Where we start
Six sessions, one product. Cadence is a fictional AI note-taker with two ways to buy it: a solo professional signs up alone, and a team lead buys the Teams plan for a group. It is the same Cadence other courses on the shelf build a brand and a launch campaign for; here we make the decisions underneath those: who, what promise, which channel, what a customer is worth, and the plan on one page. Every Cadence number is constructed and says so. Alongside it runs something real: the public Olist marketing funnel, 8,000 genuine leads and the 842 that became paying sellers, which session 4 puts on a bench.
★ What you walk out with tonight. Five candidate segments for Cadence scored on one sheet, a beachhead you can defend in one sentence, an ideal customer profile with a trigger event, and a "not for" list that is as specific as the "for" list. The same sheet works on your own product tomorrow.
A market is people who buy alike
5 min live"People who have meetings" is a description, not a market. A market, in the sense that matters for spending money, is a group you can sell to the same way: they want the same thing, they buy it the same way, and they tell each other about it. Bill Aulet's version, from the MIT entrepreneurship programme, names exactly those three conditions, and the third one is the one most segment lists skip.
LiveThe three conditions, and why the third one decides budgets4 min▶
Aulet's MIT material defines the conditions for a group to count as one market: the customers "all buy similar products"; they "have a similar sales cycle and expect products to provide value in similar ways"; and "word of mouth communication exists between customers in the market".
Similar products means one sales message works. A similar cycle means one sales motion works: a card swipe on a team budget is a different business from a nine-month procurement with a security review, even if the user is identical. Word of mouth means each customer you win makes the next one cheaper, because the group passes your name around for free. Without it, every sale costs the same as the first one, forever.
Segment on situation and behaviour before demographics. "Managers aged 30 to 45" fails all three tests; "user researchers who record five or more customer interviews a week and share the notes with product" passes all three, and you can find them.
Self-studySegment is not persona, and neither is the ICP3 min▶
Three words get used as if they were one. A segment is a group of buyers that behaves as one market. An ideal customer profile (ICP) is the description of the accounts inside a segment you most want: size, situation, trigger, disqualifiers. A persona is one person inside one of those accounts, the end user or the buyer. Aulet's advice is to build the persona from "one actual real end user" rather than a composite, which is the difference between a person you can phone and a poster on a wall.
On a team product the user and the buyer are often different people. Cadence's user is the researcher; the buyer is the research lead with a team budget. Write both down.
Olist sorted its own sellers by what they sell. The 842 sellers who closed in the public Olist funnel carry a business segment, and the top of the list is not random: home decor 105, health and beauty 93, car accessories 77, household utilities 71. Those are groups that buy the same service from a marketplace, onboard the same way, and sit in the same supplier forums. A marketplace that knows which segments close can write a message per segment instead of one message for "anyone with stock".
Win one, then move next door
5 min liveA small company cannot be everywhere, so it picks one market to win outright and uses that win to reach the next. Aulet calls it the beachhead: "the place where, once you gain a dominant market share, you will have the strength to attack adjacent markets". Geoffrey Moore's Crossing the Chasm popularised the same idea with a D-Day analogy and a sequence of bowling pins, each won market knocking into the next.
LiveFour questions that pick a beachhead5 min▶
Aulet's own summary is short: "Select one market segment where you feel you have the highest odds of success while also giving you legitimacy and strategic assets to grow further." In practice teams score their candidates on four questions, the same four you will use in the build-along:
- Pain. Do they already spend time or money working around the problem? A workaround is evidence; a nod in an interview is not.
- Reach. Can you get in front of them cheaply? Do they gather somewhere: a community, a conference, a newsletter they all read?
- Word of mouth. Will a happy customer tell three others like them, without being paid to?
- Adjacency. Once you win it, what does it make cheaper? A reference customer in the next market, a feature the next market also needs.
Size is a fifth question, and the trap is to ask it first. The biggest segment is usually the most contested and the least uniform. A small market you can dominate beats a large one where you are the eleventh option.
Self-studyWhat Moore actually added, and what is reported3 min▶
Crossing the Chasm (1991, third edition 2014) is about the gap between early adopters, who buy a vision, and the mainstream, who buy a proven solution from peers. Moore's prescription, as it is widely summarised, is to concentrate everything on one niche in the mainstream until you own it, then move to adjacent niches in sequence: the bowling pins. The D-Day framing and the bowling-alley sequence are reported here from summaries; the book was not opened for these pages, so no sentence of his is quoted.
The two authors agree on the move and differ on the reason. Moore's reason is credibility: mainstream buyers need references from people like them. Aulet's reason adds economics: word of mouth inside a tight market lowers the cost of every next sale. Both are reasons to say no to tempting customers outside the beachhead for a while.
The deals that close tell you who your beachhead already is. In the Olist funnel, of the 842 sellers who closed, 332 were tagged "online_medium", sellers already trading online at a middling scale, against 14 tagged "online_top". The marketplace was winning a specific kind of seller: big enough to have stock, small enough to need a marketplace. When your own closed deals cluster like that, the beachhead is not a theory to choose; it is a fact to notice and then double down on.
The ICP, and the list of who it is not for
5 min liveAn ICP turns the beachhead into something a salesperson, an ad platform or a partner can act on: which accounts, in what situation, triggered by what event. It has no single canonical author; it is working vocabulary across B2B sales and marketing, so treat any definition, including this one, as a tool rather than a standard. The part teams skip is the disqualifiers, and they are the part that saves money.
LiveFive fields and a not-for list4 min▶
Account: the kind of organisation, with a size band. Situation: the observable behaviour that means the pain is present (Cadence: they already record interviews). Trigger: the event that turns a fit into a buyer this month, such as a new hire, a new tool, a reorganisation or a failed audit. Buyer and user: named separately, because the message for each is different.
Not for: the accounts you will decline or deprioritise even if they ask. Each line needs a reason: regulated legal teams need retention controls Cadence does not have; solo freelancers buy on a different cycle and belong to the self-serve motion, not this one; on-premise buyers need an architecture that does not exist. Writing it down does not stop anyone signing up. It stops you spending to find them.
Self-studyTwo motions, one beachhead3 min▶
Cadence sells two ways: solo professionals sign up on their own, and team leads buy for a group. The beachhead belongs to one motion, the Teams motion here, because that is where the three market conditions hold most tightly. The solo motion does not disappear; it keeps running on whatever spend it needs to stay alive, and it often feeds the team motion when a solo user brings the tool to work. What changes is where the deliberate effort goes for the next two quarters.
The same split appears in any product with a free tier and a paid team plan. Deciding which motion owns the beachhead is the first decision; letting each motion pretend it does is how both end up underfunded.
A missing field is also a profile. Of the 842 closed Olist sellers, 779 left "has company" blank and 773 left catalogue size blank: the fields an ICP would most want were the ones the funnel did not collect. If your CRM cannot tell you the size band and the trigger of the deals you won, the first ICP task is not writing the profile. It is adding the two fields to the form.
Score five segments and pick a beachhead
★ 22 min · everyone buildsA spreadsheet or a sheet of paper. Five candidate segments for Cadence, three market tests, four beachhead questions, one ICP card. The scores below are constructed for Cadence; your scores for your own product will differ, and the disagreements in your team are the useful part.
List five segments by situation, not by title. Each one a sentence a stranger could use to find ten of them. Avoid "SMBs" and "millennials"; write what they do.
Run the three market tests, yes or no. Similar products, similar buying cycle, word of mouth between them. One "no" removes the segment from the beachhead race. It can still be a later market.
Score the survivors 1 to 5 on pain, reach, word of mouth and adjacency. Write one line of evidence next to every 4 or 5. A score with no evidence is a hope.
Pick the beachhead and write it as one sentence with a reason. The reason should name a market test or a score, not a feeling.
Fill the five ICP fields for the beachhead. Account, situation, trigger, buyer, user. The trigger is the hardest and the most valuable: it tells sales when to call and tells session 3 which channel to test.
Write the not-for list, each line with its reason. Three lines is enough. If a line has no reason it is a prejudice; if it has one it is a budget rule.
Keep the sheet. Session 2 writes the promise for exactly this ICP, and session 3 chooses channels by asking where these specific people already are. If the sheet changes later, everything downstream should be re-read.
Before session 2
30 min- Run the sheet on your own product. Five segments by situation, the three tests, the four scores with one line of evidence each.
- Pull your last 20 closed deals and tag each with the segment it belongs to. If one segment holds more than half, compare that with the beachhead your sheet picked. Disagreement is information.
- Find the trigger in three real wins. Ask, or read the CRM notes: what happened in the month before they bought?
- Show the not-for list to sales and ask which line they would fight. The line they fight is the one worth discussing.
Three questions
Clear all three and session 1 gets stamped in your passport.
1 · Solo consultants score high on pain but fail the word-of-mouth test. What does that mean for the beachhead?
The tests come before the scores. Word of mouth is what makes each won customer lower the cost of the next one; a group without it is a list of people, not a market you can dominate. Merging two groups to pass a test hides the problem rather than solving it.
2 · Which ICP field tells sales when to call?
Fit says whether they could buy; the trigger says why now. A new research lead, a new repository, a reorganisation: each is observable from outside, which is why the trigger also points session 3 at channels.
3 · Of 842 closed Olist sellers, 779 left "has company" blank. What is the first ICP task for a team in that position?
An ICP is only useful if you can tell, from data you hold, whether a lead fits. Guessing the blanks invents a profile; dropping the field gives up on it. Collecting it is cheap and makes every later session measurable.
What this session teaches, and where it came from
Every fact on this page, and its verification tier, is recorded in the course's source map.
Session 1 cheat sheet · pin this
You know who Cadence is for. Session 2 asks what you promise them, and starts somewhere most messaging workshops never look: what these researchers would do if Cadence did not exist. April Dunford's five components of positioning, built from that answer up, for the exact ICP you wrote tonight.