Strategy · 2026-08-18 · 8 min
Using LinkedIn Outbound to Find Your First Ten Customers
The first ten customers of a B2B SaaS company are not found through a funnel. They are found one named account at a time. Here is how we approach that stage differently from the outbound we run once a company has scale.
The first ten customers of a B2B SaaS company do not arrive through a repeatable process. They arrive because a founder, or someone acting like one, had a specific conversation with a specific person who had a specific problem. Trying to build a scalable funnel before those ten exist usually wastes the months you cannot afford to waste.
LinkedIn outbound is well suited to this stage, but only if it is run the way this stage demands: slow, narrow, and personal. The version of outbound that works once you have fifty customers and a defined ICP looks nothing like the version that gets you your first ten.
Why volume is the wrong instinct here
With no customers yet, the instinct is to widen the net. Message everyone who might plausibly fit. That instinct is backwards. At this stage you do not have enough signal to know who fits, so widening the net just produces more noise to sort through with no more insight per message.
The better move is to go narrower than feels comfortable. Pick twenty to thirty accounts you can name individually, where you have a specific reason to believe the problem exists, and treat each one as its own small project rather than a row in a spreadsheet.
Building the first list
A useful first list for early-stage B2B SaaS is not built from a firmographic filter. It is built from evidence.
- People who have complained about the problem publicly. A comment on a post, a question in a forum, a job posting that describes the pain you solve.
- Companies structurally similar to a design partner you already have, if you have one. Same size band, same tooling, same market motion.
- Warm-adjacent accounts. Second-degree connections of your own network, where a mutual contact can make an introduction feel less cold even if the message itself is still outbound.
Ten to fifteen names built this way, researched properly, will outperform two hundred names pulled from a generic filter every time at this stage.
Message anatomy for founder-led outbound
At this stage the message should read like it came from a founder, because it did.
- Open with the specific reason you picked them. Not "companies like yours." The actual thing you noticed.
- Name the problem plainly, in the terms the reader would use, not the terms from your pitch deck.
- Offer something concrete and low-risk. Not a demo. A short, specific conversation, or an offer to build the first version of something with their input.
- Sign it like a person, because it is one.
A worked example: "Saw your post about the manual work your team does reconciling data between tools X and Y each week. We are building something specifically for that handoff and would value ten minutes to understand how you deal with it today, no pitch attached." That message works because it is true, specific, and low-pressure.
What a good outcome looks like
Do not judge this stage by reply rate benchmarks built for mature ICPs. A healthy outcome from thirty well-researched first messages is a small number of real conversations, of which a smaller number still become design partners or paying customers. That is not a disappointing result. That is roughly what founder-led outbound looks like everywhere, described honestly rather than with invented percentages.
| Stage | List size | Message style | Goal |
|---|---|---|---|
| Pre-PMF research | 20 to 30 | Question-led, no pitch | Validate the problem |
| First ten customers | 20 to 30 named | Specific reason, concrete offer | Convert conversation to pilot |
| Post-PMF scale | Hundreds, ICP-filtered | Templated with real personalisation | Repeatable pipeline |
Common ways this stage goes wrong
- Handing it to a junior hire or a tool too early. At ten customers, the person sending the message needs enough context to improvise in the reply thread. That is a founder's job, not a task to delegate.
- Chasing volume when conversion feels slow. The fix for slow early conversion is usually a sharper problem statement, not more names on the list.
- Skipping the follow-up. Many of the first ten customers say yes on the third or fourth touch, not the first. A single unanswered message is not a no.
Once you can point to ten customers and describe, in your own words, what they have in common, you have the beginning of an ICP worth scaling into. That is also the point at which it is worth reading our take on how it works for what a repeatable version of this looks like, and checking our case studies for how that transition tends to play out.
Next: pick the ten accounts you would be proudest to close, and write the first message to each one individually before you write anything else.