tl;dr Token personalization ({FirstName}, {Company}) is free now, which means it signals nothing. Real personalization is responding to what the person just did, and "the person" is really three audiences (leads, prospects, trial users) doing three different things. The scarce input is not data or copy. It is the moment, and moments expire in minutes. This is part 1 of 3 on personalization: the why. Part 2 is the mechanics, part 3 is the measurement.
Two emails reached me last month. Both would pass a marketing review as personalized.
The first knew my name, my company, and my title. The merge fields rendered clean. Nothing was misspelled. It was, by the standards of most outbound tooling, a success.
The second knew what I did. It referenced the pricing page a teammate and I had compared the morning before, and it answered the exact question that page had left open.
I deleted the first without finishing it. I replied to the second in four minutes.
That gap is what this note is about, and the two after it.
TOKEN VERSUS MOMENT (interactive: toggle between the two emails from the cold open) THE TOKEN EMAIL (every marked phrase came from the database): "Hi {FirstName}, I noticed {Company} is doing big things in {Industry}. As a {Title}, you probably care about growth. Do you have 15 minutes this week?" Deleted unread. Every personal word was a merge field. THE MOMENT EMAIL (every marked phrase came from behavior): "You and a teammate compared the Growth and Scale plans yesterday morning. The cap question usually decides it. Here is the honest answer, and one thing the pricing page omits." Replied in four minutes. Every personal word was an observation. Grey phrases came from the database. Teal phrases came from behavior.For twenty years, personalization mostly meant tokens. Mail merge got cheaper, the databases got bigger, and the template industry taught everyone the same trick: sprinkle {FirstName} and {Company} through a generic pitch and call it one-to-one.
Here is what took me too long to admit: I ran teams that shipped that outreach and called it personalization. The recipient's spam filter had it categorized before I did.
Then language models made the trick free. Any tool can now write a plausible bespoke-sounding paragraph about your company, your funding round, your podcast appearance. The copy is infinitely cheap.
And when something becomes free, it stops being a signal. A polished, personal-sounding email no longer tells the buyer you did the work. It tells them you have a subscription.
A token proves you have a database. A moment proves you were paying attention.
Buyers act on the difference even when they never articulate it. McKinsey's personalization research found 71% of customers expect personalized interactions and 76% get frustrated when they don't happen. The same research found companies that excel at personalization take 40% more revenue from those activities than average players. The expectation is set. The token doesn't meet it.
Part of why personalization fails is that the word collapses three different jobs into one.
A lead is a flicker. An anonymous or barely-known account read your pricing page, opened the docs, showed up at the edge of your funnel. It hasn't asked you anything. Personalization here means proving you noticed the flicker without being creepy about it: answer the question the behavior implies, not the person's biography.
A prospect is a conversation. There's a named human, a thread, maybe a call. Personalization here means continuity: remembering what was compared, what was objected to, what was promised. The unforgivable email to a prospect is the one that ignores the last thing they told you.
A trial user is telling you everything. They invited a teammate on day two, hit a usage cap on day four, went quiet on day five. Personalization here isn't email craft at all. It is responding to the product behavior: the nudge that lands when the cap is hit, not in Thursday's batch.
Same word, three different kinds of attention. A team that "does personalization" with one template strategy is doing none of them.
THREE AUDIENCES, ONE WORD (interactive: pick lead, prospect, or trial user) LEAD. What it emits: a flicker: pricing visit, docs read, edge of the funnel. What it is asking: "Is this worth my time?" A personal first touch: answer the question the behavior implies, in minutes, without the biography. PROSPECT. What it emits: a conversation: a thread, a call, an objection on record. What it is asking: "Were you listening?" A personal first touch: continuity with the last thing they told you. Never ignore it. TRIAL USER. What it emits: product behavior: teammate invited day two, cap hit day four, silence day five. What it is asking: "Does this work for us?" A personal first touch: the nudge that lands when the cap is hit, not in Thursday's batch.The oldest numbers in sales operations are still the most ignored, so let me name them by their parts.
The 2007 MIT Sloan / InsideSales.com lead response study (Dr. James Oldroyd) found that contacting a lead within 5 minutes instead of 30 makes you about 100x more likely to connect and 21x more likely to qualify. You have seen those numbers attributed to Harvard. They aren't Harvard's. HBR's own 2011 audit of 2,241 US firms found something bleaker: an average first response of 42 hours, 23% of companies never responding at all, and firms that responded within an hour qualifying leads about 7x more often than the ones that waited longer.
Two decades of tooling later, the pattern holds: signal → minutes → relevance and signal → days → noise.
This is the part the token industry never priced in. The same sentence reads differently at different ages. "Saw you were looking at pricing this morning" is attention. The identical sentence four days later is surveillance. The words didn't change. The moment did.
RELEVANCE DECAY (interactive: slide time-since-signal from 10 minutes to 4 days; the same sentence is re-read at each point) The sentence: "Saw you were looking at pricing. Here is the answer to the cap question." At 10 minutes the sentence reads like attention (the question is still live). At 1 day it reads like a follow-up (the moment cooled, the question may already be answered). At 4 days it reads like surveillance (the same words, now just proof you were watching). Meter values illustrative, not a benchmark. The words never change. The moment does.Personalization is not a copywriting problem. It is a timing problem wearing a copywriting costume.
Most teams climb toward relevance one rung at a time, and each rung has a failure mode.
{FirstName} personalization fails because it's free. Firmographic personalization ("companies like yours in fintech") fails because it describes a category, not a customer. Persona personalization ("as a Head of Growth, you probably...") fails because it guesses at pain instead of observing it. Moment personalization is the top rung: it responds to a thing that actually happened, while it is still happening.
THE PERSONALIZATION LADDER (static figure). Four rungs, bottom to top, each with its failure mode. 1. {FirstName}: "Hi Brad, I noticed Civic is doing big things." Fails: it is free, so it signals nothing. 2. Firmographic: "Companies like yours in fintech are moving fast." Fails: describes a category, not a customer. 3. Persona: "As a Head of Growth, you probably care about pipeline." Fails: guesses at pain instead of observing it. 4. The moment: "You and a teammate compared plans yesterday. Here is the answer to the cap question." Responds to what happened; cannot be batched. The first three rungs can be templated, which is why they are crowded.The first three rungs can be batched, scheduled, and templated, which is exactly why they're crowded. The top rung can't be batched. That's why it works, and why almost nobody operates there.
Here is the operational truth the deck-stage version of this argument skips: responding to moments is a staffing problem.
To personalize to the moment for every lead, prospect, and trial user, someone has to watch every signal source, decide which flickers matter, match each one to the right response, and act inside the window while it's still a moment. For a Growth team already paying for the pipeline it isn't converting, that someone does not exist. I've watched teams solve this with heroics for exactly one quarter.
This is where Bryn sits. Bryn is not another dashboard to watch. It is the governed execution layer that runs Plays through your stack. Bryn watches the signals you already own, scores what fired against your definition of a good account, and runs the Play you approved while the signal is still warm. The pricing-page flicker gets the pricing answer in minutes. The trial user who hit a cap gets the cap conversation the same hour. Every run writes its receipts: the signal, the score, the Play, the approval, the outcome.
Personalization stops being a copywriting job and becomes a routing decision you make once, at Play design, the way we run it ourselves. Bryn watches. You decide the Play. Bryn runs it.
Yesterday Chris wrote about the buyer's clock: accountability enforced deal by deal, faster than any regulator. The personalization clock is the same shape, just faster still. The buyer's patience for irrelevance is measured in minutes.
This note was the why: tokens are free, moments are scarce, and the word "personalization" hides three different jobs.
Part 2 is the how: the mechanics of stage-aware Plays, what "personal" concretely means at each handoff for leads, prospects, and trial users. Part 3 is the proof: measuring personalization as pipeline and revenue instead of open rates and vibes.
If your outreach still opens with a rendered token, you don't have a personalization program. You have a database with manners.
Stop watching signals. Start running them.