Personalisation Technology: Testing & Optimisation
Last time I began this series on what to look for in personalisation and optimisation technology with my Personalisation Technology Manifesto. Now I’m going to look at each of the items in the manifesto – the technology promises – and expand on what brands should be demanding from their personalisation technology.
At the core of personalisation are questions of passive preference: the fluid perceptions that subconsciously drive consumers towards one product or another. In this article, we’re going to start with the entry point for most organisations in to the world of optimisation – the basics of teasing out those preferences with accuracy and ease. The simplest approach to personalisation is a healthy routine of optimisation through content testing, and the simplest tests are the basic A/B test and the slightly more complex multivariate test (MVT).
A/B testing is just what you’d expect: providing visitors with two variations on a piece of content and testing which results offer better metrics- more clickthroughs, longer time spent on the page, the completion of a full video. There’s nothing fancy about it, but if it’s done properly, an A/B test can give you surprisingly accurate data on what works and doesn’t work: just remember to keep the variations simple enough to pinpoint what variable is affecting the results.
A similar approach, MVT takes the black and white comparisons of the A/B test and complicates them by mixing and matching variables, increasing the possible variations of any given page, while also offering more accurate results. While an A/B test might contrast two header images, or button placements, or whole page layouts, a multivariate test will try out each combination — now instead of looking at the value of individual aspects of design, you’re getting information on how they all work together. More concretely, A/B testing lets you know that one thing works better than another: MVT is the process of learning _why _those things work better. Don’t be afraid of MVT – it can often be helpful to identify the predictive elements in terms of conversion, which you can then refine through further A/B/n testing between many variations of that element.
So what should you look for in your testing and optimisation technology:
Ease of implementation: IT simply can’t be caught up on such a critical but straight forward project, nor can you allow the pages affected to be cluttered with obsolete or bloated code. Consider a suite like Adobe’s: a single-line of code included on every page of the site or referenced in each screen of your app is all you need to implement – this is made even simpler with the dynamic tag management core service, which allows you to implement and activate tags and campaigns effortlessly, and makes removing them equally as simple.
Ease of configuration: Consider the many hands involved: who’s looking to test what, and how these various teammates will stay organised. Tests should be easy to deploy and easy to quality assure, and it should be simple to understand what tests are running where on your properties. The Adobe Target Visual Experience Composer enables your team to create, deploy and maintain optimisation activities with ease. Issues with digital governance and prioritisation are often overlooked but your solution should show you, at a glance, a summary of all of your activities in a dashboard and provide alerts about potential activity collisions. Adobe Marketing Cloud’s Assets service also provides organisations with a central hub from which creative and technical solutions can be shared, stored, edited and published, free from silos.
Ease of maintenance: Be prepared to create and change activities on the fly. Solutions that provide uncomplicated editing and publication put the control in the hands of the marketer and enable you to test and optimise your sites and apps in real-time. If a successful marketing activity starts driving large volumes of traffic to your site then you should be able to quickly and easily optimise the experience for that traffic and allow campaign managers to adapt to information received while the marketing campaigns are still running.
Ease of reporting: Ultimately, your data means nothing if you’re not able to show statistical confidence in your results. Platform integration with an analytics solution is key to making the most out of your campaigns and tests, regardless of the content. Adobe Target enables you to report in the platform itself or see your results directly in Adobe Analytics, so that you can seamlessly move between analysis and optimisation.
Ease of improvement: This is the sweet science of testing and optimisation, and it can be a real challenge. Improvement of testing procedure is about constant iteration, learning when to make tests more specific for purposes of accuracy or broader towards the end of discovery. Your optimisation technology must make it easy for you to understand what is and isn’t working and deploy the next stage of your testing strategy.
Taking these guidelines from page to practice is best illustrated with a case study. Conrad Electronic manages a digital marketplace containing hundreds of thousands of individual products, servicing a vast selection of European countries. With so much content and so many consumers, Conrad’s challenge was streamlining their digital experience and uncovering key advances in informational display and navigational ease.
With Adobe’s real-time analytics and optimisation tools, Conrad pinpointed the optimal discount display, driving up participation in sales and specials by a significant margin, especially for new users. And this distinction would prove useful in predicting contradictory behavior to come: Conrad further discovered what kind of additional product information positively impacted conversion for returning customers, allowing them to comfortably target both groups. The result of this is that Conrad created a 4% uplift in cart additions and navigation use, all based on tests and templates that could be edited and implemented in mere seconds, and you can learn more about Conrad’s approach by watching this Personalisation & Optimisation breakout session from the recent Adobe Summit (Adobe Target in ‘real-life’: driving optimisation with Target)
Conrad provides us with a solid image of what a successful testing campaign looks like- focused discovery of extremely specific channels to uplift. With the right tools, exploiting these discoveries becomes effortless. With A/B and MVT explored, we’ll be moving on to the next chapter of the personalisation manifesto: business rules and geo-targeting.