Increasing Revenue – Reducing Return Rates – Optimizing Customer Experience
Customers are more digital than ever before and their expectations of the buying experience are rapidly increasing. However, online shopping is not yet able to completely replace local shopping. Online shops often cannot provide enough information about a product for customers to make an informed decision about which size and fit is ideal for them.
The wrong product size or fit can lead to disappointed customers, which results in high return rates and bad reviews. What most consumers often lack when shopping online is individual professional advice and customized products and recommendations. This goes hand in hand with the increasing demand for personalization. According to Accenture, 91% of consumers are more likely to purchase products from brands that recognize and remember them and that provide relevant offers and recommendations.
3D avatars help uncertain customers to find the right product
Which size is right? Which product best suits the individual customer’s preferences? These questions can only be answered in e-commerce if the complete body posture and form of each customer is considered individually.
In an online store, these characteristics can be best represented with a 3D avatar that exactly replicates the body’s measurements. To create the avatar, customers upload full-body photos and, if necessary, add personal data such as height. Keypoint extraction and silhouette estimation are among the factors that then provide input into a deep-learning model. The model in turn creates a personal 3D avatar in just a few seconds.
To create a recommendation for the customer, the corresponding product, for example a mattress, is virtually merged with the 3D avatar. This objective method is very precise and increases the accuracy of the recommendation, thus increasing customer satisfaction and significantly reducing product return rates.
Individual advice in online stores: which products benefit?
3D scanning is already being used in the fashion industry to accurately measure body shapes, provide size recommendations and sell fashion in a more targeted way. This is because size specifications and the associated measurements of a garment often differ from brand to brand. As a result, customers often order and return an item of clothing in multiple sizes, cutting profits for online retailers. For this reason, the fashion industry is increasingly turning to solutions that recommend the right size to customers from the start, resulting in fewer returns.
The 3D avatar solution takes on even more importance for advice-intensive products, where a misguided purchase can bring with it not only constant frustration but can even damage your health. Typical examples include mattresses and bicycles. These are products that require individual consideration – especially if the customer has to spend a lot of money for a product. For instance, when purchasing a mattress, various parameters need to be taken into consideration, for example pressure on the body and the optimal position of your spine in order to lie in an ergonomically correct way. A mattress should be neither too soft nor too hard, and the shoulder and pelvic areas of the mattress should be perfectly tailored to you as an individual.
That is why more and more consumers are looking for individual expert advice when shopping online. With a 3D avatar, many different products can be tested virtually online in seconds to find the perfect match from hundreds of products.
Our powerful technology transforms the industry
By integrating our 3D avatar solution, retailers can offer their customers a sophisticated way to obtain recommendations for the ideal product within seconds. With just a little information such as height, weight and full-body pictures (from the front and from the side), the cloud-based system can calculate a person’s size and best fit.
The AI-based technology takes into account individual usage and compares many products in the background. This means that it creates a physical simulation in which the 3D avatar interacts with various products. In effect, the avatar acts just like a customer in a retail store. It tries out many different bicycles to determine the correct frame size and position of the seat and handlebars. Or it lies on a mattress to test it for various health and comfort aspects.
How a 3D avatar helps you to get healthy and comfortable sleep
Once the 3D avatar has been created, our solution is able to determine the best mattress for any particular customer using this representational avatar. Mattresses are measured using a sophisticated, custom built machine that records precise pressure measurements at certain indentation depths over the surface of each mattress, which allows us to create high-fidelity digital representations. In order to find the perfect mattress for a customer, we simulate an avatar in a preferred sleeping position (i.e. on the side, the back, or the stomach) on mattresses that have previously been measured and digitized. Using finite element analysis, we are able to simulate the interaction of thousands of springs and dynamic soft body collisions with millimeter precision of the collision bodies and envelope margins.
Put more simply, we create a physical simulation that evaluates how the customer’s 3D avatar, which incorporates their unique characteristics, responds to an entire night of sleep on each mattress. To calculate the optimal mattress for a customer, the software takes into account several parameters including the angles of the spine, the avatar’s indentation depth, the pressure response on the body, torso angles, and more. With these simulated measurements we can recommend the right mattress for each customer.
Finding the right bike size with biomechanical technology
We can use the 3D avatar to support customers with the complex task of choosing the best frame size and configuration for their individual body proportions and seat position. A poorly fitting frame can lead to a reduction in power transmission as well as increased or imbalanced strain on the muscles.
To obtain a recommendation, the customer can select a bicycle model whose geometry is stored in a database along with the various frame sizes and bicycle dimensions. To make an optimal bike recommendation, we don’t look at the 3D avatar in a static state, but rather simulate riding the bike using a biomechanics engine.
For this purpose, the different body segment lengths are extracted from the 3D avatar and used together with such scientific parameters as the elbow angle or ankle joint angle to transform the 3D avatar into a bicycle rider. We consider not just one adjustment option of the bicycle, but iterate using different settings, e.g. saddle positions, to optimize the seating position. This allows us to simulate the joint angles and positions as well as the segment angles for each pedal position.
After that, typical bike fitting parameters can be determined, such as the knee angle at the lowest crank angle point or the maximum hip angle. From the data extracted from the simulation and established scientific knowledge, we calculate a score for each adjustment option and frame size. Based on these, we are able to offer each customer a confident recommendation for their optimal bicycle.
More effective online shopping with the latest technology
Simulation provides the most effective way to suggest the best product fit during the purchase process. Our machine-learning-based recommendation finds the right product based on each customer’s avatar, but also based on information from other similar consumers. This means that the system can find consumers who share the same attributes and can make recommendations based on users with similar parameters, e.g. height and weight. This approach utilizes the self-feeding database connected to the customer platform or online store. In this way, the user gets a perfectly fitted product or selection of different products to choose from. In addition, time-consuming product comparisons and frustration when buying online are a thing of the past, along with excessive returns and the associated negative impacts on the environment.
