I used to have a one-page brief. Industry, role, pain point, value prop, tone. The standard form. I filled it out for the first founders I worked with and I watched what came out the other end, and it was the same problem every time: the message technically said the right things and the prospect still archived it.
It took me longer than I'd like to admit to figure out why. The brief was capturing what the founder wanted to say. It wasn't capturing what the prospect needed to hear in order to want to reply. Those are two completely different jobs, and I was doing the first one and pretending it was the second.
The four-layer brief is what I built when I finally understood what the brief actually had to do.
The layers stack. You can't skip one. You can't fill in layer four until you've done layers one through three, and you can't write the message until layer four is in place. Each layer answers a different question, and a message that skips any one of them fails in a specific, identifiable way.
Layer one — what she says in public.
Her bio, her website, her last three posts, the category she puts herself in, the keywords she uses when she's being asked what she does. This is the easiest layer and the most misleading. Stated priorities are written for strangers — for recruiters, for the press, for the version of her audience she's never met. They're calibrated for safety. They are the safest possible version of what she does, and they are the wrong material to write a message from.
I read layer one and then I deliberately set it down. Not to ignore it, but because if I let layer one drive the message, I write the message a hundred other people are also writing this week. The prospect can spot a layer-one message in under three seconds because she wrote layer one herself and she remembers what it sounds like.
Skip every other layer and write only from layer one, and the message reads as: "this person read my bio." She gets twelve of those a day. She archives all twelve, without thinking about it, because pattern-matching is faster than reading.
Layer two — what her behavior implies.
This is where you find out what she's actually working on, as opposed to what her bio says she works on. Hiring posts are the loudest signal. A VP of Sales who's posted three "we're hiring AEs" updates in five weeks is not primarily thinking about strategic outbound scaling. Whatever her bio says, her week is structured around interviews and onboarding and ramp. A message that lands on the hiring reality — "you've added four AEs in two months, and the ramp math alone is brutal right now" — sits in a different category than a message that lands on her bio.
Other things that count as layer two: board changes, executive departures, product page edits, the way the company's category description shifts over time, conference appearances she's accepted recently, podcast appearances she's chosen to do. Anything where she made a decision rather than wrote a sentence. Behavior is harder to dress up than text. That's why it's where the real signal lives.
A message that includes layer two reads completely different from one that doesn't. It reads as: "this person has been paying attention for a while," rather than "this person Googled me this morning." That difference, which sounds soft, is actually the hardest line in outbound to fake.
Layer three — what she's protecting against being seen as.
This is the layer almost nobody reads, and it's the one that decides whether she replies.
B2B buyers are not buying products. They are buying defenses against being seen as the one who chose wrong. Every senior buyer is running, in the back of her head, all the time, a calculation about how each decision is going to look. To her CEO. To her board. To her peers in the same role at peer companies. To her team. Every email she answers is a tiny version of that calculation. If replying to your email could make her look like the person who got conned by a vendor, she won't reply, even if she likes the message.
A CRO who just took a new role is defending against being seen as the person who couldn't fix the pipeline in her first ninety days. A VP of Marketing whose last campaign underperformed is defending against being seen as the person who keeps spending into the same channel. A founder who just raised a Series B is defending against being seen as someone who raised on a story she can't actually deliver. The defensive priority is real, and it changes by quarter, and it determines what she can publicly engage with versus what she'll only engage with privately.
This layer is the one that takes the longest to read for. It's the one that requires the podcast moment, the long-form interview, the place her cadence broke. It requires holding the hiring page next to the investor letter and asking yourself, "what is she trying not to look like right now?" There is no shortcut. There is no AI prompt that produces a clean answer here, because the answer lives in things the company has not said.
If you skip layer three, the message gets a polite read and a polite non-reply. The prospect thinks, good message, and then thinks, I don't have a good answer to this that I can give in public, and archives it. You'll never know that's what happened. You'll just see the reply rate.
Layer four — the actual reason to write.
One sentence. The specific thing I noticed that nobody else would have noticed.
This is what the whole message is built around. Not the value prop. Not the category. Not the offer. The specific thing. It might be a sentence from minute seventeen of a podcast. It might be the gap between her hiring page and the investor update. It might be that her last three posts all use the same word and her company's last three press releases all use a different one.
Layer four has three tests, and they're the tests I run on it before I write a single sentence of the message.
First: would another vendor doing the same homework also have surfaced this? If yes, it's not layer four yet. It's still layer two with extra steps. Go back, throw it out, find another one.
Second: does it neutralize the defensive priority from layer three, or does it poke at it? If it pokes at it, she won't reply. The right layer-four observation gives her a reason to engage that defends the act of engaging. The wrong one makes her feel exposed, and exposed people don't reply to strangers.
Third: if her CEO walked past her desk and saw her reading my message, does the message defend her time spent reading it? Or does it look like a vendor pitch she should have already archived? This is the test almost nobody applies, and it's the one I apply hardest. The senior buyers I want to write to are not reading email in private. They are reading email with peripheral awareness of everyone around them. If a message doesn't defend their attention publicly, it gets archived publicly, and fast.
When a layer-four sentence passes all three tests, the message writes itself. Sentence five lands. The close is easy. The whole thing comes in around seventy words and it lands clean.
That's the brief.
The reason I take it this seriously is that I've watched what happens when one of the layers gets skipped. Skip layer one and the message sounds like you don't know who she is. Skip layer two and it sounds like you read her bio. Skip layer three and she'll like the message intellectually and won't reply, and you'll never know why. Skip layer four and you've written a perfectly polished version of the templated message everyone else sent her this week, which is the same outcome as not having written anything at all.
The brief is not the form the founder fills out before the work begins. The brief is the work.
Everything that comes after — the message, the cadence, the second touch, the follow-up — is downstream of whether the brief got built right. If the brief is wrong, no amount of writing craft saves the message. If the brief is right, even a moderately good writer produces a message that lands.
The reason most outbound fails is not that the writing is bad. It's that the brief stopped at layer one and someone wrote a perfectly competent message from the wrong starting material.
I get that wrong sometimes too. The difference is I know which layer I missed, and I can go back and fix it. Most outbound shops don't even know there's a layer three to miss.