Two people send outreach the same week, to similar companies, with similar offers. One email reads like it was written by someone who did homework. The other reads like a template wearing a costume. The difference usually isn't effort, word count, or even quality of writing — it's whether the sender actually found a signal, or just a fact, and mistook one for the other.
Most "personalization" advice says "find something about them." That's the root of the problem. There are three different things you can find about a prospect, and only one of them is a reason to write today.
A fact, a signal, and a reason are three different things
A fact is verifiable and true, but says nothing about timing. "They're a 12-person company." "They sell to dental clinics." "Their HQ is in Rotterdam." Facts are necessary — you have to know who you're talking to — but a fact alone produces emails like "I saw you're a 12-person company," which is the corporate equivalent of "nice weather we're having."
A signal is a recent, observable event: a new hire in a role that touches your problem, a changelog entry revealing a workflow you improve, a pricing page that changed, a public complaint in their own community. Signals have a date and a direction.
A reason is the connective sentence between them: because this signal happened, the problem you solve just became slightly more urgent for this company than it was last quarter. The signal is the news; the reason is why the news is theirs.
Fictional example (illustration, not a real success story): a founder selling docs tooling notices a company's help center. Fact: they have a help center. Signal: last month's job posting says "first technical writer." Reason: a team about to hire its first writer is about to formalize docs workflows the same month. The email writes itself from that one sentence — and no amount of "loved your recent post" could have produced it.
Outreach that stacks all three lands differently because it proves a human looked. Outreach that stops at facts proves only that a scraper worked.
The five tests every claimed signal must pass
Before a "signal" earns a place in your first line, run it through five quick tests:
- Recency. Can you date it to the last few months? "They raised a Series A in 2023" is history, not news. If you can't find when it happened, you don't have a signal — you have trivia.
- Ownership. Is this the recipient's problem, or the company's? A Head of Support owns a support-tooling signal; the CEO of a 12-person company owns nearly everything, which is both an advantage and a reason their inbox is a war zone. If the change clearly belongs to someone else's function, write to that person — or don't write.
- Relevance. Does this specific change touch the specific problem you solve? "They launched a new product" is only a signal if launching products creates the workflow you improve. A genuine but unrelated signal is small talk with a timestamp.
- Specificity. Could your first line quote it back in one sentence? "I saw your changelog" fails; "your September changelog says the migration tool moved out of beta" passes. If you can't phrase it concretely, you haven't read it closely enough.
- Verifiability. Can you name where you saw it, today, if they reply "where did you see that?" No source, no email. This one rule quietly eliminates most invented "signals."
Four or five passes: strong signal — build the email around it. Two or three passes: usable as a quiet opener, not as a headline claim. One pass: context for your notes, not for your inbox.
Noise that dresses up as signal
The most common failures are things that feel researched:
- Awards and press mentions. Usually months old, and relevant to the marketing team at best — and they've heard the congrats already.
- Podcast guest spots without a dated connection to your problem. If their episode said something that is your problem, that's a real signal. If the episode was about their vacation rental hobby, it's not.
- Hiring in general. "I saw you're hiring!" is not a signal. "You posted a first-ever lifecycle marketing role eleven days ago" is.
- Funding with no strategic angle. A raise only matters if it changes something your offer touches — headcount plans, a new market, a tooling budget.
- Anything outside work. Where they vacationed, what their kids do, posts from years ago. The line is simple: use what they'd happily tell a work acquaintance; never use what they'd be startled a stranger knows.
- The recycled "signal." If your last five first lines could each swap into the next email unchanged, you don't have signals — you have one sentence with different names typed in.
And then there are disqualifiers — observations that should end the research, not start the email: the wrong buyer role entirely, no plausible budget path, or any indication you'd be the third person this month to pitch the same obvious angle. Skipping a bad-fit prospect costs you one email; sending anyway costs you the address.
Signals have a shelf life
Because a signal is a dated event, it decays. A trigger that was sharp three months ago is now backstory — useful for understanding the company, useless as a "noticed this week" opener. Two practices keep your work honest:
- Write the signal down with its date and source the day you find it. A one-line ledger per prospect — signal · date seen · where · why it matters to them — beats a folder of browser tabs you'll no longer trust next month.
- Never reuse a stale observation in a live email. If the only signal you have is old, say so quietly and honestly ("I've been following your docs since the spring migration") rather than repackaging it as fresh — or fall back to a non-time-based observation you can still verify today. What you must never do is invent urgency that the record doesn't show.
The payoff, either way
Run this filter and one of two things happens consistently. Either the email gets built on a reason a busy stranger can verify in four seconds — which is the entire competitive advantage available to a sender with no brand — or you skip the prospect and save an hour. Both outcomes beat what most outreach does: turning facts into noise at scale, at speed.
AI disclosure: this post was drafted with AI assistance and reviewed by me before publishing.
Top comments (0)