“High-quality content that drove 1M+ views and real leads.” - Marketing Manager, No Stress (Pulsetto)      “High-quality content that drove 1M+ views and real leads.” - Marketing Manager, No Stress (Pulsetto)      
    Demand GenerationJuly 18, 2026Earworm

    How to Choose a B2B Demand Generation Agency (Questions to Ask)

    Every agency now calls itself demand generation. Most sell MQLs. Ten questions, three archetypes and the red flags that reveal what you are actually buying.

    Two people in conversation across a table, representing the questions to ask before hiring a B2B demand generation agency.

    Hiring a demand generation agency looks straightforward until you take three sales calls and notice that the same two words mean something different in each one. The first agency pitches paid social and a landing page. The second pitches a squad of outsourced SDRs. The third pitches a content engine. All of them call it demand generation, all of them have a deck, and only one of them is selling what you think you are buying.

    This guide is for the person sitting through those calls. It covers what demand generation actually means (and the lead generation trap wearing its name badge), the three agency archetypes you will meet, ten questions that reveal which one you are talking to, and the red flags that should end a meeting early. Full disclosure before we start: Earworm is a demand generation agency, so we have a horse in this race. We will point out which one when we get there.

    3
    agency archetypes sold under one label
    10
    questions that tell them apart
    1 quarter
    the minimum fair pilot

    Demand Generation vs Lead Generation: The MQL-Mill Trap

    Lead generation collects contact details. Demand generation creates the preference to buy from you specifically, then captures it when the buyer is ready. These are different jobs, and most agency selection mistakes come from vendors doing the first while charging for the second.

    The mechanics of the trap are familiar. An agency gates a PDF, runs ads at it, and every form fill becomes a "marketing qualified lead". The dashboard fills up, the monthly report looks vigorous, and your sales team quietly stops calling the list, because the people on it downloaded a checklist eight weeks ago and have not thought about you since. Marketing hits its number. Revenue does not move. Eventually someone senior asks why, and the answer is that you paid for lead generation and booked it as demand.

    The failure is structural, not a matter of execution. At any given moment the overwhelming majority of your market is not in a buying cycle; the Ehrenberg-Bass Institute's much-quoted 95:5 heuristic exists for a reason. A form fill does not move anyone into market. What moves the needle is being the company a buyer already trusts, understands and prefers when budget appears. That is a memory-building job, which makes it a content and distribution job, and it runs on a longer clock than a campaign. Any agency promising demand on a lead generation timeline is quietly redefining one word or the other.

    The Three Agency Archetypes

    Nearly every vendor using the label fits one of three shapes. None is dishonest by default; the trouble starts when you buy one believing it is another.

    ArchetypeWhat you are actually buyingWhere it worksWhere it breaks
    Ads-and-landing-pages shopMedia buying, creative variants, conversion pagesCapturing existing demand fast; clear offers; short cyclesPlateaus with spend; nothing compounds; the audience is rented, never owned
    SDR outsourcerOutbound sequences and booked meetingsTesting a new segment quickly; founders who need calendar volume nowCold outreach into a market that has never heard of you; brand damage at scale
    Content-flywheel operatorFlagship content, distribution, and the demand it accruesConsidered purchases, long cycles, markets bought on trustSlower to first result; needs genuine expertise to feed it

    Declared bias: Earworm is the third type. Our demand generation service runs on flagship video content, podcasts and series that earn attention in their own right, rather than ads pointed at landing pages. The same engine does the job of a content marketing agency, because the flywheel produces the content as a by-product of doing its main work. The honest corollary: we are the wrong choice for some buyers. If you need forty meetings booked by March, an outbound shop will serve you better, and we will say so on the first call.

    The structural difference shows up over time. Media-led programmes track spend: turn the budget off and pipeline follows it down within a quarter. Content-led programmes track time: the library, the audience and the trust are still working next year, whether or not you add another pound.

    What compounding looks like
    Q1Q2Q3Q4020406080
    Illustrative pipeline contribution, indexed, from a content-led demand programme. Not real client data. Media-led programmes tend to track spend rather than time.

    Ten Questions to Ask Before You Sign

    Ask all ten. The answers matter, and the speed and comfort of the answers matter almost as much.

    1. Where, precisely, do the leads come from?

    Not "our proprietary methodology". The actual mechanics: which channels, which audiences, what a buyer saw before their details reached your CRM. An agency doing real work can walk you through one lead's journey end to end. An agency that cannot is either buying lists or does not know, and neither is charming.

    2. Who owns the content and the audience?

    If the agency produces content, confirm in writing that you own it: the episodes, the clips, the ad accounts, the subscriber list, the audience data. Some agencies keep the audience inside their own accounts, which converts your marketing budget into their asset and makes leaving expensive by design.

    3. What is the channel mix, and why?

    The right answer connects channels to your buyer, not to the agency's staffing model. Paid distribution has a genuine role here; a good paid social agency puts proven content in front of named accounts rather than spraying gated PDFs at cold audiences. But if every road leads back to the one channel the agency happens to sell, you have found a hammer describing you as a nail.

    4. How honest is your pipeline attribution?

    The wrong answer is a dashboard claiming credit for every deal that ever scrolled past an ad. The right answer is an adult conversation about what can be measured directly, what surfaces in self-reported attribution ("heard you on the podcast"), and what stays genuinely unknowable. Ask for case studies that talk about pipeline and revenue rather than impressions. An agency that overstates certainty in the sales process will overstate results in the reporting.

    5. What is the fee model?

    Flat retainers align the agency with outcomes. Percentage-of-spend models pay the agency more for spending more of your money, which is a strange incentive to fund. Per-lead pricing pays for volume, and you have read the section above. No model is perfect, but the agency should be able to defend its incentives without flinching.

    6. What are the contract terms?

    Twelve-month lock-ins with auto-renewal deserve scrutiny. Demand programmes do need patience, but confident agencies keep clients through results, not clauses. A quarterly commitment with a sensible notice period is a reasonable middle. Whatever the agency refuses to shorten after the pilot tells you what they expect the results to do.

    7. Who actually does the work?

    The people in the pitch are rarely the people in the delivery. Ask who runs your account day to day, how senior they are, how many other accounts they carry, and what is subcontracted. A strategist spread across twelve accounts is a template with a calendar.

    8. What does reporting look like, and how often?

    Monthly is standard; the content is what matters. Ask to see a real, anonymised client report. It should lead with pipeline and named-account movement, not impressions, and it should include what underperformed. A report containing no bad news is not a report. It is an invoice cover letter.

    9. What happens when it is not working?

    Every programme has flat quarters. The tell is whether the agency has a rehearsed answer: which leading indicators they watch, at what point they change course, and what they have killed for other clients. An agency that has never stopped doing something has never measured anything.

    10. What happens when we leave?

    Offboarding reveals character. Confirm the hand-back of accounts, assets, audiences and documentation, with timelines, before you sign. The agencies that make leaving easy tend to be the ones you never need to leave. The ones that go vague are telling you where the exits are bricked up.

    A lead is a contact detail. Demand is a reason to call you. Only one of them shows up in revenue.

    — Earworm

    Red Flags That Should End the Meeting

    • Guaranteed MQL volumes. Nobody can guarantee qualified interest, only form fills, and those are a different product.
    • Black-box audiences. "Proprietary intent data" that cannot be explained in plain English is usually a lookalike audience with a marketing budget of its own.
    • Rented email lists. Cold-mailing a purchased list is not demand generation, and in the UK it is also a conversation with the ICO waiting to happen.
    • Case studies without revenue. Impressions, clicks and "engagement" as headline results mean the pipeline numbers did not survive contact with the client's CRM.
    • Results promised inside a month. Demand compounds. Anything instant is capture, arbitrage or fiction.

    Running a Fair Pilot

    A pilot protects both sides, provided it is designed for the thing being tested. Judge a demand programme like a direct-response campaign and it will fail on schedule. Judge it like an investment and you will actually learn something.

    1. Scope one segment. One audience, one flagship content stream, one distribution plan. Small enough to run properly, large enough to matter.
    2. Agree leading indicators upfront. Pipeline is a lagging measure. In a first quarter, watch the quality of engaged accounts, self-reported attribution in inbound, and whether sales conversations start warmer than they used to.
    3. Give it a quarter, minimum. Two is fairer for content-led work. Week-six verdicts select for agencies that have learned to game week six.
    4. Define success and kill criteria in writing. Both sides should know, before anything starts, what "continue", "adjust" and "stop" look like.

    The pilot's most useful output is not the numbers. It is watching how the agency behaves while the numbers are still early and ambiguous, because that behaviour is what you are actually hiring.

    Choose the Engine, Not the Dashboard

    The agencies worth hiring will take all ten questions without discomfort, show you pipeline rather than impressions, and tell you plainly whether they are the right archetype for your situation. If a content-led engine is the shape you are after, with flagship video doing the demand work and paid distribution amplifying what already earns attention, that is precisely what our demand generation agency builds. Book a call and we will answer the ten questions ourselves, in order.