What does a B2B lead gen agency typically cost in 2026?

Monthly retainers of $3,000 to $15,000 for SDR work and $5,000 to $25,000-plus for full service; percent of sourced pipeline at 10% to 20% of opportunity value; pay-per-meeting of $150 to $600 mainstream and $800 to $2,500-plus enterprise; pay-per-lead $25 to $500; setup $1,500 to $5,000; three to six month minimums. All vendor self-reported. No independent price index exists.

How do I measure ROI from a lead generation agency?

Track cost per opportunity and per closed deal rather than cost per lead, over a window longer than your sales cycle, since First Page Sage puts sales-qualified lead to closed-won at around 12% for B2B SaaS, on its strict definition where an SQL has already met a rep. Control for your own speed-to-lead first: in a 2011 study of 1.25 million leads, sub-one-hour response was roughly seven times likelier to qualify a lead, though a co-author of that study sold lead-response software.

What are the pros and cons of hiring vs in-house lead gen?

In-house gives control and product depth at an agency-estimated $110,000 to $175,750 fully loaded per rep, three months to ramp and 40% median annual attrition. Outsourcing gives a two to six week start and claimed 25% to 65% first-year savings. Practitioners in the SaaStr poll report weaker product knowledge, only 7% of voters said it really worked, and no controlled study has compared the two.

How do I choose the best B2B lead gen agency?

Insist on a written definition of qualified, held-versus-booked meeting terms, your ownership of domains, mailboxes and data, quarterly exit checkpoints, and proof of list provenance. Treat Clutch and G2 scores cautiously, since they are agency-solicited. The G2 and Clutch rankings are built from reviews agencies solicited, and no independent per-agency satisfaction survey exists.

Which industries benefit most from outsourced lead generation?

The evidence here is thin and agency-published. Agencies claim traction in financial services, insurance, professional services and wealth management, where one deal covers many appointments, and I could not find a non-agency source for that claim. Practitioners report the worst results on complex, high-ACV sales generally; the SaaStr commenters describe high-AOV products with deep buyer expectations, not SaaS alone. No independent segment-level study exists.

When does outsourcing actually work?

When you have already proven the motion in-house and want extra capacity, a new region or a new segment, and when your deal size leaves headroom over the cost per meeting. It fails most reliably before product-market fit, on high-ACV complex sales, and in regulated categories where you inherit the compliance exposure.

Who is liable if the agency's emails break the rules?

You are, in most jurisdictions. The UK regulator puts responsibility on the sender or instigator of the message, and CAN-SPAM makes no exception for business-to-business email. So the contract needs a data-processing addendum and documented list provenance, and you should register the sending domains yourself.

← Blog

Outbound·20 min read

Is Hiring a B2B Lead Generation Agency Worth It?

The one independent data point is a LinkedIn poll: 7% of teams said outsourced SDRs really worked. I sell outbound, and I am publishing that anyway.

TathagataFounder, ParaphrasePublished August 31, 2026

The only large independent data point is a public LinkedIn poll, and in it 7% of teams said outsourced SDRs really worked. Almost every page answering this question is published by an agency selling the service. I sell outbound work, and I am publishing both numbers anyway.

Last updated: 31 August 2026

The question people actually type is blunter than the headline. From Quora: "Is there such thing as a good B2B lead generation company?" The doubt is already there before the research starts.

Here is the problem with the research. On 24 August 2026 I logged every source six AI engines cited on this exact question. Twenty-nine domains came back. Twenty-three of them sit on the study's own 206-domain scoreboard, the domains cited often enough across my forty test questions to classify, and the other six were cited too rarely. Nineteen of those twenty-three are lead-generation agencies selling the service. One is Reddit. Three are software vendors. That is 83% of the classifiable set, and just under two-thirds (19 of 29) of the domains cited.

One square per classifiable source. This is not a conspiracy, it is just who bothers to publish, but it means most of what you find is evidence about what sellers claim.

That is not a conspiracy. It is just who bothers to publish. But it means most of what you find is evidence about what sellers claim, not about what buyers got.

I sell outbound work. The number in the next section is against my interest. It is also the best evidence that exists.

The strongest independent signal in the category, and it is a three-year-old LinkedIn poll of self-selected voters that nobody has repeated. Read it as a signal, not a settled fact.

That poll ran on Jason Lemkin's personal LinkedIn account in May 2023, and he wrote it up on SaaStr on 16 May 2023. His own summary: "what I personally haven't seen is an outsourced SDR team replace an in-house one." Even the engagements he counted as wins only "sort of" worked, and neither startup stuck with the agency past four to six months.

The practitioner comments Lemkin reproduces from the LinkedIn thread, his selection rather than the unfiltered thread, are more useful than the headline number, because they say when it fails. James Rehm, COO of Skuuudle: "Hasn't worked for us, and we've trialed a number of well-managed, diligent organisations. For a high AOV and a complex product, there is no substitute for having your own team and investing in their expertise in the product and ICP over time." Collin Cadmus, formerly VP Sales at Aircall, names the pattern that matters most: "most of the people who pay to outsource SDRs are usually the same people who have not figured out how to make the SDR model work for them internally. I've actually never seen a company that's cracked the outbound model then switch to outsourcing it." Eric Harrington of TeamSupport adds the product-depth problem: "Customers expect more product knowledge, even at the appointment setting engagement."

Read together, those three describe the conditions rather than a verdict. Outsourcing tends to work when you have already proven the motion in-house and want extra capacity, a new region or a new segment, and when your deal size leaves headroom over the cost per meeting. It reliably fails before product-market fit, on high-ACV complex sales, and in heavily regulated categories where you inherit the compliance exposure.

One caveat on the evidence itself: that poll is three years old, it was answered by a self-selected slice of SaaStr's LinkedIn following rather than a sampled population, and no one has repeated it. It is the strongest independent signal available, not a settled fact.

Model	Published range
Monthly retainer, SDR or appointment setting	$3,000 to $15,000
Monthly retainer, content/inbound or full-funnel programme	$5,000 to $25,000+
Pay per meeting, mainstream ICP	$150 to $600
Pay per meeting, enterprise (Cleverly)	$800 to $2,500+
Pay per lead, list and data vendors	$25 to $500
Percent of sourced pipeline	10% to 20% of opportunity value
Setup fee	$1,500 to $5,000

Each bar is the published span, not a market rate. There is no independent price index for this category, so read these as asking prices.

There is no independent price index for this category. I looked. Every range above traces to an agency or a tool vendor, so treat them as asking prices rather than market prices.

The number that surprises buyers is not the retainer. It is what the retainer excludes. Contact data, tooling seats, sending domains, mailboxes and creative are routinely billed separately. Ask for the total cost at your target meeting volume, in writing, before you sign anything.

Two pricing structures deserve suspicion. Commission-only and percent-of-pipeline as the sole model is unstable, because the agency carries all the risk and reprices or walks when the maths stops working. And any per-meeting price below roughly $150 should prompt a question about what is being counted as a meeting.

Same word, four products, a twentyfold spread. The stricter the unit, the higher the number, which is why an average cost per lead tells you almost nothing.

That definitional problem is the whole story. A "lead" in one dataset is a form fill, in another an MQL, in another a booked meeting, in another a sales-accepted opportunity. WordStream's 2026 figure of $66.69 comes mostly from search-ad form fills. Belkins puts its own cold email cost per lead at $250 to $300, and its average across all client campaigns at $770 spanning 17 industries. It put cybersecurity at $750 to $1,500 in August 2025 and at $1,750 to $2,800 in its March 2026 update, the same seller and the same category doubling in a year. Those are not the same product, so comparing them is meaningless.

First Page Sage, whose industry table is the most cited in this category, concedes the point itself: cost per lead "is a valuable but slippery metric because quality varies widely by acquisition cost", and notes that "certain SDRs we've come across are selling cold leads that convert a low percentage of the time."

The usable rule, and it is agency-published rather than independent, is to keep cost per lead under 10% to 20% of your average contract value, then stop tracking cost per lead and start tracking cost per opportunity and cost per closed deal. A $200 meeting is cheap against a $50,000 deal. Against a $4,000 one it is ruinous, not because $200 is 5% of the contract, which passes the rule, but because you pay for every meeting and win on only some of them, and the rule as stated leaves that step out.

Two circulating numbers do not survive checking. The "average B2B cost per lead is $84" figure has an undefined unit blended across every channel. And Belkins' "$10 return for every $1 invested," the most repeated pro-agency claim in the category, is self-reported marketing with no independent methodology behind it. Belkins' own lead generation ROI calculator says its projections "are estimates, not guarantees."

The mechanics behind the first mistake are simple. If you have never run the motion, you cannot write the qualification criteria, so "qualified" gets defined by whoever is paid per unit. That is not dishonesty. It is what happens when the definition is left open.

The conditions have also got harder, which changes what a fair result looks like. No single dataset tracks cold email reply rates over time, so what exists are disconnected snapshots. Backlinko and Pitchbox found in 2019 that 8.5% of 12 million link-building and blogger outreach emails drew a reply overall; in the study's own breakdown a single message converted at roughly 12.5%, a three-message sequence (initial plus two follow-ups) reached about 17%, and one follow-up alone lifted replies by 65.8%. Instantly's 2026 benchmark report, covering 1 January to 18 December 2025, puts the platform-wide average at 3.43% across billions of interactions, with top performers still clearing 10%, and Instantly itself calls that stable rather than falling. Belkins' 2026 study of 7.5 million cold emails sent across its client campaigns in 2025, which excludes open-rate inflation and measures net-new cold outreach only, reports an average reply rate of 0.45%, counting unique replies against every email sent. Both divide by emails sent and they still differ roughly eightfold, which tells you more about how loosely reply rate is defined in this category than about anyone's sending. An agency quoting you 5% is quoting a number no current platform benchmark supports, which means it is either measuring something warmer than cold, or measuring it generously.

Then there is the risk you inherit rather than buy. Since February 2024, Google and Yahoo have required bulk senders, meaning 5,000 or more messages a day, to authenticate with SPF, DKIM and DMARC, to offer one-click unsubscribe, and to keep spam complaints below 0.3%. Google alone recommends staying below 0.1%. Microsoft announced authentication requirements for Outlook.com on 2 April 2025 and began rejecting non-compliant mail outright from 5 May 2025, with no junk-folder grace period. Those rules cover consumer Outlook.com, Hotmail and Live addresses, not Microsoft 365 business mailboxes. At 10,000 emails, roughly 30 complaints breaches the threshold. If the agency sends from domains it owns, that damage stays with them. If it sends from yours, it stays with you.

That mismatch is the quiet reason so many engagements end badly. With sales-qualified lead to closed-won put at around 12% for B2B SaaS by First Page Sage, an agency rather than an independent body, working from its own and its clients' sales data with no sample size published, and multi-month cycles on top, a six-month contract is judged before a single deal could plausibly have closed. Both sides then argue from incomplete data. The two startups Lemkin invested in dropped their agencies at four to six months, which is to say before the answer existed.

Before you attribute a bad result to the agency, check the control you own. Oldroyd, McElheran and Elkington reported two datasets in Harvard Business Review in March 2011. In a study of 1.25 million sales leads at 29 B2C and 13 B2B US companies, firms that tried to make contact within an hour were nearly seven times as likely to qualify the lead, which they defined as a meaningful conversation with a key decision maker, as firms that tried an hour later, and more than sixty times as likely as those who waited 24 hours or more. Their companion audit of 2,241 US companies found only 37% responded within the hour, 23% never responded at all, and among companies that responded within 30 days the average response time was 42 hours. Two cautions. The multiples come from the separate 1.25 million-lead analysis, not from the 2,241-company audit, and the study is from March 2011, before modern outbound and privacy conditions. Nobody independent has replicated the multiple since, and the vendor benchmarks that exist reproduce the direction but not the magnitude. Use it for the direction, not the number.

The gap nobody fills is incrementality. No published methodology separates agency-sourced pipeline from pipeline that would have closed anyway. Until one exists, hold back a control segment yourself, or accept that your ROI figure includes deals you would have won without paying for them.

Those in-house numbers come from The Bridge Group's 2025 SDR report, published 6 February 2025, covering 351 B2B companies with a $47M median revenue, and they are the most independent figures in this entire article. Two of them deserve attention. That 40% breaks down as 16% promotions, 13% involuntary and 11% voluntary, so the largest single share of SDR churn is reps being promoted, not fired. And SDR pay has grown 0.56% compounded annually over the past decade, against CPI growth of 43.9% over fourteen years, which tells you something about how the role is valued, though the report's two windows are not the same length.

The fully-loaded cost of an in-house rep, once you add employer burden, tooling, management, data and ramp, is put at $116,500 to $175,750 by Cold Call Me and at $110,000 to $160,000-plus by SalesHive. Both are lead-generation agencies selling the outsourced alternative, though the salary inputs underneath their estimates are corroborated by independent compensation research. Outsourced dedicated engagements are quoted at $36,000 to $60,000 a year, which is the bottom of the $3,000-to-$15,000 monthly range in the table above. At the top of that range, $180,000 a year, outsourcing costs more than the in-house rep it is meant to replace. Agencies claim 25% to 65% first-year savings, but the cost figures those same firms publish imply 45% to 80%.

The cost comparison is real, though both sides of it are seller-published. The performance comparison is missing, and anyone presenting one has built it from vendor self-reporting.

Here is the limit of all of it: no controlled study has compared whether in-house or outsourced teams book more meetings, or better ones. The cost comparison is real, though both sides of it are seller-published. The meeting-level performance comparison is missing. Anyone presenting one has built it from vendor self-reporting.

Take the ownership point seriously, because it is the one buyers concede without noticing. If the agency registers the sending domains and mailboxes, you inherit the deliverability consequences during the engagement and keep nothing when it ends. Register them yourself.

The compliance exposure works the same way, and it is not theoretical. The UK regulator is explicit that responsibility sits with the sender or the instigator of a direct marketing message, and says it would usually take enforcement action against the instigator, which means you, even when a third party pressed send. In the United States, CAN-SPAM "makes no exception for business-to-business email" and carries penalties up to $53,088 per violating email with no total cap; the record settlement is Verkada's $2.95M in August 2024, for more than 30 million emails that failed to honour opt-outs and lacked a valid postal address. Canada's CASL is stricter still at up to $10M per violation for a company and $1M for an individual, and it has been enforced against a B2B training company, Compu-Finder: the CRTC's first notice of violation proposed $1.1M in March 2015, but on review it imposed $200,000 in October 2017, for 317 non-compliant commercial emails. The scary number circulating in agency blog posts is the proposed one. In Germany, unsolicited B2B email is presumptively unlawful and competitors can sue directly, which is a faster risk than a regulator.

If the agency sends from domains it owns, the damage stays with them. If it sends from yours, it stays with you, and the regulator looks at the instigator either way.

So the contract needs a data-processing addendum and documented list provenance, not because it is good hygiene but because the liability lands on you regardless of who sent the mail.

On selection itself: treat Clutch and G2 scores carefully, since they are agency-solicited. G2 permits incentives up to $100 and labels those reviews; Clutch says it may itself offer incentives, and allows provider-side incentives only with disclosure. Neither publishes how many reviews are incentivised. Rankings do exist, G2's Grid Report for Lead Generation Services and Clutch's Leaders Matrix, but both are computed from reviews the agencies themselves solicited, so they rank solicitation effort as much as delivery. What does not exist is a per-agency satisfaction survey run by anyone independent. The 2023 SaaStr poll rated the category as a whole, not individual agencies, and no one has repeated even that. What you can verify is specific: ask for the definition of qualified, the held-meeting rate rather than the booked rate, the no-show policy, who owns the infrastructure, and what happens on exit.

What does a B2B lead gen agency typically cost in 2026?: Monthly retainers of $3,000 to $15,000 for SDR work and $5,000 to $25,000-plus for full service; percent of sourced pipeline at 10% to 20% of opportunity value; pay-per-meeting of $150 to $600 mainstream and $800 to $2,500-plus enterprise; pay-per-lead $25 to $500; setup $1,500 to $5,000; three to six month minimums. All vendor self-reported. No independent price index exists.
How do I measure ROI from a lead generation agency?: Track cost per opportunity and per closed deal rather than cost per lead, over a window longer than your sales cycle, since First Page Sage puts sales-qualified lead to closed-won at around 12% for B2B SaaS, on its strict definition where an SQL has already met a rep. Control for your own speed-to-lead first: in a 2011 study of 1.25 million leads, sub-one-hour response was roughly seven times likelier to qualify a lead, though a co-author of that study sold lead-response software.
What are the pros and cons of hiring vs in-house lead gen?: In-house gives control and product depth at an agency-estimated $110,000 to $175,750 fully loaded per rep, three months to ramp and 40% median annual attrition. Outsourcing gives a two to six week start and claimed 25% to 65% first-year savings. Practitioners in the SaaStr poll report weaker product knowledge, only 7% of voters said it really worked, and no controlled study has compared the two.
How do I choose the best B2B lead gen agency?: Insist on a written definition of qualified, held-versus-booked meeting terms, your ownership of domains, mailboxes and data, quarterly exit checkpoints, and proof of list provenance. Treat Clutch and G2 scores cautiously, since they are agency-solicited. The G2 and Clutch rankings are built from reviews agencies solicited, and no independent per-agency satisfaction survey exists.
Which industries benefit most from outsourced lead generation?: The evidence here is thin and agency-published. Agencies claim traction in financial services, insurance, professional services and wealth management, where one deal covers many appointments, and I could not find a non-agency source for that claim. Practitioners report the worst results on complex, high-ACV sales generally; the SaaStr commenters describe high-AOV products with deep buyer expectations, not SaaS alone. No independent segment-level study exists.
When does outsourcing actually work?: When you have already proven the motion in-house and want extra capacity, a new region or a new segment, and when your deal size leaves headroom over the cost per meeting. It fails most reliably before product-market fit, on high-ACV complex sales, and in regulated categories where you inherit the compliance exposure.
Who is liable if the agency's emails break the rules?: You are, in most jurisdictions. The UK regulator puts responsibility on the sender or instigator of the message, and CAN-SPAM makes no exception for business-to-business email. So the contract needs a data-processing addendum and documented list provenance, and you should register the sending domains yourself.

SaaStr, Jason Lemkin, "Only 7% of You Have Really Gotten Outsourced SDRs to Work," 16 May 2023, reporting a public LinkedIn poll with 1,209 votes, including comments Lemkin reproduces from the LinkedIn thread by James Rehm, Collin Cadmus and Eric Harrington.
The Bridge Group, 2025 SDR Report, 10th edition, 6 February 2025 (n=351 B2B companies).
First Page Sage, average cost per lead by industry, 2026 report (data January 2022 to June 2025). The $1,357 cost-per-SQL figure for 2024 is attributed to First Page Sage in secondary sources; I could not locate it on First Page Sage's own site.
Belkins, "What are B2B Cold Email Response Rates? (2026 Study)", updated 26 June 2026 (7,530,489 emails, 2025 campaigns); Belkins cost-per-lead figures, published 12 August 2025 and updated 20 March 2026.
Instantly, 2026 cold email benchmark report, measuring 1 January to 18 December 2025; Backlinko and Pitchbox, "We Analyzed 12 Million Outreach Emails", 16 April 2019, link-building and blogger outreach.
WordStream via LeadSpot, 2026 search-ad cost per lead (13,000+ US campaigns).
Pricing ranges published by Cleverly, SalesHive, Bullseye, Toplead, Leads at Scale, ColdCallMe and OutboundSalesPro, retrieved 30 August 2026.
Harvard Business Review, "The Short Life of Online Sales Leads," March 2011 (Oldroyd, McElheran and Elkington). Two datasets: a response-time audit of 2,241 US companies, and a separate study of 1.25 million leads at 29 B2C and 13 B2B US companies. Co-author David Elkington was chairman and CEO of InsideSales.com, which sold the lead-response software the finding implies. HBR is the publisher, not the researcher.
FTC, CAN-SPAM Act Compliance Guide; FTC and DOJ v. Verkada, 30 August 2024.
ICO, business-to-business marketing guidance (PECR and UK GDPR).
CRTC, CASL penalties and enforcement, including Compu-Finder.
Google and Yahoo bulk sender requirements, February 2024; Microsoft Outlook.com sender requirements, announced 2 April 2025, enforced by rejection from 5 May 2025.
Paraphrase Labs AI Visibility Baseline, measured 24 August 2026: 40 B2B buyer questions across six AI engines. On this question, 29 domains were cited; 23 were classifiable from the study's own 206-domain scoreboard and six were cited too rarely to classify; of the 23, 19 are lead-generation agencies, three are software vendors and one is Reddit.

#outbound#lead-generation#b2b#agencies#sdr