
Last week I wrote about the attribution trap: how cutting content at budget review time often removes the thing warming buyers up before they ever showed up in your data.
A few of you replied with a version of the same question. Something like: "We know the content matters. But when pipeline is thin, what are we actually supposed to do?"
It's the right question. And it almost always leads back to the same confusion.
When pipeline feels thin, the instinct is to generate more leads.
So founders reach for the lead gen lever. Paid social, paid search, more events, another SDR. Things that produce names and fill the top of the funnel. The names arrive. Conversion stays flat. CAC climbs.
This is probably the most expensive pattern I see in B2B scale-ups. And it comes from treating demand gen and lead gen as the same thing.
They're really not.
Lead gen finds people who are already in-market. Already aware of the problem, already looking at options, already open to a conversation. It captures demand that exists. Good lead gen puts you in front of the right buyer at the right moment.
Demand gen creates that awareness in the first place. It's the content that makes a founder realise they have a problem worth solving. The newsletter that reframes how they think about a decision. The LinkedIn presence that positions you as someone who understands their world before you've ever been introduced.
If the demand isn't there, lead gen finds people who aren't ready. More spend, same conversion.
Across the B2B scale-ups we're working with at the moment, conversion rate is the problem that comes up most.
Paid acquisition is running. MQLs are arriving. But conversion through the sales cycle is doing the real damage to CAC and to the economics of scaling.
In almost every case, top-of-funnel volume isn't the issue. The issue is that the people arriving in the pipeline haven't been warmed up to the problem. They don't already understand why this matters. They don't have a view on what good looks like. So the sales team carries all of that education themselves, across every single conversation.
This is an expensive problem and even more damaging with a small TAM. Sales cycles tend to be long. Deals need multiple stakeholders aligned. If you're re-educating from zero in every discovery call, you're burning resources that compound: longer cycles, more touchpoints to get to a decision, more drop-off as the buying committee loses momentum.
The sales conversations at our best-performing clients feel different.
They start later in the buying decision. The prospect already knows the framework. They've read the newsletter, seen a few LinkedIn posts, maybe heard the founder on a podcast. By the time they get on a call, they're past "what do you do?" They're asking "how does this work with our team?"
The conversation is about implementation, not convincing.
That's what demand gen actually does. It moves the work that used to happen in the first two or three sales calls upstream, into content and consistent presence. The buyer has had time to absorb a point of view before they ever raise their hand.
This is specifically what we build for clients. The demand layer that makes the rest of the funnel cheaper to run, because the people arriving in it are already partway there.
If CAC is climbing but lead volume looks fine, check where the drop-off is happening.
Drop-off between MQL and SQL: likely a targeting or messaging problem. The leads aren't the right people.
Drop-off inside the sales cycle: usually a demand gen gap. Buyers are arriving without context. The sales process is carrying too much of the education.
The fix for a lead gen problem is better lead gen. The fix for a demand gen gap is content that does the education earlier, so your salespeople don't have to.
A quick diagnostic: pull your last ten discovery call recordings. Listen to the first five minutes. How much of it is spent explaining the problem? If it's most of the call, the demand layer has a gap. If they're explaining the problem back to you and asking how you'd solve it, the demand engine is working.
Drop me a line if you want a second set of eyes on where the drop-off is happening in your pipeline. I'm happy to take a look.