Stitch Fix UK, where customers have been subscribing since its 2019 launch, provides a unique perspective on the personalization challenge facing retailers. Unlike the prior experience with an in-person stylist who had the advantage of observing customers’ shapes, coloring, and style preferences during dress shopping sessions, Stitch Fix operates remotely.
Customers create profiles with their measurements, budget, and fabric preferences (such as an aversion to polyester). They then request a ‘Fix,’ which consists of a box containing five selected items, accompanied by a note to the stylist detailing their desires and style. The stylist curates this selection and sends it to the customer’s home for trial. Items they like are kept and billed for, while the rest are returned. A £10 styling fee is applied but credited towards purchased items, and there’s a discount if all five items are retained.
The departure of Stitch Fix from the UK emphasizes the immense challenge of achieving effective personalization in the retail industry. It sheds light on the disparity between what consumers express as their preferences and what they ultimately choose, making the pursuit of true personalization a formidable task.
“The Complexity of Achieving Personalization Stitch in Retail”
First and foremost, it’s important to note that some of the most cherished and frequently worn items in customers’ wardrobes originated from Stitch Fix, including pieces from their own labels. When Stitch Fix got it right, they truly excelled.
However, despite possessing a wealth of information about customers like those who subscribed, there were numerous instances where the experience didn’t feel completely tailored to their unique taste and preferences. It’s worth noting that Stitch Fix had access to a more substantial and detailed dataset about their likes, dislikes, and style compared to any other retailer they engaged with.
Why, then, couldn’t Stitch Fix consistently deliver boxes filled with clothing that customers absolutely adored and couldn’t resist keeping?
For starters, it seemed that Stitch Fix struggled to grasp slightly more “edgy” personal styles, despite customers’ best efforts to convey them. Emails were inundated with images of clothing that had the opposite effect – they didn’t excite customers to schedule a Fix.
Part of the challenge was Stitch Fix’s limitation on the number of characters allowed for making requests or providing feedback to stylists. This constraint made it difficult to effectively communicate the nuances of their styles and the types of items that would seamlessly integrate into their wardrobes.
Upon signing up, Stitch Fix requested that customers rate a limited selection of 14 outfit styles to gauge how well they aligned with their personal styles. However, for a service centered around personal style, it felt like an in-person video call during onboarding might have been more beneficial. Such a call would have allowed a stylist to observe how customers dress, explore their existing wardrobes, and engage in a more in-depth discussion about their desired look and feel.
Remarkably, Stitch Fix’s stylists never even saw a photograph of customers. This highlights a fundamental flaw in contemporary retail personalization – achieving genuine personalization is resource-intensive and time-consuming. It necessitates the development of ongoing customer relationships, which don’t always yield the desired results. This is why many retailers often resort to archetypes, demographics, and mass email campaigns.
The issue with this approach is that consumers are diverse, complex, and challenging to predict. Even individuals who share similar traits, features, or tastes are distinct in their preferences.
Moreover, it’s challenging to assess the quality of the data provided by customers. For example, Stitch Fix’s website featured a ‘Style Shuffle’ feature where customers were presented with outfit images and asked to express their style preferences with a thumbs-up or thumbs-down. While this gamified method of communication was convenient, it revealed a significant data challenge. When customers encountered a top they adored but disliked the pattern, how could they effectively communicate this? Would giving it a thumbs up imply they liked all aspects of the top? Conversely, would a thumbs down jeopardize their chances of receiving a different top with a similar style in their Fix?
The persistent challenge in personalization and customer data lies in the quality and interpretation of that data. The extensive data that Stitch Fix believed they possessed about their customers may not have held as much value as anticipated, making their task considerably more difficult.
Additionally, it’s not just about the accuracy of the data but also about the ability to derive meaningful insights from that data. This raises questions about who is interpreting and utilizing the data – artificial intelligence or human beings?
Stitch Fix employs a combination of both in its model, as do many other retail businesses. However, the model may not have adequately accounted for customers’ unpredictable human nature.
Consumer Behavior Gap:
Stitch Fix encountered significant challenges when translating customers’ written feedback and requests into actionable guidelines. The reality is that customers sometimes don’t precisely know what they want. Or what they thought they wanted might change when presented with an alternative.
For instance, according to Stitch Fix’s AI, it appeared that customers exclusively desired black clothing, as they repeatedly indicated in their profiles, feedback on outfit suggestions, and the items they retained from each Fix. However, they would occasionally include a personal note to their stylist expressing a desire for a pop of color. The data suggested one thing, while their verbalized preferences said another. How could the stylist reconcile these conflicting signals to make a successful selection?
The Stitch Fix model perpetually grappled with the disparity between what customers thought and said versus what they actually did.
Ideally, Stitch Fix aimed for customers to schedule regular recurring Fixes, yet they only ordered a couple each year. Customers rationalized that they didn’t require monthly clothing deliveries since they didn’t shop that frequently. While it’s true that they didn’t intentionally make a new purchase every month, they came to realize that they did buy items more often than they initially thought. These unplanned purchases, whether new or second-hand, often occurred when they spotted something appealing on social media, during a store visit, or in response to brand emails.
The gap between customers’ perceived shopping frequency and the reality turned out to be more substantial than anticipated. The issue was that because customers had to consciously choose to receive a Fix, Stitch Fix was limited by their perceived behavior, rather than benefiting from the actual behavior. This is a common scenario for many retailers.
Similar to many shoppers, customers have become increasingly mindful of their purchases and the duration for which they retain items. Stitch Fix’s choices perfectly aligned with these criteria.
However, operationally, Stitch Fix grappled with some of the most challenging and costly aspects of retailing, particularly home delivery
Read our more blogs on retailinsights