Split testing using Vypr

What is split testing?

Split testing allows you to compare different versions of one element of a product, whether it be packaging, images, marketing claims, materials, colourways, etc. It is typically used by online companies such as Google to optimise their marketing strategies, but can be a useful tool for retailers.

It works by segmenting your audience, and showing one group of consumers one element (let’s say a version of packaging), and another group (or groups) another variant. You can then see how well each variant is performing, based on the data your testing reveals.

Why use split testing?

Split testing is a key tool in behavioural economics to understand impulsive decision making. These “spur of the moment” decisions are part of consumers’ “System 1” thinking (largely unconscious and automatic), as opposed to System 2 (deliberate, conscious, calculated thought). Split testing through the Vypr platform accesses this System 1 process, and allows you to understand what consumers will do as opposed to what they say they will do. You can find out which of two or three variations is the most popular, without your audience knowing what is being tested. This works better than showing each consumer two or three options and asking them to choose one.

For example, if you want to test whether a “low in fat” claim should be added to packaging, if consumers can see the two side-by-side (one with the claim, and one without) they will focus in on the claim as the key element and this will influence the result. But this behaviour doesn’t represent a real impulse purchase situation, where the consumer will only have one version to either buy or not buy. By doing a split test you can tease out whether the “low in fat” claim leads to higher purchase intent without consumers realising that the claim is the element you are testing. The results from split testing are consistently more representative of real-world behaviour than giving users an explicit choice of options.

Split testing is relatively inexpensive, and a valuable asset to your data collection. Our consumer panel is made up of thousands of participants, available to give you feedback on your products.

It is important to note, however, that split testing doesn’t tell you the reasons why one element is more successful than another. To get the best results, it’s best to test one element at a time. Changing too many elements at once for each segment means you won’t know what is responsible for success. Try tweaking which font you use for one test, or the placement of the product images for another.

How can I use Vypr to run split tests?

Using Vypr, a split test will show you which version of each element works best for your potential customers. This means you can not only identify the best way to market and sell your product, but gain an insight into the elements of the product itself. You can test different flavours, shapes, ingredients, nutritional claims… in fact any variable of your product can be tested.

We offer two different forms of split testing: Split by description, and Split by image. Split by description lets you show the same image to the panel, but with a different description. Split by image lets you show the same description, but with a different image. You can split your audience into two or three groups.

This eliminates the need to physically make every product concept. By showing one part of your audience one idea, and a different part a different idea, you can determine which collection of ideas will amount to a product that will perform more favourably on the shelves.

© 2018 Vypr Validation Technologies Ltd | Registered No. 8589377 England & Wales | Privacy Policy