B2B Lead Generation: Channels, Costs, and What Actually Converts

B2B lead generation is the work of getting people at other businesses to raise their hand: a reply, a booked meeting, a form fill. The question that decides whether it works is not which channel is best. It is which channel you can run consistently at the headcount you actually have.

That reframe matters because every channel in this article works for someone and fails for someone else, and the difference is rarely the tactic. A channel that needs 100 dials a day fails at a firm with one person who also delivers the work. A channel that needs six months of compounding fails at a company that needs meetings this quarter. So this guide goes channel by channel, with the numbers we could trace to a real study, what each channel's ceiling is, and how to measure whichever you pick. Where a popular statistic has no traceable source, we say so and leave it out.

What B2B lead generation means in practice

Definitions drift, so fix one. A lead here means a person at another business who has shown they might want what you sell. A meeting is a lead who has agreed to talk. A held meeting is one that actually happened with someone who has a real problem and the authority to act on it. An opportunity is a held meeting that goes into your pipeline.

Most reporting stops at the first of those, because it is the easiest to count, and that is where the damage is done. A thousand form fills and ten held meetings are very different results, and two channels that look equal on leads can be far apart on meetings. The rest of this article judges every channel on how it performs further down that chain, not just at the top.

One boundary worth drawing now: lead generation is not the same as demand generation. Demand generation builds awareness and preference before anyone is buying. Lead generation captures and converts people showing interest. We cover the difference, and why it matters for the shortlist, in a separate piece, because it changes which channels deserve your first dollar.

The buying context every channel operates inside

Before choosing a channel, understand what you are up against. 6sense surveyed nearly 4,000 B2B buyers for its 2025 Buyer Experience Report (46% North America, 20% Europe, 20% UK and Ireland, 14% Asia-Pacific; 42% technology, 38% services, 14% manufacturing). The median purchase cost was $200,000 to $400,000, so this skews toward larger, longer deals than many small firms sell, but the pattern is instructive.

The same report found first contact with a seller happens about 61% of the way through the buying journey, down from 69% the year before, and that the average cycle is 10.1 months. Put plainly: most of the decision is made before a seller is invited in, and the buyer, not the seller, usually starts the conversation.

That does not make outbound pointless. It means lead generation has two jobs on two timelines. One is being known and trusted before anyone is shopping, which is slow and compounding. The other is reaching the right people at the moment a need appears, which is fast and repeatable. The channels below split between those two jobs, and a sensible plan has at least one of each.

The seven channels at a glance

Cold email: the economics and the ceiling

Cold email is the cheapest channel per touch and the most abused, so the numbers are worth reading carefully.

Belkins analyzed 7.5 million cold emails sent across 2025 and tracked 34,393 replies, an average reply rate of 0.45%. Note the denominator: that figure is replies divided by total emails sent. Older benchmarks, and many vendor blogs, divide by opens or by delivered messages, which produces much higher percentages for the same behavior. Belkins stopped tracking open rates because they became unreliable. When two sources quote very different cold email reply rates, check the denominator before assuming one of them is wrong.

By seniority, founders and owners replied at 0.57%, C-level executives at 0.42% and VPs at 0.32%. Morning sends (8am to noon) did a little better than late evening, 0.54% versus 0.40%. The size effect is the one to build a plan around: the smaller the company you target, the better the reply rate, which suggests cold email may suit firms that sell to small businesses better than to large enterprises.

How many emails does a meeting take? Belkins reports it booked over 1,200 appointments across those 7.5 million emails. Dividing one by the other gives roughly one appointment per 6,250 emails or fewer. That division is our arithmetic, not a figure Belkins publishes, and it spans many campaigns of varying quality, so treat it as an order of magnitude.

The ceiling is deliverability and infrastructure, not creativity. Google requires authentication and a low spam rate for bulk senders, and in Saleshandy’s analysis of 53.1 million cold emails (January to June 2026), fully authenticated domains (SPF, DKIM and DMARC together) were 2.7 times more likely to land in the inbox than unauthenticated ones. Our cold email operator's guide covers what moves that number, and the compliance guide covers what the law requires in the US, EU and Canada. The annotated templates post shows what a three-touch sequence looks like when it is built from data.

Cold calling: the economics and the ceiling

Belkins' calling benchmarks come from more than 175,000 dials: a 9.9% per-dial connect rate, 58% of connects becoming a conversation, 4.6% of conversations booking a meeting, and about one meeting per 370 dials end to end. Dialing the same prospect about three times lifted the connect rate from 9.9% per dial to 24.5% per prospect.

The ceiling is rep time. At 370 dials per meeting, calling only pays when a single meeting is worth a lot, and it only scales with people. Bridge Group's survey of SDR teams found a median of 112 daily activities per SDR (44 phone, 41 email, 19 LinkedIn, 8 other), so a full-time rep running a phone-led motion produces a fixed amount of outreach regardless of how good the list is. Cold calling also carries its own legal and consent rules, which differ from email's, especially if AI voices are involved. We worked through the full data and the rules in our cold calling analysis.

LinkedIn outreach: the economics and the ceiling

Expandi's H2 report covers 13.2 million connection requests across 13,302 accounts. The average acceptance rate was 28.5%, with morning sends accepting at roughly 32% to 37% against 23% to 24% in the evening. Messages of 151 to 200 characters produced an 8.5% reply rate, three follow-ups produced a 9.8% reply rate, and five or more dropped to 5.0%, worse than a single message. AI-hyperpersonalized messages lowered acceptance by 12% in within-account comparisons.

Those are strong response numbers next to cold email. What is missing is the number that matters: no credible public source we found publishes a connection-to-meeting rate, so any cost per meeting for LinkedIn outreach you see quoted is a guess. Track it yourself.

The ceiling is the platform. LinkedIn limits the volume an account can safely send each day, so growth means more accounts or more patient sequencing, and both carry account risk. We broke down the limits in our LinkedIn outreach guide, and the 13.2 million-request analysis has the timing and length data in full.

Paid: the economics and the ceiling

Paid is the one channel where you can buy volume on demand, and it has the best public price data, with one important catch.

Metadata.io analyzed 153 B2B advertisers and $57.6 million of 2025 ad spend, which produced 211,000 leads.

The click prices are almost identical, about $9 to $10. The gap is conversion: 5.9% of LinkedIn clicks became leads against 1.9% of Google clicks. Within LinkedIn, lead-gen forms cost $193 per lead and landing pages $346. Two cautions. Metadata's own pages describe the $202 figure differently, one as a spend-weighted aggregate and one as the typical advertiser's median, and the spread across advertisers is wide, so read $202 as a mid-point, not a promise. And Metadata notes that click-through rate is close to useless for comparing channels, because a search click and a feed click mean different things.

The catch is that these are leads, not meetings. Nobody publishes a reliable lead-to-meeting rate for B2B paid, because it depends on how you qualify. So here is the arithmetic with the unknown made explicit.

The ceiling is your budget and your follow-up speed. Paid scales as far as you will fund it, and then the question becomes how fast someone talks to the lead. On the speed side, a famous Harvard Business Review audit of 2,241 companies reported that contacting a lead within an hour made it far more likely to qualify, though the study is from 2011 and the exact multiples should be read as indicative.

Referral and partner: the economics and the ceiling

Referrals are the channel everyone says is best, so it is worth being strict about what is actually known.

The B2B referral statistics that circulate widely, such as referrals converting at around 26% or closing 69% faster, mostly trace to vendor blogs quoting each other, and we could not find a primary source for them. We have left them out. What we can trace is narrower.

Schmitt, Skiera and Van den Bulte tracked roughly 10,000 customers of a German bank for almost three years and published their findings in the Journal of Marketing in 2011. Referred customers had higher retention and, over the period, a lifetime value at least 16% higher than similar non-referred customers, with the range reported as 16% to 25%. That is a consumer bank, not a B2B services firm, so it is directional evidence, not a B2B benchmark.

The B2B evidence that is solid points the same way from a different angle. In 6sense's data, 97% of buyers had prior experience with at least one vendor on their shortlist, 85% had previously bought from the vendor they chose, and 68% personally knew sellers from the winning vendor. Relationships decide a great many deals before a seller is invited in.

The ceiling is that you cannot turn referrals on. They arrive in proportion to the work you have already done for people, and you cannot buy more of them next month. Partner channels (agencies, consultants and tool vendors who serve the same buyer) follow the same logic, and we found no credible public benchmark for partner-sourced pipeline. Treat referral and partner as a base layer worth protecting, not as a lead generation plan.

Content and SEO: the slow compounding channel

Content is the channel that works on the first of the two jobs, being known before anyone shops. 6sense found 61% of the journey happens before a seller is contacted, which is where search and content do their work.

It is also the channel with the most ways to fail quietly. A study of 912 million blog posts found 94% earn zero backlinks, and we broke down what the posts that do earn them have in common. Ruler Analytics, tracking over five million conversions across 13 industries, put organic search at a 4.9% conversion rate, but that spans very different business models and a loose definition of conversion, so it says little about B2B services specifically.

The ceiling is time and consistency. There is no published, reliable timeline for when a B2B content program starts producing meetings, and anyone who quotes you one is guessing. What you can say is that it compounds only if you keep publishing, and that it pays off best in firms whose sales cycle is long enough to benefit. Our content strategy framework and the guide to closing the loop between content and leads go deeper.

AI search: the newest channel, and the least measured

A growing share of research starts in an AI assistant, which makes being cited there a lead generation channel in its own right. G2's buyer behavior report, as reported by the research aggregator Omnibound, found 29% of buyers now start research through an LLM more than through Google (we could not trace the primary report, so treat it as indicative).

We measured how that plays out. In our own AI visibility audit of 29 real B2B companies across six industries, average visibility was 42.8%, with marketing agencies averaging 77% and insurance brokers and management consultancies averaging 17% to 27%. A later, larger audit of 30 B2B SaaS companies averaged 67.3% (see it here). AI assistants answered comparison queries 69% of the time but problem-first queries only 21% of the time. The full audit is here, and how to get cited is the practical follow-up.

One more data point, with a caveat: Ruler Analytics reported AI referral traffic converting at 5.8% against 4.9% for organic search, in a cross-industry dataset with a loose definition of conversion. It is suggestive, not proof.

The ceiling is that the channel is new, volatile and unevenly measured. Results swing sharply by industry, so check where you stand before you invest. Our free AI visibility audit runs five buyer-style queries and shows who gets named instead of you.

Outbound or inbound: which one you can afford to start with

Put the two jobs side by side. Outbound (email, calls, LinkedIn) reaches people at the moment a need appears and produces meetings in weeks, but it needs a list, infrastructure and consistent execution. Inbound (content, SEO, AI search, referrals) compounds and costs less per meeting over time, but it needs patience and gives you no control over timing.

The practical test is cash and time:

• If you need meetings this quarter, start with outbound on a narrow, well-defined list. It is the only family of channels that can produce them that fast. Budget for the setup: warming sending domains takes 4 to 6 weeks in practice, so the first meetings do not arrive on day one.
• If you have a longer runway, start building inbound alongside it, because the 61% of the journey that happens before a seller is contacted is only reachable that way.
• If you have neither the time nor the staff for the work, the decision is not which channel but who runs it, which is the next section.

We lay out how a one-person firm should approach outbound in our guide for consultants, and how signal-based triggers change when you reach out.

Buy, build, or outsource

Once you know the channels, the operating model is a separate decision: build it in-house, buy tools and run it yourself, or outsource it.

The most useful numbers on the build option come from The Bridge Group's 2025 SDR research, a survey of 351 B2B companies (78% North American, 83% B2B SaaS, median revenue $47 million, median deal size $50,000). It is observational survey data and skews to larger software companies, so a small services firm should read it as a guide to the shape of the problem.

Add the rest of the picture: median SDR on-target earnings of $80,000 (a $55,000 base and $25,000 variable), an average ramp of 3.0 months, an average tenure of 1.9 years, and 4.1 quality conversations per SDR per day. On our arithmetic, a 3.0-month ramp against a 22.8-month average tenure means about 13% of an average SDR's time in the seat is spent getting up to speed. And on-target earnings are not the cost: benefits, tools, management time and recruiting come on top.

The takeaway is not that hiring SDRs is wrong. It is that one hire is a real fixed cost with real failure odds, which is exactly why outsourcing exists. Outsourcing converts that into a variable cost and hands you a team that has already built the infrastructure, in exchange for less direct control. We compare the two common outsourced models in our done-for-you versus done-with-you piece, and our own outsourced offerings are outsourced SDR and managed outbound sales. Our Inbound package (AEO, GEO and SEO) covers the content, search and AI-search channels above, and Full Hybrid joins both lanes in one CRM record. Pricing for all three is on the pricing page. Before you sign with anyone, get two things in writing: what counts as a qualified meeting, and who owns the data and the sending domains.

Is it worth it? A worked example

Whether a channel pays comes down to one comparison: what a held meeting is worth against what it costs. Here is the arithmetic with every assumption visible.

Suppose a new client is worth $20,000 in first-year revenue, and 20% of held meetings turn into clients. Each held meeting is then worth $4,000 in expected first-year revenue. A channel that delivers held meetings for less than that is profitable on first-year revenue alone, before any repeat business. In the earlier paid illustration, both LinkedIn Ads ($1,010 per meeting) and Google Ads ($2,620 per meeting) clear that bar at a 20% lead-to-meeting rate, though Google has much less headroom once you count no-shows and unqualified meetings.

These inputs are made up for illustration. The point is the method. Replace $20,000 and 20% with your numbers, then compare the result with each channel's real cost per held meeting. If the channel loses money at your numbers, no tactic fixes it.

How to measure it: the six numbers that matter

Check them in order and the leak shows itself. If you cannot reach your list, nothing downstream matters. If you reach them and they do not respond, the message or the timing is wrong. If they respond and meetings do not happen, the problem is the offer or the follow-up. If meetings happen and nothing enters pipeline, you have a qualification problem, which is the one that makes outsourced bookings look better than they are.

If you want to see where your own pipeline stands before choosing a channel, QualiPi is a short self-assessment of how predictable your revenue and pipeline are.

Sources

• 6sense, The B2B Buyer Experience Report for 2025, nearly 4,000 buyers plus a companion survey of 766
• Belkins, B2B Cold Email Response Rates, 7.5 million emails, 2025
• Belkins, B2B Cold Calling Benchmarks, 175,000+ dials, 2025
• Expandi, State of LinkedIn Outreach H2, 13.2 million connection requests across 13,302 accounts
• Metadata.io, B2B Advertising Benchmarks and What LinkedIn Ads Cost, 153 B2B advertisers, $57.6 million of 2025 spend
• The Bridge Group, 2025 SDR Models, Motions and Metrics Report, 351 B2B companies
• Schmitt, Skiera and Van den Bulte, Referral Programs and Customer Value, Journal of Marketing, 2011, about 10,000 customers of a German bank
• Ruler Analytics, Conversion Rate Benchmarks, 5 million+ conversions across 13 industries
• Oldroyd, McElheran and Elkington, The Short Life of Online Sales Leads, Harvard Business Review, 2011
• Deligatr, AI visibility audit of 29 B2B companies, our own analysis, n=29
• Saleshandy, Email Deliverability Statistics, 53.1 million emails
• G2 buyer behavior figure as reported by Omnibound, B2B Buying Statistics, primary report not traced; the 912-million-post study is cited with its original source in the linked Deligatr article

FAQ

Related reading

• Cold email: the complete operator's guide
• AI SDR: what it is, what it does, and where it fails
• Demand generation vs lead generation: what is the difference
• Does cold calling still work? What 175,000 B2B dials show
• LinkedIn outreach: what 13 million connection requests show
• Signal-based outreach: how it actually works
• Done-for-you vs done-with-you outbound
• How to run outbound as a consultant