Testing Techniques for Desire
Leveraging lean techniques to discover desire before designing experience
In my previous post, I argued that UX practitioners need to test for desire before testing for delight and other usability considerations. This is because the only way of creating a breakthrough product or a feature is to quickly arrive upon the most desirable concept from a large consideration set.
Traditional usability tests are designed for this purpose. While methods like RITE testing are popular they are only still based on rapidly refining a single concept. A/B testing is meant to compare different concepts but there is no testing that is designed to unearth a winning concept. In short, traditional usability studies and testing is great for deeply learning about the pros and cons of single concept but cannot be employed to broadly learn about a broad range of concepts.
For example, let’s say personalization through better recommendations on an e-commerce site is thought to be something that will lead to a better UX & greater revenues. Typically, design will start with a competitive review of recommendation systems and then explore concepts that make sense for the business. From there on, the UX & product teams will iterate and test to find the most promising recommendation system for the site. This is usability research and A/B testing phase. During testing, questions will include- “What do the stars mean?”, “How can you write a review?” , “Where do you expect the recommendation to surface?”, “What happens when you click on the submit button?”, “Did you like X better than Y?”. Now, contrast this with how a lean entrepreneur treats the hypothesis & tests for desire. The questions that arise are quite different — “What if I served up these type of recommendations here?”, “What if you could talk to another user at this point in the experience?”, “What if after this purchase, I texted you similar products?”, “What if a stylist gave you advise on pairing items?” . In other words, the assumption that recommendations improve UX is tested in a variety of different ways with a variety of different techniques to deliver them. The aim is clearly to cheaply and quickly test for the desire to have recommendations in the first place instead of to test for the best recommendation experience that can be delivered.
This stems from the fact that lean entrepreneurs are trying to establish their business models before building solutions or even writing out their business plans. In the case of UX testing, business models are generally not at stake but user behavior certainly is. The difference is in testing to ascertain that a product will be used as opposed to testing to ascertain that a product will be understood & liked.
Several businesses have used lean testing like concierge testing or wizard of oz testing to discover desire and demonstrate demand before designing the experience. One famous example of this type of testing is Zappos.com which started out as a portal on which pictures of shoes from the local shoe store were uploaded which customers could order. Upon receiving the order, the shoes would be purchased from the store and shipped to the customer. The founders of Zappos were only interested in proving that customers have a desire to shop for shoes online. No better proof of that than actual user behavior which in this case was actual orders.
Ultimate victory for designers is when the design is used by a large number of users. So, the next time a feature request comes in, instead of jumping in to survey data, competitive reviews or sketching experiences, we can be more strategic with our first step & test for the desire for the feature instead.