At the Marketo Marketing Nation Summit this year, our slogan was ‘A/B Testing is Dead,” and it drew a lot of reactions from show attendees. You can imagine why – many of the event attendees spend hours if not days per week on A/B testing different elements of their emails, landing pages, and digital ads. But as we explained to our booth visitors and others who noticed the slogan, when it comes to generating the best business results from your marketing efforts, A/B testing is being supplanted by multivariate testing.
The Status Quo
A/B testing has been the norm in B2B marketing for some time. There are good reasons why B2B marketers adopted the practice.
A/B testing started to apply actual science to marketing and took a lot of the guesswork out of how to get more clicks on your web pages and digital ads, higher conversion rates on landing pages, and more opens and clicks on emails. Forward-thinking marketers adopted A/B testing, and the marketing technology community responded. As a result, most marketing automation systems, email marketing platforms, and landing page optimization tools now ship with A/B testing functionality built right in.
But smart marketers and successful companies don’t just reach a goal and then rest on their laurels – they move the bar higher and seek out ways to improve business results even further. This is especially the case now that CMOs are under constant pressure to deliver not just vanity metrics like email clicks and social follows, but actual sales pipeline and revenues.
Too Much Work, Too Little Payoff
The problem with A/B testing is that it’s finite in what it can do for marketers. You can set up and test two binary choices and see which one garners the better results. However, if you want to test multiple options, you have to perform additional A/B tests.
This can get extremely time- and labor-intensive, as mentioned in this KissMetrics article that breaks down a lot of the benefits of multivariate testing for business analytics. A/B testing takes a lot of manual effort from marketing teams, effort that multivariate testing can eliminate, and given the amount of time A/B testing can take even well-staffed marketing teams, the campaign can often end before you have a chance to enjoy more than incremental performance improvements.
Enter Artificial Intelligence and Machine Learning for Marketers
Concurrently, artificial intelligence and machine learning are exploding in the business world – as we continue to make computers smarter and faster, businesses have been applying AI and machine learning to business intelligence, enterprise operations and marketing in all sorts of interesting ways.
With multivariate testing, however, we have an opportunity to apply artificial intelligence and machine learning to marketing in a whole new way – marketing campaign execution. By employing machine learning principles and utilizing models based on artificial neural networks, marketers can now automate multivariate testing and the get nuance and detail that need, without tying up precious marketing team resources for days or weeks at a time.
A demand generation engine powered by AI can test a variety of advertising campaign elements and channels and see what combinations convert the best. It can also connect to the CRM and automatically discern and ultimately drive the most pipeline and revenues. The benefits for CMOs are twofold: first, they free the marketing staff from the mundane tasks of setting up UTM parameters and other campaign elements, so they can focus on more creative, value-added pursuits. Second, they can perform the rapid experimentation that CMOs are seeking today as they adopt agile marketing methodologies.
Companies who leverage multivariate testing very quickly understand how fonts and layout work on a reader’s mind – where people’s eyes go according to how data or images are presented, and what they pay attention to on the Web. All of this is actionable, real marketing intelligence that drives conversions and pipeline.
So, is A/B testing truly dead? Yes and no. The concept of testing variables in your marketing campaigns is not dead at all – it’s evolving and expanding. The point of our slogan is, A/B testing is too slow and manual labor-intensive to meet the demands of modern B2B marketers. A/B testing isn’t enough – B2B marketers need “A through Z” testing and beyond, multivariate testing that encompasses everything they’re doing to reach customers so that they can deliver predictable and scalable top-of-funnel leads and return on investment for their business.