One of the holy grails of retail tech is the personalisation of a customer’s shopping experience, so customers feel like they have their own digital personal shopping assistant who knows their tastes almost better than the customer does. This involves a significant investment in tech and time as retailers build data warehouses and use advanced analytics to build up a data picture of customer behaviour.
There are no shortcuts, and while a retailer can say they’re personalising by adding a customer’s name onto an email and/or sending discount voucher on a customer’s birthday, it won’t result in customers having a noticeably different retail experience.
Larger US retailers have responded by acquiring tech start-ups to improve data gathering and analytics capabilities. Other online retailers have invested in product-recommendation engines that factors such as customers web browsing and purchase history. Attempts at personalisation are not just about driving customer loyalty through a better and tailored retail experience but to also encourage customers to find new products they wouldn’t normally consider – if the recommendation engine can give them the right recommendations.
US online clothing retailer, Stitch Fix, has personalisation at the very heart of its business model to better predict what clothes customers will be interested in, lessening the amount of information the customer must provide before Stitch Fix can match them to a product that they’ll purchase.
While the theory sounds elegant, the reality is that learning a customers’ tastes is more complicated than you’d think.
Stitch Fix asks customers to complete an 80-question survey which asks how a range of clothes matches the customer’s styles preferences. This data is added to the information on which items a customer keeps or returns.
The downside of this approach is that providing detailed information about fashion preferences is a long-winded process that can alienate time-poor customers – especially if the recommendations made isn’t right the first time. Navigating this hurdle saw Stitch Fix creating a mobile app called Style Shuffle which asks customers to give a thumbs up or thumbs down to a range of clothes.
Since its 2017 launch, 80% of Stitch Fix’s 3.1 million customers have played Style Shuffle, rating a boggling two billion clothes. Using the data generated, Stitch Fix created a machine learning algorithm that tries to predict what items the customer might be interested in, based on what other Style Shuffle users liked. The upside of this approach has been that Stitch Fix is now better placed to forecast inventory demand at a more precise and detailed level.
Augmented Reality: The perfect shopping aid?
Chances are that you’ve probably already heard of virtual reality, where you put on a headset with built-in screens so you can immerse yourself in a virtual world. Augmented reality takes things one step further by overlaying the real world with virtual objects. Sheer geekiness aside, augmented reality is also a boon for retailers who are using it in increasingly innovative ways.
Take a task as simple as choosing a pair of glasses. This can be a long and tiresome experience. Customers often end up trying on dozens of pairs before they find glasses that they’re happy with. Last year, optician and retailer, Specsavers launched a service in the US that takes the hassle out of choosing glasses.
The tech Specsavers use is called the Frame Styler. It is an augmented reality app that runs on an Android tablet. It uses the tablet’s built-in cameras to produce a 3D model of the customer’s face. Using machine learning and AI smarts, it then selects which glasses would best suit the person’s face, based on the model of their face, their gender, and age. Customers get to try on multiple glasses styles in 3D, and because they can see how the glasses look, they can quickly and easily make a purchase decision.
Tech writer Pat Pilcher shares the latest gadgets for work, rest and play in every issue of FMCG Business. To subscribe visit http://www.fmcgbusiness.co.nz/magazine-subscribe/
