How to find your first B2B customers: five real sources

The warm network is a grant of a few free conversations. It hides, for about a quarter, the fact that you have no repeatable way to start one.

There is a specific day in most new B2B companies when the warm list runs out. You have called everyone who owes you a favour, everyone from the last job, and the two people who said they would introduce you to someone. Three of them were polite, one bought, and the rest went quiet. That day arrives sooner than anyone plans for, and what you do in the week after it decides whether the business has a pipeline or a founder with a phone.

The warm network is not a strategy. It is a grant of a few free conversations that hides, for about a quarter, the fact that you have no repeatable way to start one.

Start with what the law lets you do

This is the step that gets skipped, and in the Gulf it is the one with teeth. If you are selling into Saudi Arabia, the Personal Data Protection Law reaches you whether or not you are in the Kingdom. Article 2 of the PDPL states that the law applies to processing of personal data of individuals residing in the Kingdom "by any means from any party outside the Kingdom". Buying a list in London and mailing it from Cairo does not put you outside it.

The regulator is unusually direct about what bad practice looks like. In SDAIA's guide to the PDPL for controllers and processors, one of the listed examples of bad data protection practice is a business that collects contact details and then "starts regularly sending marketing information to collected emails" without having identified a lawful basis for doing so. The guide is explicit that you must identify a lawful basis for each purpose of processing, as specified in Articles 5 and 6.

Read that as a design constraint rather than a reason to stop. It does not say you cannot do outbound. It says you need to know which basis you are relying on before the first send, and you need a record of it. Founders who treat that as paperwork end up rebuilding their entire prospecting operation in year two, usually at the request of a large customer's procurement team.

The five places first customers actually come from

In roughly the order they pay off for a company with no brand yet:

• People who have already bought this problem solved. Not your network, their category. Someone who bought a competitor two years ago and churned knows the problem exists and has a budget line for it.
• The referral your warm list can still give you. The favour is exhausted, the introduction is not. Ask for one specific name rather than a general keep-me-in-mind.
• Direct outreach to a narrow, named list. Twenty accounts you can describe individually beats two thousand you bought.
• Places your buyers already gather. Industry associations, procurement portals, sector events. Slower, and the conversations start warmer than anything cold.
• Partners who already sell to your buyer. The longest to build and the only one that compounds.

Notice what is not on that list: content, ads and a website rebuild. Those work, and they work later. They are demand generation, and at this stage you do not have enough traffic for them to produce a single conversation this month.

Narrow further than feels comfortable

The instinct with an empty pipeline is to widen. Anyone with a budget, any sector, any size. It feels like more shots on goal and it produces fewer, because a message written for everyone reads as a message written by a machine, and because you cannot learn anything from a hundred rejections that have nothing in common.

Pick one sector and one company size for ninety days. The point is not that this segment is the right one forever. The point is that twenty conversations inside one segment teach you the vocabulary, the objection and the buying process, and twenty conversations spread across nine segments teach you nothing you can reuse.

Qualify for the buying process, not the interest

Early interest is cheap and misleading. People will take a call with a new company out of curiosity, and curiosity looks exactly like a lead in a spreadsheet. What separates the two is whether there is a route to a signature.

Three questions do most of the work. Who else has to agree. What has to be true for this to be bought this quarter rather than next year. And what happened the last time they bought something similar. The third one is the most useful and the least asked, because the answer tells you the real procurement path rather than the one your contact imagines.

Deals that skip this stage do not fail immediately. They fail four months later in a place nobody scoped, which is the pattern we set out in the five places B2B deals actually die.

Volume is not the variable you think it is

The standard advice is to send more. It is not wrong so much as it is the wrong end of the equation. Doubling volume doubles your rejections at the same rate, and it burns a finite list of named accounts you cannot buy back.

The lever with more travel in it is relevance: whether the first sentence demonstrates that you know something specific about that company. That is not a personalisation token. It is a reason the message could not have been sent to anyone else. On the channel side, the mechanics of doing this without getting your account restricted are covered in what the LinkedIn rules actually allow, and there is a good complementary piece on the non-ad routes to first conversations in Khaled Badr's write-up of direct outreach as a growth channel, written from the practitioner side rather than the sales-team side.

Count conversations, not activity

The first metric worth having is not emails sent or connections made. It is the number of conversations with someone who could sign. That number is small at the start, which is exactly why it is the honest one, and it is the only input that predicts revenue in a business with no history to forecast from.

Once there are enough of them to see a shape, the forecasting discipline in building a pipeline you can actually forecast starts to apply. Before that, a CRM full of stages is theatre.

The honest limit

None of this is fast. A first B2B customer in a considered-purchase category typically takes longer than a founder's runway assumes, and the delay is usually procurement rather than persuasion. If your plan needs revenue inside sixty days, outbound to net-new accounts is the wrong instrument, and the honest alternatives are narrower: sell a smaller scoped engagement to someone who already knows you, or accept a longer runway.

We would also rather say plainly that the first ten customers are usually won by the founder, in conversations that do not scale, using judgement that is not yet written down anywhere. Hiring a salesperson to do this before you can describe how it works reliably produces an expensive quarter and a departure. If the early months produce activity but nothing that advances, the cause is more often structural than personal, which is the argument in a system problem rather than an effort problem.

A sequence that holds up

Decide the segment and write down why. Build a named list of twenty accounts, not a bought list of two thousand. Establish and record your lawful basis before the first message goes out. Spend the first two weeks on referrals and prior-category buyers, because they answer faster and teach you the vocabulary. Then start direct outreach with what you learned, and measure conversations rather than sends.

At the end of ninety days you will not have a machine. You should have a segment you understand, a message that gets answered more than it did in week one, and a small number of real opportunities. That is what a first pipeline looks like, and it is worth considerably more than a large list of people who have never heard of you.