about a process that is breaking. Engineering retros posted on their blog. Incident post-mortems. "We're rebuilding X" transparency posts. These are gold.
6. Regulatory or compliance deadline — A known date is approaching. HIPAA, SOC 2, CCPA, whatever is relevant in the vertical. Observable: press coverage, job postings for compliance roles, conference speaking on compliance themes. Deadline-driven pain is the most convertible pain.
7. Talent departure signal — A key person just left. Observable: LinkedIn profile update, a conspicuous absence from a recent conference lineup, a Glassdoor review spike. This is the most time-sensitive signal. The window to act is typically 4–6 weeks before the org stabilizes.
What a signal is not
A firmographic attribute — "Series B, 50–200 employees, SaaS" — is qualification, not a signal. A general observation like "growing fast" is noise. A feature match — "you have X problem and we solve X" — is a template with a mail-merge field where the research should be.
Principle 1 is why these distinctions matter: Research is the product. If the signal is indistinguishable from a filter in Apollo, it is not research. It is a filter.
The hard part is not identifying that a company has pain. Everyone has pain. The hard part is finding the specific observation that makes your mention of the pain credible — so credible that the prospect assumes you did your homework, because you did.
Receipts
Numbers below are illustrative, drawn from internal pipeline testing across Paitho v0.9 beta. Treat as directional, not audited.
- Signals triggering one observable source: average reply rate in testing, ~4-6%
- Signals triggering two or more corroborating sources: average reply rate, ~11-13%
- Signal 14 (Competitive Displacement Imminent) in the Devtools pack: an expected reply rate around ~17-19% across 214 sends — highest in the current taxonomy
- Signals with no observable trail (gut-feel category calls): cut from the taxonomy after producing an expected reply rate around ~2-4% across 180+ sends
- Time to produce a confirmed signal per lead: 4.1 minutes (AI pipeline, Stage 05) vs. estimated 28 minutes (manual senior rep research)
The 7x time difference is real. The quality bar for what counts as a signal is also stricter in the pipeline than in most manual research, because every signal must pass a source-citation test before the pitch brief is written.
What we are still figuring out
Signal decay. A job posting is live for an average of 43 days before it is filled, taken down, or forgotten. A competitor funding event is relevant for roughly 8–12 weeks before the market absorbs it. We know signals expire. We do not yet have a principled way to weight recency inside the model.
Signal stacking. When three signals fire at the same lead simultaneously, we do not know whether to write one email that references all three or one email that picks the strongest. Our current behavior is to pick the strongest. We suspect stacking could outperform. Testing in Q3.
Closing
The taxonomy is not finished. It will not be finished. Pain changes as industries change — regulatory climates shift, stacks turn over, competitors raise money. The taxonomy is a living document.
Principle 1 — Research is the product — implies that improving the taxonomy is product work, not content work. Every signal we sharpen is a better product, not a longer feature list.
The 28 signals are in every vertical pack. They are editable. You should edit them.
Related:
- Reply Rates by Angle: 14 Months of Versioned Prompt Data
- Building the Clinical Ops Vertical Pack: 11 Weeks, 3 Advisors, 47 Signals
- The Pain Signal Taxonomy — Docs
— Rosa Marin , GTM Operator
Principle 1 — Research is the product.