A/B testing for B2B software

B2B software is the hardest place to split test well, and it's better to know that before you start. Traffic is modest, most visitors aren't buyers yet, and the money arrives months after the visit through a sales team. A test judged on closed revenue would need years.

That doesn't make testing useless. It means choosing a metric your volume can support, testing only changes big enough to show up, and knowing when to skip the test entirely.

Why testing is different for B2B software

Sales cycles outlast tests. If a demo booked today closes in four months, a six-week test ends long before any revenue from it exists. You need a number that arrives sooner and still tracks revenue.

Volume is low at every step. A site with 6,000 visitors a month might book 60 demos and close 12 deals. Twelve outcomes a month can't separate two versions of a page.

The visitor isn't the only buyer. A champion reads your site, then a committee decides. A page can win the champion and lose the finance lead.

Lead quality varies. A shorter form can double demo requests and fill the calendar with students and competitors. Raw bookings flatter changes that make it easier to book, whether or not the right people book.

Pick the metric your volume can support

From most to least reliable:

If you have a self-serve plan, it's the best place to test, because money moves the same day. See revenue per visitor.

The tests to run first

1. Show prices for smaller plans

Replace "Contact us" with real prices for your entry plans, keeping contact-us for enterprise. Measure self-serve revenue per visitor and sales-qualified demos per visitor. The pricing guide covers the setup.

2. A self-serve plan next to "Book a demo"

Add a card checkout for a starter plan beside the demo button. Measure self-serve revenue per visitor plus pipeline value per visitor.

3. Book a time on the page

Let visitors pick a slot in a calendar on the page instead of submitting a form and waiting for an email. Measure demos held per visitor, not demos requested.

4. Fewer form fields

Cut the demo form to work email and company, and look up the rest. Measure sales-qualified demos per visitor, since raw requests will rise either way.

5. A headline that names the role

Replace a category slogan with a headline naming who uses the product and the job it does for them. Run it on every page with the same call to action to pool traffic. Measure qualified demos per visitor. More in the landing page guide.

How much traffic you need

A made-up example. Your product and pricing pages get 6,000 visitors a month, about 200 a day. 1% book a demo and 20% of demos close.

With standard settings (95% confidence, 80% power):

Even the fastest line takes over six months. To get an answer sooner, test changes that could plausibly move demos by 30% or more, pool every page that shares a call to action into one test, and test on self-serve revenue where you can. If none of that fits, skip the test, make the change you believe in, and track before and after with the caveats that come with it. The low traffic guide covers these options, and the minimum detectable effect calculator tells you what your traffic can detect.

Common mistakes

• Calling a winner on raw demo requests.
• Stopping a test after two weeks because 14 demos beat 9.
• Testing a gated content form and counting ebook downloads as pipeline.
• Running a test while sales changes its qualification rules, which moves the metric on its own.
• Spending a year on a website test when ten customer interviews would answer the question.

How Outtest fits

Outtest judges every website and pricing test on revenue per visitor from a payment tool such as Stripe, Paddle or Chargebee. It reads payments, not a CRM, so it can't judge a test on pipeline or demos; signups and similar actions are shown only for context.

That makes it a good fit for B2B products with a self-serve plan or card checkout, where tests like showing prices or adding a starter plan move revenue it can see. For a sales-only funnel where money arrives months later by invoice, most tests would end as draws at Outtest's 42-day limit, and a CRM-based setup will serve you better. Where it fits, pricing and checkout changes always wait for your approval, and a winner needs at least 7 days, a 90% chance of beating the original and a lift of at least 10% by default.