Define who's worth talking to
An ideal customer profile is a written description of the kind of buyer your product is genuinely good for. Without one, "generate leads" has no target and every list is as good as any other — which is why untargeted outreach performs so badly.
Who: Regional travel agencies, 2-15 staffDoing: Building custom itineraries for individual clientsToday: Rebuilding documents by hand each time a booking changesTrigger: Peak season, when volume makes the manual process hurtDecider: The owner, usually also one of the agentsNot for: Large tour operators with in-house systems Solo agents doing under ~5 itineraries a monthThe two most useful lines are the last ones. Writing down who it's *not* for is what makes a list buildable, and it's the part people avoid because exclusion feels like lost opportunity.
Building the first list
- Directories and trade associations for the industry — unfashionable, public, and accurate
- Map and local business data, when your ICP is geographic
- Professional networks filtered by role and company size
- Reviews of competing products, where people describe their problem publicly with a name attached
- Communities and forums where the work is discussed — join, read, and note who asks the recurring questions
- Existing customers' referrals, which convert better than any of the above
Aim for a first list of fifty specific, named organisations rather than a thousand rows. Fifty is small enough to research individually, which is what makes the outreach land — and large enough for a pattern in the responses.
Outbound and inbound
| Outbound | Inbound | |
|---|---|---|
| You control | Exactly who hears from you | Nothing — you control what you publish |
| Speed | Replies this week | Months |
| Cost shape | Your time, per contact | Upfront effort, then compounding |
| Lead quality | Variable — they weren't looking | High — they came looking |
| Scales by | Doing more of it, or hiring | Publishing more, with a lag |
Early on, outbound is nearly always the right answer, because you need conversations now and you need to learn what language works. Inbound is what you build with the language outbound taught you. Running them in the other order is how people spend six months publishing content aimed at nobody in particular.
Outreach that gets a reply
- Research for two minutesFind one specific, true thing about them. That sentence is the entire reason the message gets read.
- Lead with their situation, not your productThe first line should be about them. If it's about you, it reads as a template.
- Ask for something smallFifteen minutes, or a single question they can answer in one line. Not a demo, not a trial.
- Keep it under a hundred wordsLonger messages get read less, not more carefully.
- Follow up twice, then stopTwo follow-ups spaced a week apart is diligent. Five is harassment, and it costs you the referral too.
Tracking it without software
Below roughly twenty active conversations, a spreadsheet is genuinely sufficient. Company, contact, stage, last touch, next action, next date. The only column that matters is next action — a pipeline is lost through forgotten follow-ups more than through lost deals.
- One row per organisation, not per email
- Every row has a next action and a date, or it's closed
- Review the whole sheet once a week, on the same day
- Move to a proper CRM when you can't hold the state in your head — usually somewhere past twenty active conversations
Common mistakes
- Skipping the ICP, so every list looks equally good
- Buying a list instead of building fifty rows by hand
- Investing in inbound before outbound has taught you the language
- Follow-up sequences that continue well past a reasonable stop
- A pipeline sheet with no next-action column, so deals go quiet unnoticed
Key takeaways
- An ICP that names who it's *not* for is what makes list-building possible
- Fifty researched organisations beats a thousand purchased rows
- Outbound first for speed and learning; inbound second, built on what you learned
- The next-action column is the pipeline; everything else is decoration
Try it yourself
Write your ICP including the "not for" lines, then build a list of twenty named organisations that match. If you can't find twenty, either the ICP is too narrow or the market is smaller than assumed — and both are worth knowing now.
