How to start a B2B newsletter that books sales conversations
A founder's newsletter earns its place when it gives you a reason to email the people you want to sell to. Judge it on the conversations it books with them, and decide before the first issue when you will stop.
Start a B2B newsletter as a reason to email the people you want to sell to, and judge it by the conversations it books with them. Ask one named buyer at a time to be featured in an issue: a feature interview is a request with nothing for sale in it, strangers answer it, and the interview is a sales conversation as well. Fix one format with a word budget, publish weekly for eight issues, send it from a person, and write down before issue one how few conversations will make you stop.
The example below is one plan for a weekly newsletter aimed at the people who own new business at B2B agencies. Put your own buyer in their place and the method stays the same.
Subscribers are the wrong measure
The usual advice is to grow a list and launch into it. Three founders who launched that way, two of them to audiences of twenty to thirty thousand, got two customers, a plateau at $7,000 to $8,000 of monthly recurring revenue, and a stall near $1,500. An audience built by a founder talking about building also attracts the wrong reader: other founders, price sensitive, quick to churn and mostly trying out new tools.
Email is a relationship platform, and the list is yours. Its size is still the wrong number to judge it by. A newsletter platform's own growth features, such as recommendation swaps with other publications, work by volume, and volume brings in exactly the reader you do not sell to. The plan put the trade as a sum: two hundred agency owners beat five thousand founders, and the second number is the one that feels like progress. So leave the subscriber count out of every update: in the plan's words, it is the metric that will "quietly replace the real one".
A publication is a reason to email a stranger
"Can I show you my product?" goes unanswered. "I write a weekly piece on how B2B agencies win new business. Can I feature your desk?" is a legitimate, flattering ask, and strangers answer it.
The feature interview and a customer conversation are the same twenty minutes. One request produces an issue and a conversation with a buyer, so the issues are the by-product and the conversations are the output. Ask what you would ask on any first call, and for the feature, ask how they run new business: where the list comes from, what gets sent and what they have stopped doing.
The reason to bother is arithmetic. A kit for forty first conversations, working backwards from rates it marked as assumptions, put cold email at 4 to 8% from send to conversation. At that rate forty conversations need 500 to 1,000 researched sends, which over four weeks is 125 to 250 a week while also holding the calls. The kit concluded that the constraint was finding people to talk to, and that the answer was warmer ways in rather than more cold volume. A feature request is one of them.
The newsletter tests your words as well. Each issue reaches the segment you sell to, so a phrase that gets replies there can move to your website after it has been tried. That costs nothing, where paid ads buy the same iteration with a budget you can lose.
Write the feature request before issue one
A twenty-minute call takes about two weeks to land in a calendar, so send the first three requests on day one, before the first issue ships. Here is the plan's request, with the sign-off left as a placeholder:
Subject: Your new-business desk: 20 minutes?
Hi {First name},
I write a weekly teardown for people who own new business at B2B agencies: one company a week, the research a first meeting deserves and almost never gets.
Every third issue features how one agency runs its new-business desk: where the list comes from, what gets sent, what they have stopped doing. I would like that to be {Agency}.
Twenty minutes, your words, and you see the piece before it goes out. {One specific line about why them: a client win, a niche, something they published.}
Worth a slot?
{Sender}
Three rules for sending it. Personalise the line about why them or do not send it, because an unpersonalised version is a worse cold email than an honest one. Send it from a personal address, never from a marketing tool. Log every reply as it lands, because each call is one of the conversations you are counting.
A yes gives you an issue, a conversation, a named relationship and a quote to open the next request with. A no costs one email and often returns a referral.
One format, with a word budget
Choose at most three formats and let one of them carry most issues. The plan had a spine, a feature every third issue, and a short piece built on one number for a week that runs short.
Its spine was a teardown of one real company: the research someone should have done before a first meeting with them. Seven sections, always in this order, each with a budget:
| Section | What goes in it | Words |
|---|---|---|
| The line | Name, what they do, one number | 30 |
| Why now | The trigger, with dates | 120 |
| The one thing | The single fact the approach rests on | 30 |
| Two signals for, one against | The against makes the other two believable | 220 |
| Who to approach | Roles, because a name needs a source, and why that seat owns the problem | 120 |
| The opening line | The sentence you would send | 80 |
| Sources | Plain links | links |
The reader gets the verdict before the evidence and the evidence before the caveats. An issue runs 600 to 900 words, and a teardown that runs to 1,400 is a different product and a worse one. Everything in it is checkable, or it does not run, and the against is the section that gets forwarded.
Pick each subject by four tests, and require all four. A reader recognises the name without being told who they are. There is a dated public trigger in the last 90 days: funding, a hire, a launch or an acquisition. Your reader could plausibly sell to them. Nothing in the piece needs a source you cannot link. For the first issue, which every later subscriber reads first, a company with a boring, well-documented trigger beats an exciting one that rests on a claim you cannot link.
Weekly, eight issues, and the stop decided now
Publish weekly, on a fixed day. The plan chose Tuesday and called fortnightly "the lie a busy operator tells themselves". Name all eight issues before the first one ships, because a queue decided in advance is what stops a weekly publication dying on the question of what to write this week.
Then set the stopping rule while it is cheap. The plan's: fewer than six conversations with agency owners traceable to the newsletter by issue eight, and it has failed as a way to find people to talk to, so stop or change the request. Subscribers appear nowhere in that rule. Make the eighth issue the review, and publish the decision in it.
Set it up so people answer
Say who it is for. Naming the buyer gets more replies from the right people and fewer readers overall, which is the right trade when the goal is booked calls. "I write about how B2B agencies win new business" gets a reply, and "I write about go-to-market" does not. The line under the newsletter's name names the format and the cadence and claims no category. The plan's was The research a first meeting deserves. One company, every Tuesday.
- Send it from a person. Use a first name and a surname as the sender, never the company. A newsletter from a person gets opened, and one from a brand gets filed. Set the reply address to an inbox you read, because the replies are the point.
- End the welcome email on one question. It is the one email every subscriber opens. The plan's asks What are you trying to win right now? and invites the reader to name the account, or the type of account, because subjects are picked from the replies. A reply is a conversation, and the answers fill the queue with companies your buyers care about.
- Test delivery before issue one. Send yourself a draft, open it in the two mail systems your buyers use at work, check it lands in the inbox rather than a promotions or junk folder, and read it on a phone.
- Write issue one before you announce anything. An empty publication converts nobody.
- Leave out a paid tier. A cheap subscription beside the thing you sell competes with it for the same attention and attracts the price-sensitive reader.
- Give the newsletter original writing. If the platform will not let you set a canonical URL, republishing an article that already ranks on your site points a high-authority duplicate at your own page.
Five things
- Judge it on conversations booked with named buyers, and leave the subscriber count out of every update.
- Ask one buyer at a time to be featured, with three requests sent on day one.
- Fix one format with a word budget, and a linkable source for every claim.
- Publish weekly for eight issues, named in advance, with the stopping rule written down before the first.
- Send it from a person, and end the welcome email on a question.
Keep reading
- How to define an ideal customer profile you can build a list from Most of an ideal customer profile can be looked up. The half that decides who buys is the half only you hold: the customers you already have, why they bought, the person it is for, and who you will never approach.
- How to prepare for a sales call before you dial The preparation that changes a first call is done before it starts: make the thing you would otherwise describe, send it ahead, and write the call down in minutes, from the opening line to the one thing you will ask for.
- How to clean a prospect list before the first email goes out The number on a list is the part everyone checks, and the errors sit in the rows: the wrong company under the right name, one brand counted three times, research that says nothing, and people who should never hear from you.