B2B lead generation on a market you can count by name

TL;DR. We analysed the 2024 annual accounts of the 732,239 Romanian companies that reported sales. Only 18,421 of them, or 2.52%, sold more than 3 million euros, and those same companies account for 79.2% of all reported sales. If you sell to businesses in a market like this, lead generation means getting chosen by a few thousand companies you can list by name.

In a market this small, the method that works is an analysis of each company, built from public data and sent to the decision-maker before the first call. The cost of a lead is then counted with the whole system that produces it, meaning people, tools, data and time.

Lead generation is the process by which a company finds potential customers and brings them to a first sign of interest, such as a reply, a request for a quote or a meeting.

A lead is the company, or the person inside it, who showed that interest. In B2B, where you sell to other companies, there is almost always a group of people behind a lead, deciding together.

The best-known B2B lead generation playbooks assume a large market like the United States, where a seller can work through tens of thousands of contacts and still have a market left.

Many European markets do not work that way. Romania is a clear example. The companies worth selling to are few, their owners know each other, and a careless mass email gets talked about.

Maybe in your industry everything sells through people you already know and through referrals, so advertising seems to change nothing.

Or maybe you have spare capacity on the factory floor, yet orders arrive in waves, and what happens between the quote and the order no longer depends on you.

Both remarks are fair. And both lead to the same question, which is how you reach customers who do not know you yet without damaging your name in a market where everyone talks to everyone.

Our guide to B2B sales systems covers the whole machine, from the funnel to the numbers leadership should watch. This article stays on its first floor, the moment new companies enter the funnel, and on the cold email method that suits a market you can count. Cold email means emails sent to people who do not know you yet.

What is B2B lead generation, and how is it different from prospecting?

B2B lead generation is everything a company does to get first signs of interest from other companies that could become customers. Prospecting is only one part of it, the active part, where you pick the company and reach out first.

The rest runs the other way, because the customer finds you, through an article, a referral, a Google search or an answer from an AI assistant.

The two directions have established names. Outbound means you make the first move, by email, phone or a LinkedIn message. Inbound means the customer makes the first move after finding something of yours. A healthy company runs both, in different proportions.

The difference matters first for the calendar. Prospecting can bring the first replies a few weeks after the first sends. Blog articles work on a different clock, over months and years, but they keep bringing people in long after you stop sending.

One more term is worth settling before the numbers. A qualified lead is a lead that matches your ideal customer profile, the ICP, and has already taken a step toward the next stage, for instance by asking for an analysis or agreeing to a call.

How we build the ideal customer profile explains why that profile gets written before the first email, not after the first replies.

The table below puts the four main sources of leads side by side. It carries no numbers, because results depend on the industry and on what each company brings to the table. It only shows who moves first and what each source asks of you.

The last row looks fastest, and in small-volume B2B it is fast, but only up to the first click. The smaller the market, the more a wrong message costs you. The rest of this article starts from there.

How big is the B2B market you can actually reach?

In Romania, the B2B market can be counted by name. Of the 732,239 companies that reported sales for 2024, only 18,421 sold more than 3 million euros, and they made 79.2% of all reported sales. The figures come from the annual accounts filed with the Romanian Ministry of Finance, which we collected and analysed ourselves at UNRIVALS.

We chose the 3 million euro threshold because from that size up a company usually has a budget for strategy and, more importantly, a team that can carry out what gets decided. Of the 18,421 companies, 7,602 had at least 50 employees in 2024, which also means a structure where the buying decision passes through several people.

The red card shows the real size of your market if you sell to large companies. The list is finite, and it also moves.

Of the 18,421 companies, 17,940 also filed accounts for 2025, and 1,581 of those fell below the threshold. In the other direction, 2,678 companies that had sold less than 3 million euros in 2024 crossed it in 2025.

For lead generation, this changes how you think about volume. In a market with millions of buyers you can afford to lose people along the way, because others keep coming.

In a market of a few thousand companies, every wasted contact is a customer lost for a year. And a few thousand new names a year mean the list has to be rebuilt every year.

Our doctrine, SIGNAL DOCTRINE™, the set of principles we work by, has a principle for exactly this situation, which we call the customer profile as a pattern.

It says there are patterns of customers with the same recurring need, and places where you find them grouped together, so you target the community where decision-makers already gather, to shorten the time to your first customers.

In practice, you work on named accounts. You pick companies on data anyone can check, such as sales, profit, industry and headcount, and then you look for the person who decides in each of them. Only after that do you write the first message.

Why does the American lead generation playbook struggle in a small market?

The American playbook assumes a large market, where you can test thousands of messages and lose half your contacts without worrying. A market like Romania’s is small, tied together by personal connections and sensitive to mass mailings. On top of that, the B2B buyer now covers most of the journey alone, before speaking to any salesperson.

International data confirms that second part. A Gartner survey published in March 2026, run with 646 B2B buyers in August and September 2025, found that 67% of them prefer to buy without talking to a sales rep. In the same survey, 45% said they had used AI during a recent purchase.

The 6sense 2026 Buyer Experience Report, based on more than 4,600 B2B buyers, goes further. The vendor a buying group favours before it talks to sales wins the deal 82% of the time. Roughly 60% of the buying cycle passes before the buyers talk to a seller. Until then, the group gathers information and builds a shortlist.

Who decides matters too. According to Forrester’s State of Business Buying 2024, an average of 13 people inside the organisation take part in a buying decision, and 89% of purchases involve two or more departments. The same report found that 86% of purchases stall at some point along the way.

The panel above shows where a B2B lead is really won. If the favourite is chosen before the first call, your email cannot sell on the first try. It has one job, which is to get you onto the shortlist.

The small-market layer sits on top of all this. In many industries, potential customers know each other and find out quickly who sent a careless message.

Founders are right when they say business comes through referrals. The trouble is that a referral cannot be scheduled, and a factory with spare capacity cannot wait for the phone to ring on its own.

This is where a second principle from the doctrine comes in, sequencing, where the market is educated gradually. It states that every stage has its own objective, and skipping a stage means burning it and losing.

The first contact works like a first meeting. In lead generation, that means the first email asks for a small step the decision-maker can take in ten seconds, and the long conversation comes later.

What works in B2B prospecting in a small market?

In a small market, account-based prospecting works, where every company receives an analysis built from public data and sent to the decision-maker before the first call. The method follows from the conclusion of the previous section. If the favourite is chosen before the first contact with sales, the first contact has to bring the decision-maker something useful already.

In a market of 18,421 large companies, the method also has a simple arithmetic argument. On a list of a few thousand companies, an analysis for every company is doable, and the work can be spread over months and quarters. We believe the American playbooks choose volume precisely because, in their markets, that kind of work would never end.

A finite list also changes how you treat each company. Since thousands of other names are not coming up behind it, every company becomes a separate account, with its own history, the person who decides and the reason you would write to that company in particular.

The method has five stages, in this order.

1. The list from the registry. Romanian companies above the 3 million euro threshold, filtered by industry and by your ideal customer profile, with last year’s accounts checked. The industries you do not work with leave the list from the start.
2. The decision-maker’s address. You look for the address carrying the name of the founder or managing director. Generic addresses such as office@ are handled separately, with different copy.
3. One observation about their company. Every email starts from a public fact about that particular company, and the observation is written for that company alone.
4. One small step. The email asks for one thing, a reply from the decision-maker saying they want the analysis.
5. The analysis, then the call. Whoever says yes receives an analysis of their company’s market position, built only from public sources, before any conversation.

The fifth stage is what makes the difference. The diagnosis happens before any contact, from public sources, and the analysis itself becomes the proof.

If this much came out without a single conversation, what could come out of sitting down together? The method has a drawback too. Some decision-makers may find the analysis too general, because nobody has spoken to them yet. The drawback is worth accepting, since the analysis only goes out after the decision-maker has asked for it.

The order of the stages matters as much as the stages themselves. An analysis requested with a one-line reply reaches a decision-maker who is expecting it and already knows why it is arriving.

The first email matters most. Instantly’s 2026 cold email benchmark report found that 58% of cold email replies come from the first email in a sequence. That is why stages 3 and 4, the observation and the small step, deserve the most care.

Checking addresses in the second stage also matters for delivery. In a purchased, unverified list, many addresses may no longer exist, and the emails sent to them never reach anyone. Email providers quickly penalise a sender with many dead addresses, so even good emails can end up in spam afterwards.

Our guide to B2B cold email goes into the message itself, from subject line to follow-up. Here we keep only the order, which is the list, then the decision-maker, then the reason to write, then the small step.

How much does a B2B lead cost?

The cost of a B2B lead is everything the system spends on people, tools, data and time until the first sign of interest, divided by the number of companies that reached that stage.

Counting only subscriptions and lists gives a cost that is too low and tells a false story. A B2B lead costs whatever the whole system that produces it costs.

The table below shows what goes into each of the four cost components and how to measure it in your company.

People are the component most easily left out of the calculation, because salaries already sit in another budget. In account-based work, every company receives an observation and an analysis written for it alone, and that work is done by people, even when a language model (AI) helps them research faster.

Data looks free, because annual accounts are public. But they have to be collected, cleaned and updated every year, since the list changes. The 2,678 companies that crossed the threshold in 2025 show how quickly new names appear.

Time is the component that never shows up on an invoice. While the list is being built and the message tested, expenses keep running and calls have not yet arrived.

The reply rate changes the whole calculation, because it tells you how many companies you have to contact to get one reply. Instantly’s 2026 report puts the average cold email reply rate at 3.43%, the top quartile at 5.5% and the top 10% of senders above 10%.

At the other end, Belkins measured an average reply rate of 0.45% across 7,530,489 emails sent in 2025. Belkins measures campaigns it runs for its own clients, while Instantly aggregates everyone using its platform, so the two figures describe different populations.

The gap between the two reports shows why cost per lead cannot be copied from a study. The same monthly expenses, divided by far fewer replies, give a far more expensive lead.

Comparing yourself with these benchmarks also calls for care with the maths. Instantly divides replies by emails sent. If you divide the companies that replied by the companies contacted, the figures cannot be compared directly.

To find your own cost, first define a lead strictly. For instance, a lead is a company that replied, received an analysis or held a call, counted once, in the month of its first sign of interest. A clear refusal can count too, because it is also information about the market.

Then divide all of the system’s expenses for a month by the number of companies that reached each stage, meaning the companies contacted, those that replied, those that received the analysis and those that held a call.

The better question for you is a different one, namely how much you can afford to pay for a call.

The answer comes from three of your own numbers, the average contract value, the margin you keep on it and how many calls it takes, on average, to close one contract. The marketing KPIs that change a CEO’s decision show how to tie them to your acquisition budget.

A hypothetical example makes the maths visible. An equipment company has average contracts of 100,000 lei, the Romanian currency, keeps 20% of each as profit and closes one contract out of every five calls.

Each contract leaves 20,000 lei, so a call is worth 4,000 lei to that company. Below that ceiling a call is a good deal, while one at 6,000 lei would be a loss.

What is the biggest mistake in B2B lead generation?

The biggest mistake in B2B lead generation is counting leads as the final result. A lead is only the entrance to the funnel, and what matters to the company is whether it moves to the next stage. A campaign that reports hundreds of leads and no calls has shifted work from marketing to sales without doing it.

Our doctrine calls this the double funnel of marketing and sales.

It says there are two overlapping funnels, one of differentiation, run by marketing, which takes a company from suspect, a company that merely fits the profile, to prospect to analysis, and one of synergy, run by sales, which takes it from negotiation to contract, retention, development and profit. Marketing materials must match the concrete sales stage.

The rule that follows applies directly to lead generation. The objective of every piece of material is the next stage, not the final sale. The email aims for the reply, the analysis aims for the call, and the call aims for a meeting with everyone who decides. The B2B double funnel shows in detail how the two halves connect.

The second mistake is about the list. A purchased database, full of generic addresses and companies outside your profile, produces bounces, complaints and a tarnished name. In a market where decision-makers know each other, a careless message sent to thousands of companies stays attached to your brand long after you sent it.

The right column takes more work per company, and that shows straight away in the cost per contact. It also shows in the later stages, because a reply to a message written about their company already carries a reason to talk. In the left column, the work moves to whoever has to sort through the wrong replies.

The third mistake is about the message. The temptation is to write about yourself, your years of experience, your portfolio and your services. A decision-maker reads the first email to find out what you know about them. If the first line could be sent unchanged to hundreds of other companies, they recognise it as a mass mailing and close it.

Which layer fixes lead generation, from L1 to L4?

Our methodology arranges the work on four layers. L1 is performance marketing, and L4 is positioning, the reason a company would choose you.

Between them sit L2, revenue and the commercial process, and L3, which we call the AI Brain, the system built on artificial intelligence that keeps the company’s data and decisions. Lead generation can break on any of them.

The red layer in the figure, positioning, decides what a decision-maker says after reading the analysis. The other three layers only carry the message to them. A perfect sending system cannot save a company that does not know why it deserves to be chosen.

A fourth principle from the doctrine applies here, infrastructure before marketing. It says a company rarely has a marketing problem, and more often lacks infrastructure, meaning systems, channels, positioning and a brand foundation. Marketing built on a weak base burns money.

In lead generation, infrastructure takes a very concrete form. It means an updated list, verified addresses, one place where every email sent and every reply is recorded, and a person who answers the same day. Without them, a campaign that looks good on paper gets lost in inboxes nobody reads anymore.

L3, the AI Brain layer, makes it possible to prospect a large number of companies while keeping the individual approach. In a system like this, every email sent, every reply and the status of every company sit in a single database.

The observation about each company is written from its own data, and a language model (AI) helps prepare observations for many companies without turning them into a template.

Without that memory, every person on the team would start from scratch. They would write to a company that said no a month ago, or repeat a message that did not work. The system’s memory stands in for the relationship until the relationship exists.

L1, the advertising layer, has a role too. Our view is that in account-based B2B, ads do their job after the decision-maker has already received the email and the analysis. Run alone, at the start, they pay for clicks from people outside your profile.

How do you build B2B lead generation step by step?

A B2B lead generation system takes about 90 days to build, on one product and one channel. Our methodology says a position, the reason a customer would choose you, gets tested on one product and one channel for ninety days, and only then generalised. The order below follows the stages of the method described above.

1. Write your ideal customer profile on one page. Industry, size in sales, headcount, region and the problem you solve for them. Also write down which companies you do not want, because a list without exclusions grows fast and goes bad just as fast.
2. Build the list from public data. In Romania, the annual accounts filed with the Ministry of Finance give you sales, profit and headcount for every company, and the trade registry tells you who runs it.
3. Find the address of the person who decides. The founder, the managing director or the commercial director, depending on company size. Check every address before sending, so it does not bounce.
4. Write one observation for each company. A fact from its accounts or its website, tied to the problem you solve. If you cannot find anything to say about a company, it is probably not for you.
5. Ask for one small step. A short reply is enough. The big step comes after the decision-maker has received something useful from you.
6. Prepare what people get when they reply. An analysis, a case study or a comparison with the market, made for their company. This is where the call is won.
7. Measure every stage. How many companies replied, how many received the analysis, how many held a call and what each stage cost, using the full cost of the system.

Steps 4 and 6 take the most time, which makes them the easiest to skip. They are also the only ones a competitor cannot copy overnight. An observation written for a single company passes the competitor test. The test asks whether a competitor could send the same sentence unchanged, and here the answer is no.

Our B2B sales process for premium services picks up where this article ends, from the call to the contract. Good lead generation hands sales a company that already knows who you are and what you found out about it.

Frequently asked questions

Can ChatGPT do lead generation?

ChatGPT can help with lead generation, but it cannot do it alone. It helps you write faster, summarise a website or a set of accounts and prepare an observation for each company. The list, address verification, sending and same-day replies remain jobs for a system and for people.

What are some effective AI tools for B2B lead generation?

The most useful tools work on your own data, meaning the list, the email history and the replies. A language model (AI) helps with researching each company and with the first draft of the message. Sending and address verification tools keep delivery under control.

What is a good reply rate for B2B cold email?

Instantly’s 2026 report gives an average of 3.43% replies per email sent, 5.5% for the top quartile and above 10% for the top 10% of senders. Belkins measured 0.45% across 7,530,489 emails sent in 2025. Compare your own figures with these benchmarks only if you calculated them the same way, meaning replies divided by emails sent.

How long before lead generation produces the first calls?

Outbound brings the first replies within a few weeks of the first sends. A stable system, with a clean list and measurement at every stage, takes about 90 days to build. Content and organic search bring results over months and years.

What is a qualified lead in B2B?

A qualified lead is a company that matches your ideal customer profile and has already taken a step toward the next stage. One example is a company that asked for the analysis or agreed to a call. A reply from a company outside your profile is still a lead, but not a qualified one.

What to do on Monday morning

In a market of a few thousand companies, lead generation means being on the shortlist before the first phone call.

Start with the list. Pull from public accounts the companies that fit your profile and count them. If you end up with a few hundred, you are looking at work you can finish in a quarter, with a message written for each company. A finite list tells you exactly how much work there is.

Want to see what a decision-maker learns about your company before you ever speak? Ask for a diagnostic and we will show you what the market can already see. In a small market, the first company to find out what customers think of it wins the shortlist.