Hey Product Hunt 👋 Get 30% off your first 3 months if you sign up using this link

Better Done Than Perfect · Season 8 · Episode 2

Proven Ideas for A/B Testing with Sahil Patel

You'll learn why you should get inspiration from the tests run by other companies, how early you should invest in A/B testing, where to use disruptions in your website design, and more.

Sahil Patel

How do you come up with new ideas for A/B testing in B2B SaaS? In this episode, we talk to Sahil Patel, CEO of Spiralyze. You'll learn why you should get inspiration from the tests run by other companies, how early you should invest in A/B testing, where to use disruptions in your website design, and more.

Show Notes 📝

Thanks for listening! If you found the episode useful, please spread the word on Twitter mentioning @userlist, or leave us a review on iTunes.

Key Learnings 💡

Prior to joining Spiralyze, Sahil was the founder and CEO of ER Express, a healthcare SaaS business that he ran for 11 years. In 2016, he became an early client of Spiralyze and it opened his eyes to the power of A/B testing.

After selling ER Express to Constellation Software in 2021, Sahil was looking to start the next chapter of his journey:

"The founder at Spiralyze, Gajan Retnasaba, was looking for a partner to help him take the business to the next level. And it was the right time for me to do it."

Sahil joined the team in 2022 and currently serves as its CEO.

Going with "proven winners"

As an A/B testing company, Spiralyze helps their clients run tests using proven winners:

"There are 34,000 websites that run A/B tests somewhere on their site, and we scrape all of them. So if there's a company out there that runs a test and it wins, we know about it. And then we find the tests that work, which we call 'proven winners', and we run those tests for our clients."

Sahil shares that A/B testing is a very cruel game:

"Only 11% of tests will beat the control. You have to run about 10 tests just to find 1 win."

Which is why it's better to test out the proven winners:

"If you run tests that have already worked for another company, your rate of winning is twice as good. And if you run a test that has worked for many companies, your likelihood of beating the control is three times higher."

How early should you invest in A/B testing?

If you're a pre-Series A or a Series A B2B SaaS company, it might be too early for you to invest a lot in A/B testing:

"Should you do some? Yes. Can you get some benefit? Yes.

But if you want a real CRO program where you're continuously testing, you need a certain maturity of your marketing function. In our experience, that's post-Series A."

Sahil shares that you can use conversions as guidelines to determine if you should be investing more in A/B testing. Conversions will depend on what model you're using:

  • Product-led growth: Free trial signups
  • Sales-led growth: Demo requests or booked calls with the sales team
  • Hybrid

"If you're getting at least 50 conversions/month, you're in a good place and you should be doing some A/B testing. If you're doing at least 100 conversions/month, it's a no-brainer. You should absolutely be doing it."

How mature should your marketing tech stack be?

Sahil says not to overthink about your tech stack before doing any A/B tests:

"Series A and beyond companies will be running one of the big CMSs to host their website or are using tools like Marketo for their landing pages or HubSpot for forms. Certainly bigger companies will have a very robust tech stack.

But they're not prerequisites to testing. I don't think you have to be super sophisticated to get some benefits from A/B testing."

The three bulletproof A/B testing ideas for B2B SaaS

Based on tens of thousands of primary and secondary data that Spiralyze has worked with, here are the three bulletproof testing ideas:

  • Showing your product
  • Skimmability
  • Addressing objections

Showing your product

The hero section of a typical B2B SaaS website would likely show:

  • Lifestyle images. These are usually stock photos of people who are shown to be enjoying using a product.
  • No image, just pure text. A lot of text and very hard to read.
  • Product screenshots.

"There's a lot of data that shows you'll get more people interested in your product if you let the image do the heavy lifting."

Sahil gives the example of car ads. While people wouldn't really use a car to go off-road or climb up a mountain on a typical day, these visuals are what draw people in:

"They do it because it captures people. What are people buying? They're buying the spirit of the car."

The same principle applies to B2B SaaS software.

Tips on showing your product

Tip 1. Use a stylized version of your interface

This works best if you have a complicated or "boring" product. For a complicated product, it helps your audience get a clear idea of what your product does. For a boring one, it creates something that's interesting and exciting.

"Have a designer digitally draw the idea of what your product does. You can put little callouts of little featurettes of what your product does."

Tip 2. Add a little bit of blur to the product image

"This creates a little bit of enticement. Think of it as, you're looking at a cool car and the windows are slightly tinted. That makes you want to look in and see what's inside the car."

Tip 3. Showcase your best-selling product or the one that has the biggest "wow" appeal

It could be pretty overwhelming and confusing to decide if you have a multi-product suite. Sahil advises to pick the best-selling one or the one that has the biggest appeal.

Tip 4. Be careful not to mislead your audience

"Don't show them something that they're never going to get. If the reality is dramatically different than the screenshot, it's going to be jarring and you'll get the opposite effect: a lot of people will sign up and they won't activate because they'll feel misled."

Tip 5. Show your hesitant CEO what your competitors are doing

"Some client's CEO says, 'Hey, this is our IP. It's a secret sauce. Our competitors are going to steal it.'"

While there's no answer in A/B testing that can address this concern, Sahil says what you can do to address this hesitation is to put together a swipe file of your close competitors:

"Put together a swipe file of your 15 closest competitors and find product screenshots. Show them to your CEO and say, 'Hey, these are great! Let's go to the product team and just tell 'em to copy all of these,' then you do it."

Sahil says that product screenshots don't entirely reveal the core of how your product works, and that's why it's okay to use them:

"Everyone has a product roadmap and a product that works well. Particularly in B2B SaaS,the power is in the workflowof how everything is interconnected and works.

Yes, you might have these unbelievable signup pages, unbelievable dashboards in your product. Someone like your competitor will see one single static image on your website and they're just unlikely to copy. And even if they do, it's not going to create a lot of value for them."


Sahil says that as marketers, we're biased towards what we've put on the website:

"We imagine that the audience is pouring over our website, reading it word for word because we've spent so much time on the copywriting and re-writing."

But the reality is, the audience is multi-tasking:

"While they're reading your website, they've probably got Slack up in one window. Or they're texting with their spouse about the grocery list, and they're quickly going to your website, reading something and moving on.

If you're not convinced, go get a heat mapping tool installed and you'll see just how quickly people scroll."

Which is why the web rewards easy to skim and easy to read copywriting.

So how exactly do you improve the skimmability of your website?

  • Use short and sweet bullet points, not paragraphs
  • Make a bold claim in the headline, and use three bullet points to support it

Where to use disruption on your website design

Before thinking about trying running A/B tests on disruptions, Sahil emphasizes that you should be clear on what your product does:

"You should first make sure that you and your audience have extreme clarity on what your product does. If you're so disruptive that your audience doesn't know what it does, the disruption is going to fail."

Sahil says to go conventional on your copywriting:

"Conventional is good. Easy to skim, avoid complicated language, and reduce the cognitive load on your audience."

On the other hand, you can play around and be unconventional with your overall page composition:

"For example, most B2B websites have some copy on the left in the hero section and a form on the right. A great test to run is to flip those two (form on the left, copy on the right). It sounds really simple but it's a high-performing test.

This doesn't mean it'll win for everyone, but it's a great way to try something that disrupts the pattern without creating something that's so bizarre."

Handling objections

Objections are the little whispers in your audience's heads that keeps them from signing up such as:

  • "I have to put a credit card in. So if I put it in and I forget to cancel, it's going to keep charging me."
  • "Are they going to make it impossible for me to get out of the free trial?"
  • "If it works and I start using it, am I getting the full product? Or am I getting a kneecapped version of the product?"
  • "What features am I getting and not getting?
  • "If it works and I love it, is it going to work with my tech stack?"

Based on thousands of data points, Sahil says to pick the most common objection to improve your conversions. How exactly would you narrow down on it?

"The best way is to interview your customers or potential customers.

You can also do exit polling on the site. When someone navigates away, you can ask them why they're leaving. But I always recommend there's no substitute for talking to human beings."

Addressing your audience's objections is what Spiralyze calls the "Prozac approach":

"It's a little piece of assurance that solves a nagging little/big issue."

Objection: "I don't want to put my credit card in for a free trial."

For example, you found that your audience's most common objection is putting in a credit card for the free trial:

"A really simple thing is to put 'No credit card required' below your CTA. Something as simple as that can have a really powerful effect in driving more conversions for your product-led growth business."

But for businesses that have high intent traffic, you wouldn't really need it:

"You wouldn't need this because they need your product. Maybe there's a particular pain point and they're just ready to do it. If that's your product and your audience is all there, then that's wonderful."

Objection: "Am I getting the full or kneecapped version of the product during the free trial?"

For audiences who are concerned about what version of the product they're getting in the free trial (full version or kneecapped version), Sahil says that you should tell them upfront:

"You should never mislead. So if they get the full thing, you should say 'Full functionality in the free trial.' Really simple and it addresses a pain point."

Sahil shares their experience of running tests for RocketReach, a platform where you can find contact information for your lead generation program:

"In the free version, you'll get a certain number of lookups every month. So under their CTA, we just showed how many free lookups you get when you use the free version.

So people know exactly what they're getting and if that doesn't work for them, they don't sign up and it's fine. But for most, this test showed that they got a higher conversion rate with it."

Objection: "Does it work with my tech stack? How many integrations do I get out of the box?"

"You can put 'Unlimited integrations' or '5 free connections with the starter package' under the CTA."

Final advice

Do run A/B tests that have worked for other companies.

"Run tests that there's some evidence, even if it's not wholly scientific, that have worked for other companies. Go check out 10 of your competitors, build a swipe file, and track them for a quarter. Instead of getting a single snapshot, you're getting a longitudinal picture and an N > 1."

Don't get hung up on not getting high-volume traffic before testing.

"You can have a measurable impact on your business, get more revenue, and reduce your CAC by running some straightforward principles on your site, even if you have low volume traffic."

Thanks for listening! If you found the episode useful, please spread the word on Twitter mentioning @userlist, or leave us a review on iTunes.

Subscribe to win your free shirt

Join our mailing list below, learn about new episodes as they go live, and win one of our custom-designed shirts from Cotton Bureau.