Open Google Ads and you can create a campaign type called Demand Gen. Open LinkedIn Campaign Manager and you can attach a Lead Gen Form. Both buttons exist, and one marketer clicks both in the same afternoon. Ask that marketer about demand generation vs lead generation and you will get two answers, both sincere. That is how the two words ended up meaning almost nothing.
They do mean something. The distinction is not strategic taste, and it is not funnel diagrams. It is about what you count, and when. Run a demand programme, then report it with lead generation’s numbers. The quarterly review kills the demand programme, judged by a metric that was never measuring it. We see that more often than we see a real strategy mistake.
The two words name two different jobs
Strip the labels off and look at the work.
Demand generation is the job of making more people want the thing, and making the people who already want it think of you. It operates on a market. Do it well and more people in that market know who you are, and the ones already looking think of you sooner.
Lead generation is the job of turning interest into a contact record. It operates on an individual. It hands you a row: a name, a work email, a company, and the moment that person raised a hand.
Those are not competing philosophies. One creates the wanting, the other captures the wanter. A company can be excellent at the second and have nothing to capture. That is the failure mode nobody labels correctly, because an empty form looks like a form problem.
Why the labels get swapped
Part of this is our industry being sloppy. The bigger part is that the ad platforms now sell both words back to you as products. The words have quietly become the names of buttons.
Google Ads sells a campaign type called Demand Gen. Google describes it as capturing engagement and action across YouTube including Shorts, Discover, Gmail, Maps and the Display Network. It is a real product and it does real work. It is also a line item. So “we are doing demand generation” can now truthfully mean “there is a campaign of that type in the account”. Those are not the same sentence.
LinkedIn does the mirror image. A Lead Gen Form arrives pre-filled from whatever the member already has on their profile, which usually covers name, contact details, company, seniority, job title and location. Anything not on the profile, such as a phone number, they still type themselves. LinkedIn’s own pitch is that members send you their professional details in a couple of clicks, without being taken away from the platform. Beautifully frictionless. In our experience it also means a contact record can exist before the person has read much of anything.
So the vocabulary drifted. Buying the campaign type is not the same as generating demand, and collecting the form is not the same as generating a lead worth having.
Lead generation counts a moment. Demand generation moves a market.
You find out at completely different speeds.
Lead generation produces its evidence instantly. Someone fills the form, the row appears, the cost per lead updates the same day. You can watch it move hour by hour. That immediacy is genuinely useful. It is also why lead generation wins the budget argument against a motion that cannot answer as fast.
Demand generation produces its evidence late, and mostly in places that carry no form. When it is working, you tend to see it in direct traffic and in branded search. Sales calls open differently, closer to “I have been reading your stuff for a while” than to “remind me what you do”. Someone typing your name into a search bar had a cause. Your brand-defence campaign bills for that click, and it did not create the reason behind it. In our work, the self-reported line on the form, the one asking how they heard about you, carries more truth than the attribution model does.
Neither of those is soft. They are just slow, and they land somewhere other than the campaign that caused them.
Cost per lead is a dial, not a verdict
This is the part that does the damage. Let us do the arithmetic in public.
Take a fixed spend of $10,000 and a piece of content that 1,000 people reach. Put a form in front of it. Some share of those people fill it in. A 5% fill rate gives you 50 leads, so each one cost $200. Double the rate to 10% and 100 leads arrive at $100 each. Twenty percent produces 200 leads, and the price halves again to $50. Push it to 40% and you book 400 leads at $25.
Data table
| Form-fill rate | Cost per lead |
|---|---|
| 5% | $200 |
| 10% | $100 |
| 20% | $50 |
| 40% | $25 |
Nothing about the market changed across those four rows. The demand was identical. The only thing that moved was the share of people who filled the form, and cost per lead fell by a factor of eight.
That share is mostly a design decision, not a verdict from the market. In our work, asking for four fields instead of nine tends to lift it, and so does putting a template behind the form instead of a research report. Neither change makes one additional person want what you sell. Both improve your cost per lead.
Now run it ungated. A thousand people read the thing. Zero leads. Cost per lead is not a large number here. It is undefined. The metric quietly reports the best version of that content as a total failure.
So cost per lead is mostly measuring your form, not your market. That is what the arithmetic above shows, and in our work it is what the metric does in practice too. It says very little about whether anybody wants what you sell, and that is the question demand generation exists to answer. Judge a demand programme by cost per lead and you reach one conclusion: put a form on everything. The metric agrees with you. Everything else does not.
Where lead capture actually sits
The useful mental model is not two boxes side by side. Lead capture is a layer inside demand generation, not a rival to it.
Demand has to exist before it can be captured. Every form on your site is a net, and a net does not produce fish. When the numbers are bad, the instinct is to work on the net: shorten the form, change the button, add an offer. Sometimes that is right. Often the net is fine and nothing is swimming past it. Six weeks then disappear into form optimisation while the real problem sits upstream, untouched.
There is a second reason the capture layer flatters itself. B2B buying groups are commonly reported at six or more people. Your form collected one of them. The other five still have to be reached, convinced, and given something to forward internally. You can chase them with more capture, and plenty of teams do, by multi-threading the account and asking each one to fill something in. But a form only works on someone who already knows why they are filling it. Getting the other five to that point is demand generation doing its job.
That one captured person still has to be worth capturing. We wrote that up separately in what actually makes a lead marketing qualified.
How to tell which one you are actually running
Forget what the plan says. Three questions settle it.
Data table
| Gating choice | People who read it |
|---|---|
| Gated, 5% fill | 50 |
| Gated, 20% fill | 200 |
| Ungated | 1,000 |
What happens if you remove the form? If the answer is “then we get nothing”, you are running lead generation. A demand programme survives the question, because reaching people was the point of it. In the arithmetic above, taking the form off is what put the content in front of all 1,000 rather than 200.
Who are you reaching that has never heard of you? If your targeting is retargeting, your existing list and people searching your category by name, you are harvesting demand, not creating it. That is a legitimate and often correct thing to do. It is just not the thing that grows the pool.
What would you report if you were forbidden to mention leads? If the honest answer is nothing, there is no demand programme. If you can point to branded search, direct traffic, share of voice or what buyers say on calls, there is one. Judge it on those.
Go back to the same $10,000 and 1,000 people. Gate it, get a 20% fill rate, and you book 200 leads and 200 readers. Gate it with a 5% fill rate and you book 50 leads and 50 readers. Leave it open and you book zero leads and 1,000 readers. Your reporting sees one of those three rows as excellent, one as poor and one as a catastrophe. A buyer would rank them in a completely different order.
Running both without one eating the other
Almost every B2B company should run both. The failure is rarely the mix, it is the shared scoreboard.
Split the reporting first, before the budget. Count capture work in leads, cost per lead and lead-to-opportunity rate. Review it weekly, because it answers that fast. Count demand work in reach into the target market, branded search, direct traffic and self-reported source. Review that quarterly, because that is when it says anything. Two motions, two scoreboards, two cadences.
Then protect the slow one. In our experience a shared scoreboard means the demand budget loses the argument, month after month. It is being asked a question it cannot answer inside the window. We worked through the same structural argument for named accounts against a market in ABM vs demand generation. It shows up again when you price the two motions together instead of separately, which we covered in what demand generation costs.
| Demand generation | Lead generation | |
|---|---|---|
| Operates on | A market | An individual |
| Produces | People who want the thing | A contact record |
| Evidence shows up | Late, and off-campaign | Immediately, in the campaign |
| Read it in | Branded search, direct traffic, what buyers say | Leads, cost per lead, lead-to-opportunity |
| Review cadence | Quarterly | Weekly |
| Who it can reach | People who will never fill in a form | The person who filled in the form |
If you want help drawing that line inside your own numbers, that is the work we do at COSEOM®.
FAQ
Is lead generation part of demand generation?
Yes, in practice. Lead generation is the capture layer that sits inside a demand programme. Demand has to exist before a form can collect it, which is why treating them as rivals produces the wrong fix when the numbers go soft.
Which one should a small B2B company start with?
Start by harvesting whatever demand already exists, because it is faster and it funds the rest. Then add demand creation before the existing pool runs out. Skip that second step and the warning you get is the capture numbers sliding for no visible reason, which arrives long after the cause did.
Why is my cost per lead so good and my pipeline so bad?
The cause we find most often is that cost per lead is measuring your form rather than your market. Shorten the form or offer something easier to want, and the number improves without any change in demand. Check whether those captured contacts are the people who actually buy, and whether anyone else in the buying group has ever heard of you.
Does a Google Demand Gen campaign count as demand generation?
Only if it is doing that job. Google describes the campaign type as capturing engagement and action across its surfaces, which tells you where the ads run, not what the work is for. You can point it at people who have never heard of you or at people already close to buying. Having one in the account proves the line item exists. It does not prove anybody new wants what you sell.
How do you measure demand generation without leads?
Watch four things. Branded search for your own name, direct traffic, reach into the target account list, and the self-reported “how did you hear about us” answer. None is precise on its own. Read together over a quarter, they are much harder to move with a gated PDF than a lead count is.
Should we ungate all our content?
No, and that is the wrong shape of question. Gate the things a buyer would only want after they are already interested, such as a pricing calculator or an assessment. Leave open the things that create the interest in the first place. Then stop comparing the two on the same metric.
COSEOM Team
COSEOM is a multilingual B2B growth agency. Since 2008 the COSEOM team has run international SEO, generative engine optimization, paid media and demand generation programs for B2B software and technology companies, in English, German, Spanish and Portuguese. This account publishes our research, playbooks and product guides. More about COSEOM