Demand generation is the least standardised term in business software marketing. To one agency it means paid acquisition with a lead target, to another it means a content and community programme measured in pipeline over quarters, and to a third it means the whole revenue marketing function outsourced. All three will answer the same brief, quote wildly different numbers, and describe their work in language that sounds identical in a proposal. This page defines the parts, explains why the measurement choice decides everything else, and sets out what to establish before you sign, so that month four is a conversation about performance rather than about what you thought you were buying.
The three things the term is used to describe
First, capture: converting demand that already exists. Paid search, review site presence, comparison and alternative pages, and a website that does not lose people who arrived ready. This produces results fastest and has a ceiling set by the size of your category's existing search demand. Second, creation: making buyers aware of a problem they were not searching to solve, through content, events, podcasts, community and paid social. It works slowly, is hard to attribute cleanly and is the only route in a category where nobody is searching yet. Third, operations: routing, scoring, nurture and the reporting that connects marketing activity to closed revenue. Most disappointing engagements are a capture programme bought by a company that needed creation, or the reverse, and the mismatch is visible in the first proposal if you ask which of the three is being sold.
Why the measure you pick decides the programme
Agree what you are counting before you agree what you are paying, because the metric shapes every decision that follows. A programme judged on lead volume will produce lead volume, including from job seekers, students, competitors and people who wanted the ebook. A programme judged on qualified pipeline behaves completely differently, and costs more per unit while usually costing less per closed deal. A programme judged on closed revenue needs a sales cycle to elapse before anyone can be fairly assessed, which is unworkable as a monthly management measure but essential as an annual one. Pick a primary measure that reflects your commercial goal, agree a secondary leading indicator that moves faster so nobody is flying blind for two quarters, and write both into the reporting section of the contract rather than leaving them to be negotiated in a review.
What the organic half actually requires
The content engine underneath most successful software demand programmes is not a blog calendar, it is a small set of high intent pages plus genuine depth on the problem you solve. Google's guidance on creating helpful, reliable, people first content describes what it tries to reward: material made for people, showing first hand experience or real depth of knowledge, that leaves the reader feeling their question was answered. In software that usually means material written with somebody who has actually implemented the thing, which costs more per piece and outperforms volume by a wide margin. It also means integration, migration, comparison and pricing pages that most companies leave until last, despite them attracting the readers closest to a decision. Ask any candidate how many pieces per month involve an interview with a practitioner, and the answer will tell you which kind of content programme you are buying.
Attribution honesty, and what to require
Software buying involves several people over months, most of the research is invisible to your analytics, and no attribution model resolves that. What a serious provider does is state the limitation openly and design around it: a self reported source field on the form, tracked spend against pipeline created in cohorts by month, and a small number of channel experiments with clean holdouts. What a weaker provider does is present a last click dashboard where the channel they manage takes credit for demand created elsewhere. Ask a candidate directly how they would report a quarter in which brand search rose, paid held flat and pipeline grew. The answer separates the people who understand this problem from the people who will send you a colourful report. Companies comparing business to business demand generation services generally find this one question does more sorting than any other.
Questions people ask about saas demand generation
How much should an early stage company spend?
Less than most proposals suggest, and concentrated. Before product market fit is clear, a broad programme spreads a small budget across channels that each need scale to work. Capture the demand that already exists, learn what converts and who buys, and expand once the answer is repeatable. Agencies that will take a small budget and focus it on one channel are usually more useful at that stage than full service ones.
What is a fair timeframe before judging results?
Match it to your sales cycle plus the time the channel needs to work. Paid capture gives a readable signal within weeks. Content and community programmes need at least two quarters before the pipeline they create becomes visible, and longer if your average deal takes months to close. Agree a leading indicator that moves sooner so the intervening period is not blind.
Should an agency own the whole function?
It can work for a company with no marketing team, provided one person internally owns the relationship and the measures. It works badly as a permanent arrangement, because category knowledge accumulates outside your company. The usual pattern is an agency building the engine while an internal hire is recruited to run it.
How do I compare two very different quotes?
Normalise them onto the three parts: capture, creation and operations. Ask each provider what proportion of the fee goes to each, who does the work, and what the deliverables are per month. Two quotes that differ by a wide multiple almost always differ in scope rather than in rate, and the split makes that visible immediately.