Most demand gen playbooks were written for companies that already had a marketing function. They assume budget. They assume some baseline brand awareness. They assume an existing pipeline you're trying to grow.
None of that applies when you're starting from scratch. You don't have budget to test channels. You don't have brand awareness. You don't have pipeline to optimize. You have a product and a hypothesis about who needs it.
Here's what actually works in that situation.
Phase 1: The first 90 days are about evidence
Before you spend any meaningful money on demand gen, you need three things you almost certainly don't have:
- A repeatable description of your ideal customer that you've validated with actual conversations
- Specific language those customers use to describe their problem (their words, not yours)
- A clear understanding of how those customers currently solve the problem (your real competition)
None of this comes from a workshop. It comes from 30 customer conversations in 90 days. Past customers, current prospects, people who fit the ICP but haven't bought. Open-ended questions, not multiple choice.
Skip this phase and everything downstream is theater.
Phase 2: One channel, one motion, six weeks
The instinct when starting demand gen is to pick three channels and run small experiments on each. This sounds smart. It's a trap.
Three small experiments produce three weak signals. You can't tell which channel is actually working because none of them got enough investment to know.
Pick one channel. Run it hard for six weeks. Spend more than feels comfortable. The goal isn't to optimize — the goal is to get a definitive answer about whether that channel can work for your specific business.
For most early-stage companies, the first channel should be the one with the shortest feedback loop. Usually paid search or outbound to a tight list. Both produce signal fast.
Phase 3: Build the capture infrastructure
While the channel is running, build the boring parts that catch demand once it shows up:
- A homepage that tells visitors what to do in one sentence
- A pricing page that doesn't require a sales call to get a number
- A lead form that takes 30 seconds, not 3 minutes
- An automated follow-up sequence that doesn't go cold after the first email
- A way to track which channel each conversion came from
This is the part most demand gen playbooks gloss over. They focus on channel work and assume the conversion infrastructure exists. In a from-scratch build, it usually doesn't. The boring infrastructure work matters more than the channel choice.
Phase 4: Read the data honestly
After six weeks of running one channel hard, you'll have one of three outcomes:
The channel is working. Cost per qualified meeting is sustainable. Conversation quality is high. Pipeline is real. Double down. Add a second channel only after the first is humming.
The channel isn't working, but the demand is. You're getting traffic and meetings, but conversion is broken somewhere. The fix is conversion infrastructure, not channel.
The channel isn't working, and the demand isn't there. You don't have a channel problem. You have a positioning, audience, or product problem. Don't try to fix it with more channel investment. Go back to Phase 1.
What this isn't
This isn't a brand-building playbook. It won't make you famous. It won't get you a Wikipedia page. It produces the kind of demand gen function that can show closed-won revenue in 6 months — which is what most early-stage companies actually need.
Brand-building gets to happen once you've proven the unit economics. Not before.
The sequencing
Evidence first. One channel hard, not three light. Boring capture infrastructure before exciting brand work. Read the data without flinching.
This is unglamorous in a way most demand gen content isn't. It's also what actually produces pipeline from zero.