Demand Generation vs. Lead Generation: Why Your A/B Tests Need a Different Goal for Each

‍Most teams treat A/B testing as a single discipline with a single goal: get more conversions. But what you’re optimising for — demand or leads — changes everything about how you design your tests, what you measure, and how you call a winner.

What is the difference between demand generation and lead generation?

These two terms get used interchangeably all the time, and that’s where the confusion starts.

Lead generation is about capturing contact information from people who are already interested. A formfill, a demo request, a trial sign up — these are lead gen actions. The person has shown intent and you’re collecting it.

Demand generation is about creating that interest in the first place. It’s the work that happens before someone is ready to fill in a form — the blog post that introduces your category, the homepage that articulates why your problem matters, the pricing page that turns a curious visitor into a convinced one.

In practice, both happen on yourwebsite. And both involve A/B testing. But the goal of each is fundamentally different, which means the way you run tests for each should be too.

Why do most SaaS teams test both goals the same way?

Because conversion rate is easy to measure and everything else isn’t.

For example, when you run an A/B test on your homepage and Variant B gets 30% more demo requests, that looks like a clear win. It’s a number that goes up. It’s attributable to a specific change. It’s the kind of result that gets shared in a Slack channel.

What’s harder to measure is whether those demo requests are any good. Whether the people who converted on Variant B are actually the right buyers. Whether they’re going to turn into closed deals, or whether they’re going to waste three hours of your sales team’s time before going dark.

This is the lead quality trap.You optimised for volume, you got volume, and now your pipeline looks healthy on paper while your close rate quietly drops. The A/B test wasn’t wrong — you were just measuring the wrong thing for the goal you actually had.

A 30% lift in demo requests sounds like a win. But if those demos close at half the rate, you’ve made your sales team twice as busy and your revenue exactly the same.

How does the lead quality trap play out in practice?

Here’s a scenario that plays outmore often than most teams want to admit.

A SaaS marketing team is tasked with growing pipeline. They run a test on their homepage hero — the control has a specific, outcome-focused headline aimed at their core ICP, while the variant broadens the message to appeal to a wider audience. The variant wins. More people click the CTA. More demos get booked.

Three months later, the salesteam starts pushing back. The leads don’t seem as qualified. More calls are ending with “not a fit right now.” The ACV on new deals is lower. Nobody connects this back to the homepage test because by this point the test has been called, celebrated, and moved past.

What happened? The broader message attracted more clicks, but from a less qualified pool. The demand genwork — building conviction in the right buyers — was quietly undermined by a lead gen optimisation that prioritised volume over fit.

This is not a failure of A/Btesting. It’s a failure to define the goal before the test started.

What should you actually be measuring for each goal?

The metric you optimise for in a test should reflect the goal the page is serving. That sounds obvious, but in practice most teams default to the same metrics regardless of context.

When the goal is demand generation

Demand gen pages are doing educational and persuasive work. They’re convincing someone that your product is worth their serious attention. The right tests here are ones that improve the quality of the conviction you’re building, not just the quantity of clicks you’re generating.

•     Metrics to watch: MQL-to-SQL conversion rate,demo-to-close rate, average contract value on converted leads, time spent on page, scroll depth

•     Tests that make sense: Headline positioning and specificity, value proposition framing, case study and social proof placement, content depth and structure

•     What to be cautious of: Optimising purely for CTA click rate — a broader, vaguer message will almost always generate more clicks and worse leads

When the goal is lead generation

Lead gen pages are like talking to people who are already interested. They’ve done some of the demand gen work already — now the job is to reduce friction and make it as easy as possible totake the next step.

•     Metrics to watch: Form completion rate, CTA click rate, cost per lead, page-to-conversion time

•     Tests that make sense: Form length and field order, CTA copy and placement, page layout and friction reduction, trust signals near the conversion point

•     What to be cautious of: Reducing so much friction that you attract unqualified submissions — removing the company size field might lift form completion rate while filling your CRM with leads that will never close

How should your CTA reflect which goal you’re serving?

Your primary CTA is one of the clearest signals of which goal a page is serving — and one of the most powerful things you can test.

“Book a demo” is a demand gen CTA. It’s asking for a significant time commitment from someone who needs to be genuinely interested. It filters for intent. If it’s working well, the people who click it are closer to buying.

“Download the guide” or “Start for free” are lead gen CTAs. They’re lower friction, designed to capture a wider pool, and rely on nurture to do the qualification work downstream.

Testing which CTA sits as your primary action isn’t just a conversion rate question — it’s a question about which motion you’re prioritising. A product-led growth team and a sales-led team should be running different CTA tests, measuring different downstream outcomes, and calling winners on different criteria.

The mistake is testing “Book a demo” vs. “Start for free” purely on click rate. The right question is: which CTA produces more pipeline per visitor, not which one produces more clicks.

What does a demand gen test look like vs. a lead gen test?

The practical difference comes down to what you’re changing and what you’re measuring against. Here's an example:

A demand gen test

You’re testing whether a change in your homepage headline — from feature-led to outcome-led — improves the quality of visitors who go on to request a demo. You measure this not by counting demo requests, but by tracking what percentage of those demo requests become sales-qualified leads within a specified time.

A lead gen test

You’re testing whether removing two fields from your contact form increases form completion rate without materially affecting lead quality. You measure form completion rate as your primary metric, and you set a quality threshold — if the MQL rate on submissions drops below a certain level, the test is a loss even if form completions went up.

The key difference is that demand gen tests need a longer measurement window and a downstream metric, while lead gen tests can often be called on on-site behaviour alone — as long as you’ve set a quality floor.

How do you know which goal your page is actually serving?

It’s worth doing a quick audit of your highest-traffic pages before your next test cycle. For each page, ask two questions:

•     Is this page talking to someone who is already interested in buying, or is it building that interest from scratch?

•     What is the primary action I want this visitor to take,and what does taking that action signal about their intent?

A homepage talking to cold traffic is a demand gen asset. A pricing page visited by someone who has already read three case studies is a lead gen asset. A retargeted landing page is a lead gen asset. A blog post ranking for a problem-awareness keyword is a demand gen asset.

The same page can serve different goals for different traffic sources — which is why the most sophisticated teams eventually run separate tests for organic vs. paid traffic to the same URL. But for most SaaS teams, just being explicit about the primary goal of each page before testing it is a significant step forward.

Key takeaways

•     Demand generation and lead generation are differentgoals that require different A/B testing strategies, different metrics, anddifferent criteria for calling a winner.

•     The lead quality trap happens when you optimise a demand gen page for lead gen metrics — you get more conversions and worsepipeline.

•     For demand gen tests, measure downstream outcomes like MQL-to-SQL rate and demo-to-close rate, not just on-site conversion rate.

•     For lead gen tests, set a quality floor before youstart — a lift in form completions that tanks your lead quality is not a win.

•     Your primary CTA is one of the most powerful signals ofwhich goal a page is serving. Test it with downstream metrics, not just clickrate.

•     Connecting your A/B tests to CRM data doesn’t need to be perfect to be useful. Even a basic tagging process gives you better signalthan on-site metrics alone.

‍