Category · B2B outbound

B2B lead generation agency: how to evaluate one

GrowthPal runs outbound for B2B companies that have an offer people already buy and want to find out, quickly, which message moves it. We build lists from dated public signals, write to a hypothesis, and tell you when outbound is the wrong tool. Best fit is past 3M in ARR, a team ready to take calls within a day, and a reachable market of roughly 30,000 contacts.

The buyer

Who you are really selling to

Role	What they care about	Where they pay attention	Best channel
Founder or CEO	Whether outbound can be made to work at all for this offer, before committing headcount. Runway, and the hours they are currently spending writing emails themselves.	Founder peer groups, private Slack and WhatsApp communities, posts from operators they already follow, referrals from other founders. They read an agency's own writing and judge the thinking.	Referral first, then a direct founder to founder email. A message that reads like a sequence is read as one.
CRO or VP Revenue	Pipeline coverage against a number already committed to the board. Predictability over peak output. Whether a vendor solves a reporting problem or creates one.	Board and investor networks, peer CROs, and their own AEs complaining about lead quality.	Warm introduction through an investor or peer. Cold email works when it references a public pipeline or hiring signal rather than a title.
VP Sales or Head of Sales	Rep capacity, and whether the agency sounds like their company in the inbox. Under capacity makes outbound a staffing question; at capacity makes it a nonquestion.	Sales communities, their own SDR managers, vendor demos.	Email and LinkedIn together, with specificity about their motion. They respond to control, not to volume.
Head of Growth or Demand Generation	Channel mix and attribution. Outbound is one line in a plan that already holds paid, content and events, so they want to know what it tells them that the other lines do not.	Marketing communities, newsletters, conferences, tooling vendors' content.	Content first, then email. They will read a long page end to end before booking anything.
Head of Partnerships or Business Development	Access to named accounts rather than volume. A short list of logos often matters more than the total addressable market.	Industry bodies, events, trade press, personal network.	Referral and direct personal email. They get one first touch per target account and know it.

Constraints

What makes outbound different here

Inbox access is a technical gate before it is a copy problem

Mailbox providers publish enforceable requirements. Google's bulk sender requirements, which apply to senders of more than 5,000 messages per day to Gmail accounts, call for SPF and DKIM authentication, DMARC on the sending domain, valid forward and reverse DNS, TLS in transit, one-click unsubscribe for marketing mail, and a spam rate in Postmaster Tools kept below 0.30 percent, with the same guidance separately recommending below 0.10 percent.

What it decides: Pacing and channel. Volume per domain is capped by sender reputation rather than by ambition, warmup becomes a prerequisite instead of a formality, and a complaint-generating message is now expensive, because it degrades the asset that has to deliver every future message.

The qualified market is often small enough to exhaust

Plenty of B2B offers have a few thousand genuinely qualified accounts worldwide, not a few hundred thousand. Volume-led outbound burns that universe in weeks and leaves nothing to return to in the next quarter.

What it decides: The list. Cap the universe deliberately and split it into waves instead of sending to everything at once, so the same account can be approached from a new angle later without feeling harassed. A small market makes the program a testing program by necessity.

You are writing to a buying committee, not a buyer

In most B2B purchases more than one person has to agree, and the person who replies is rarely the person who signs. A campaign aimed at one title produces replies that stall politely.

What it decides: The list schema and the copy. Build at the account level with several roles mapped, write a different message per role because the economic buyer and the operational owner have different problems, and space touches across the committee so one account does not get near-identical emails in a week.

The sales cycle is longer than the first review point

When a deal takes months to close, the only fast signals available are reply behavior and meeting counts, and neither predicts revenue on its own. Any agency, GrowthPal included, that promises a revenue outcome inside a window shorter than your own sales cycle is describing something other than revenue.

What it decides: The angle and the review. Each campaign is built as a falsifiable hypothesis so a negative result still teaches something, and the first review asks what was learned about the message rather than how many meetings landed.

Compliance is per jurisdiction and the list is usually mixed

One list routinely carries US, UK, EU, Canadian and Australian contacts under five different rule sets. The US works on an opt-out regime. The others work on consent models with narrow exceptions, which is why a US playbook does not travel.

What it decides: Jurisdiction segmentation happens before the list is built, not after, because retrofitting it means suppressing contacts you have already emailed. Some contact types, such as sole traders in the UK, get a different channel rather than a cold email.

Signals

Buying triggers, and the angle each one earns

Trigger	How we find it	The angle it earns	Proof the email needs
Pipeline shortfall discussed publicly	Earnings calls and filings where management talks about pipeline coverage, bookings softness, or a change in go-to-market investment. For private companies, a sudden change in hiring posture or a public comment from an executive.	A coverage gap this quarter, and outbound as the way to add top of funnel without adding headcount inside the quarter.	A specific reference to the stated shortfall and one concrete first move. Not a claim about results.
Repeated SDR postings that are not getting filled	The same SDR or BDR role posted again on a careers page or job board over several months, plus short tenures visible on profiles at that company. The Bridge Group's 2025 SDR report, covering 351 B2B companies, puts average SDR tenure at 1.9 years, median 2024 annual attrition at 40 percent with a 21 to 57 percent interquartile range, and quota attainment at 60 percent.	SDR attrition and SDR ramp mean the team is rarely at full strength, and an agency absorbs that cycle instead of competing with it.	The specific repeated posting, dated. The research is context for the decision, not a line in the email.
A new segment or product line with no referral base	A new page on their site, a changed homepage headline, a new pricing tier, a first job posting for a vertical they did not serve before, or a conference where they are newly exhibiting.	The existing team has no inbound and no referrals in the new segment, which is exactly where outbound has the highest marginal value.	Evidence of having read the new page, plus a named hypothesis about who buys it that they can agree or disagree with in one word.
A new revenue leader in their first quarter	Role changes on LinkedIn, company announcements, trade press. A new CRO or VP Sales reviews motion and vendors early.	They own a number they did not set and need pipeline that does not depend on the team they inherited.	Something specific from their prior role, and a view on what usually breaks in the first ninety days.
A failed agency engagement	Hard to target directly, which is worth saying. Weak public signals: a role posting that says rebuilding outbound or bringing outbound in house, or a client logo that has quietly left an agency site. Usually this surfaces in conversation.	Diagnosis before delivery, including the possibility that the offer rather than the outbound was the problem.	A specific, falsifiable claim about what usually goes wrong, and a request to read the previous sequences.
Funding round or acquisition	Registry and regulatory filings such as US SEC Form D and UK Companies House share allotments, press releases, investor portfolio pages.	Use it as a timing filter, never as the angle. A newly funded company receives a flood of identical congratulation emails in the same fortnight. Build the angle from a commitment made at the time of the raise, such as a new market or a hiring plan.	The stated plan, referenced precisely. Never the amount.
Expansion into a new geography	A new office address in registry filings, a country subsidiary, local-language pages, job postings located in a new country, membership of a local trade body.	No local network, no referral base and no inbound, so outbound is the only channel that starts from zero and still works.	The registry or posting evidence, plus awareness of that country's consent rules, which doubles as evidence of competence.
Visible strain in the buying function	Job postings that bundle several roles into one, a careers page frozen for a year alongside growth signals, public reviews about response times, an executive describing a bottleneck.	The constraint is throughput rather than demand, so more meetings would make it worse. This supports an honest conversation about waiting, which converts later more often than people expect.	The specific posting or public statement, and a reason to wait rather than buy.

Peek inside

Example angles we would test

Examples of angles to race in this market, not results. Each one is a hypothesis with its proof assembled before a single send.

Your AEs are prospecting instead of closing, because nothing reliable arrives at the top of the funnel.

Proof it needs: Repeated SDR postings, or a sales team visibly smaller than the revenue target implies.
The ask: A short call comparing current pipeline coverage against the number.

You launched into a segment where nobody knows you, and the motion that built the core business does not transfer.

Proof it needs: A dated new product page, a new pricing tier, or a first hire for that segment.
The ask: A written view on who the first fifty accounts should be, sent before any call.

The last outbound program failed and nobody established whether it was the list, the offer or the copy.

Proof it needs: A named common failure mode, plus a request to read what was previously sent.
The ask: Send the old sequences for a written read. No call required.

Deals stall because only one person at the account ever engaged, and that person cannot sign.

Proof it needs: Committee structure observable at comparable accounts, read from their own job postings and team pages.
The ask: A short conversation about which roles actually sign in their deals.

You are ninety days into owning a number you did not set, with a team you did not hire.

Proof it needs: The dated role change plus something specific from their prior role.
The ask: Fifteen minutes on what they plan to change first.

The new territory has no inbound and no referrals, so every other channel starts from zero.

Proof it needs: A registry filing, a local-language page, or a job posting located in that country.
The ask: A short discussion of the local list and the local consent rules.

Your prospects get several near-identical emails a week, because every competitor pulls the same filter-based list from the same database.

Proof it needs: A description of the exact filter set that produces that list, which the reader will recognize because they use it.
The ask: A conversation about building from signals instead of filters.

Sending volume has outgrown sender reputation, and the resulting drop is being read as a copy problem.

Proof it needs: The published mailbox provider requirements, and an offer to check their current authentication setup.
The ask: Permission to run a technical check on their sending domains.

Your constraint is delivery capacity, not demand, so more meetings would make this quarter worse.

Proof it needs: Public evidence of the bottleneck, such as a posting that bundles three roles into one.
The ask: A short call to decide whether to do nothing for now, with that outcome explicitly on the table.

Your contract with the incumbent is approaching renewal and nobody has looked at alternatives yet.

Proof it needs: A public contract award record with a stated term, or a published renewal date.
The ask: A conversation timed to the renewal window rather than to this week.

The list

How the list gets built in this market

Lists get built from dated public signals and then verified, with vendor data used for coverage rather than for the reason to write. If a competitor could reproduce your list in ten minutes with the same three filters, it is a queue rather than a list.

Registries and authoritative databases

Companies House in the UK, for legal entity, status, officers, registered address, filing dates and share allotments.
SEC EDGAR in the US, including Form D filings, and the national company registry equivalent in each target country.
Industry licensing and regulatory registers, which is where a sector page does its real work. Wattstor's lists were built from UK public sources including the ESOS register, which is the pattern rather than the exception.
Public procurement and tender portals, where contract awards show who buys what, from whom, and when the term ends.
Trade body and membership directories, a strong proxy for a company operating in a category and choosing to be visible in it.
Patent, trademark and standards filings, slow signals but reliable for product direction.

Signal sources that carry the angle

Job postings on careers pages and boards, the single most useful public signal in B2B. They reveal team structure, named tooling, new functions, geography and strain.
The company's own site over time, including archived versions: new pages, changed positioning, new tiers, removed products.
The public technology footprint observable from the site itself.
Executive and company posts on LinkedIn and in public communities, where priorities are stated in their own words.
Press, podcasts and conference agendas, where a named person says what they are working on.
Review sites and public support forums, where customers say what is not working.

Where vendor data stops being enough

Adequate for firmographic scoping, a first draft of the universe, and finding which roles exist at an account.
Not adequate for anything that forms the angle. A database records what was true when the record was written.
It does not know the role was posted twice this year, the product page changed last month, or the person started six weeks ago.
Every competitor buys from the same few providers with the same filters, which is why target accounts report near-identical emails.
Rule of thumb: vendor data for coverage, public sources for the reason to write.

How a row gets verified

Confirm the company still exists and still trades, from the registry rather than from the database.
Confirm the person still holds the role, from a source dated more recently than the database record.
Verify the address technically and drop catch-all and risky results rather than sending to them.
Run CRM suppression against the client's own systems, covering existing customers, open opportunities, partners and past opt-outs. Not just the agency's list.
Confirm the jurisdiction, apply that rule set, and confirm the dated signal that justifies the angle. If the signal cannot be cited in one sentence, the row comes out.

What makes a bad list here

Built from filters alone, with no signal attached to any row.
Bought or scraped in bulk, with no verification pass and no jurisdiction segmentation.
Contacts with no reason: the row cannot answer why this person, this month.
Titles targeted instead of responsibilities, so it catches people who match the words but not the job.
No suppression against existing relationships, so customers and open deals get cold emailed.
Stale by construction, with no record of when each field was last confirmed.
Reproducible by any competitor with the same tool and the same three filters.
Large because large was the goal. Size is the output of a definition, not an input to one.

Rules

Compliance and trust constraints

Nothing here is legal advice. Rules differ by country, change over time, and turn on facts, so read each source directly and take advice from a qualified lawyer in the relevant country before running a program.

In the US, CAN-SPAM makes no exception for business email

The FTC's compliance guide states the law covers all commercial email, including business to business. It is an opt-out regime: prior consent is not required, but the message needs accurate header information, a subject line that reflects the content, a disclosure that it is an advertisement, a valid physical postal address, and a working way to opt out, honored within 10 business days and functional for at least 30 days. Penalties apply per email in violation. The guide also states that hiring another company to send on your behalf does not remove your own responsibility, which lands directly on any agency relationship.

Read the source

In the UK, a sole trader is not a corporate subscriber

The ICO separates corporate subscribers, meaning companies, limited liability partnerships, Scottish partnerships and government bodies, from individual subscribers, who need consent unless the soft opt-in applies to an existing customer. Sole traders and some partnerships count as individuals. You may email corporate bodies, and the ICO recommends keeping a do-not-email list of businesses that object and screening new lists against it. Data protection rules still apply to a named individual's work address.

Read the source

Canada and Australia run on consent, and publication alone is not it

We could not load the CRTC or ACMA websites, so this comes from the statutes themselves, read in full on the official government legislation registries. Canada's anti-spam law prohibits sending a commercial electronic message without consent, express or implied, and requires sender identification, working contact details and an unsubscribe mechanism. Implied consent is a closed list: the conspicuous publication route needs the address published, no refusal notice attached to that publication, and relevance to the person's business role. Australia's Spam Act defines consent as express or reasonably inferred from conduct and business relationships, says consent may not be inferred from the mere fact an address was published, requires the message to identify who authorized it with contact details likely to stay valid at least 30 days, and requires a clear, functional unsubscribe address valid for the same period. It also prohibits address-harvesting software and lists produced with it. Check your own position with the CRTC and the ACMA directly.

Read the source

Mailbox provider rules decide whether any of it is read

Google's sender guidelines apply above a daily volume threshold to Gmail accounts and have been enforced since February 2024. They require SPF, DKIM and DMARC, correct forward and reverse DNS, TLS, standards-compliant formatting, one-click unsubscribe for marketing mail, and a low spam rate. These are not law, and they bind harder than law does in practice, because a domain that fails them is not read whatever it says.

Read the source

The actual document

The buyer's checklist: ten questions to ask any outbound agency

Take these to every agency you are evaluating, including this one. Each question is paired with what a good answer sounds like and what a weak one sounds like, because the questions are easy to find and the scoring is the part nobody publishes. They are deliberately uncomfortable for an agency that sells on volume.

Field	What goes in it
Who specifically will write our copy and build our list, and can I meet them now?	Good: a named person, introduced before you sign, who can describe their own process in their own words. Weak: our team, or a promise to introduce you once the contract is signed.
How many clients does that person handle at once?	Good: a specific number, with an explanation of what it means for the attention your account gets. Weak: evasion, or a number so large that nothing can be hand-built.
Show me how you would build our list. Which public sources, which signals?	Good: named registries, named signal sources, and a reason attached to every row. Weak: a database name and three filters. If a competitor could reproduce it in ten minutes, so is your list.
What is your written qualified meeting definition, and who decides?	Good: agreed in writing before launch, with a dispute process and a stated no-show and reschedule policy. Weak: we will know it when we see it, or a definition that counts a booking whether or not it happens.
What are the contract term, the notice period and the exit terms?	Good: a term matched to a realistic learning cycle, with a defined review gate and a clean notice period. Weak: a long minimum with no review gate, or an auto-renewal you have to remember to catch.
Which domains will you send from, and how is my primary domain protected?	Good: separate sending domains, authentication and warming described concretely, and explicit protection of the corporate domain. Weak: sending from your main domain, or not knowing why that matters.
How do you handle compliance across jurisdictions, and who carries the risk?	Good: jurisdiction segmentation before the list is built, awareness that the US is opt-out while the UK, EU, Canada and Australia are consent-based, and a plain statement that you retain liability. Weak: we are fully compliant, with no detail.
How do you suppress our customers, open opportunities and partners?	Good: CRM suppression that reads from your systems, runs before every send, and covers past opt-outs and do-not-contact requests. Weak: a one-time list they ask you to email over at kickoff.
What will you test first, and what result would tell you the hypothesis is wrong?	Good: a named hypothesis and a stated condition for abandoning it. This question separates testing-led from volume-led outbound better than any other. Weak: a sending volume, a number of sequences, or a target meeting count.
Under what circumstances would you tell us not to do this?	Good: specific conditions, offered without hesitation, including that the offer may not be ready or the market may be too small. Weak: outbound works for everyone, or a reframe of the question.

We publish this knowing buyers will take it to our competitors and to us. An agency that would rather you did not have these questions is telling you something, and so is one that answers all ten without being asked twice.

Process

How the engagement runs

Three stages and nine steps, the same ones in every market: build the list from public sources, race message angles against it, scale the one that wins. The nine steps from kickoff to booked meetings are written out in full.

What changes for this market

Week one is an evaluation, in both directions. Before any list exists we agree what counts as a result, in writing, and we say whether we think outbound is the right tool at all. If the answer is wait, we say wait, and that conversation is free.
List building starts from registries and dated public signals, with vendor data used only for coverage. Every row carries the signal and the public URL it was read from, and a row that cannot answer why this person, this month, does not get sent.
The list is split by jurisdiction before it is built, because the US is opt-out and the UK, EU, Canada and Australia are not. Suppression runs against your CRM, not ours, and it runs before every send.
The send calendar is paced to sender reputation rather than to a volume target, from separate sending domains with no open or click tracking. The first review asks what was learned about the message, and every campaign carries a written condition that would tell us the hypothesis is wrong.

Proof

What GrowthPal has published

What was published	What the figure counts	Source
4M+ cold emails sent for 40+ B2B companies	Emails sent, cumulative across all clients. Not one campaign.	Read it
449 interested replies from 67,934 contacts reached in one month, about 1 per 150, with 955 replies in total	Contacts reached, in one month.	Read it
83.47 percent positive-reply rate: 202 positive replies from 4,303 contacts	Share of replies that were positive, not the share of contacts. State this whenever the figure is used.	Read it
2,000+ qualified leads in 9 months	Qualified leads. The contact denominator for the full 9 months is not published.	Read it
A meeting booked with a billion-dollar company on the first day of the campaign, with lists built from UK public sources including the ESOS register	Not a rate. A single named event.	Read it
A study of 412 campaigns across 47 companies over 18 months: campaigns scoring above 0.7 on the composite message-market fit score averaged an 11.4 percent reply rate, against 1.1 percent for campaigns below 0.3. Volume, warm-up and time of day explained less than 8 percent of the variance.	Total reply rate, measured against messages in those campaigns. This is a different measure from interested replies per contact and the two must never be compared on the same page.	Read it

How these are counted

An interested reply means a human reply expressing interest in continuing the conversation. Out-of-office messages, unsubscribes, referrals to someone else and automated responses are not counted. A figure measured against contacts reached and a figure measured against emails sent are different measures: the column above says which each one is, and we do not convert between them or add them together.

The full counting method, and what an interested reply is.

Honestly

Who this is not for

Before you know who the offer is for. Outbound finds people who already have a problem you solve. It does not tell you what to build or who to sell to. If the offer moves every month, testing angles against it measures the wrong variable and burns the list while doing it.
A market smaller than roughly 30,000 reachable contacts. Testing needs enough volume to separate a message that works from one that got lucky. Under that size outbound can still work, but a small, hand-run, relationship-led approach usually beats a system, and you should not pay for a system.
Strictly tender-based buying. Where purchases run through formal procurement and framework agreements, the useful move is getting onto a supplier list and timing a renewal window. A sequence aimed at a calendar booking has nowhere to go, however good the copy is.
When delivery capacity is already the bottleneck. If the team cannot service the work it has, more meetings make this quarter worse, not better. This is the most common reason to say wait. It is also the one buyers argue with hardest, usually a quarter before agreeing.
No one ready to take the calls. Replies arrive within days of the first sends. If nobody owns follow-up inside a day, interested buyers cool off and the whole spend is wasted on timing rather than on message.

Questions

Answered straight

Is cold email dead?

Untargeted cold email mostly is, and the evidence is published by the mailbox providers rather than by anyone selling a cure. Google's bulk sender requirements apply above 5,000 messages a day to Gmail accounts and require authentication, one-click unsubscribe and a spam rate below 0.30 percent, with below 0.10 percent recommended. Volume that generates complaints now degrades the asset that delivers every future message. If an agency's answer to falling response is to send more, that answer is the problem.

Should we hire an agency, buy software, or build an in-house SDR team?

Software sends what you already know works; it does not discover the message. An in-house SDR team builds knowledge you keep, but ramps slowly and carries real attrition risk: the Bridge Group's 2025 report across 351 B2B companies puts median annual attrition at 40 percent and quota attainment at 60 percent. An agency reaches a tested message fastest and absorbs the hiring cycle, but you rent it unless the list definitions, sequences and learnings are handed over in writing. Agency first, in-house later, is a common path.

How do the engagement models differ?

Three shapes exist. A retainer buys a team and a process for a period, covering list building and infrastructure whether or not meetings land this month. Pay per meeting bills on booked appointments, which looks aligned and quietly rewards volume unless the qualified meeting definition is strict, written, and covers no-shows and reschedules. Pay per lead bills on a delivered contact and is the weakest alignment of the three. In all of them the decisive term is what counts, agreed in writing before launch.

We tried an agency before and it failed. Why would this be different?

Start by asking what was actually tested. Most failed engagements never isolated a variable: the list, the offer and the copy all changed at once, so nothing was learned and the next agency inherits no information. Before signing again, ask the new agency to read the previous sequences and the previous list definition and say what they think went wrong. Diagnosing without seeing either is the same failure about to repeat. Sometimes the honest answer is that the offer was the problem.

When should a buyer expect the first honest progress report?

Run two clocks. Campaigns are sending two to four weeks after kickoff, with warming the infrastructure inside that window, and a winning message is usually identified around week eight. That is message signal: which hypothesis gets engagement, which segments respond, which do not. Revenue signal cannot arrive faster than your own sales cycle, so a promise to show revenue impact inside a shorter window is describing something else. Agree what the first review examines before launch.

Will this damage our brand or our sending domain?

It can, which is why the question is worth asking in detail. Sending happens from separate domains, never your primary corporate domain, with authentication in place and no open or click tracking. Copy goes through your approval before anything sends. Existing customers, open opportunities and partners are suppressed from your CRM rather than from a list someone pasted at kickoff. Ask what happens when a target account asks never to be contacted again, and expect a concrete process.

Does hiring an agency reduce our compliance exposure?

No. The FTC's guide states plainly that hiring another company to handle email does not remove your own legal responsibility, and the consent regimes elsewhere work the same way: the business being promoted stays on the hook. That is a reason to ask an agency how it handles jurisdiction segmentation, not a reason to avoid agencies. An agency that cannot describe how it splits a mixed list before building it is describing a risk you will carry.

What would make you tell us not to do this?

A market too small to test in, an offer still moving month to month, delivery capacity already at its limit, or buying that runs strictly through tender. We say so before taking money, because taking it anyway produces a failed engagement that costs both sides more than the fee. Ask any agency you are evaluating for the conditions under which they would decline the work, and treat hesitation as the answer.

Who you are really selling to

What makes outbound different here

Inbox access is a technical gate before it is a copy problem

The qualified market is often small enough to exhaust

You are writing to a buying committee, not a buyer

The sales cycle is longer than the first review point

Compliance is per jurisdiction and the list is usually mixed

Buying triggers, and the angle each one earns

Example angles we would test

How the list gets built in this market

Registries and authoritative databases

Signal sources that carry the angle

Where vendor data stops being enough

How a row gets verified

What makes a bad list here

Compliance and trust constraints

In the US, CAN-SPAM makes no exception for business email

In the UK, a sole trader is not a corporate subscriber

Canada and Australia run on consent, and publication alone is not it

Mailbox provider rules decide whether any of it is read

The buyer's checklist: ten questions to ask any outbound agency

How the engagement runs

What changes for this market

What GrowthPal has published

How these are counted

Who this is not for

Answered straight

Related reading