As we approach Fit Analytics’ 10-year anniversary, we would like to reflect on our journey with you and share our learnings along the way.
It all started in 2010 with the launch of UPcload, a webcam-based body modeling service that allowed users to create a personalized fit profile using a laptop and a CD as a reference item. While the idea behind UPcload was extremely relevant to online shoppers – fit recommendations based on physical photo scans – ultimately the technology was a bit too laborious for consumer adaptation at scale.
We learned that although shoppers (in general) are willing to spend time finding their perfect size – they are more open to a technology that does the work for them. Asking customers to step away from their laptop, phone, or tablet to take photos ultimately led to page abandonment. We also found that shoppers were reluctant to pose in tight clothing in front of a camera for fear of where those photos could end up.
After dropout rates spiked to over 80%, we quickly pivoted to create a fully web-based size advisor, which we named Fit Finder. Technology had advanced considerably in the years after our webcam solution, and this enabled us to create a machine learning solution that required little work for the shopper, while providing the most accurate recommendations in the market. Fit Finder was our flagship product utilizing machine learning – this is how our Fit Analytics data platform was born.
Let the Journey be a Journey
To really understand what customers wanted while shopping online, we launched several user testing initiatives. We asked consumers to complete Fit Finder questionnaires of varying lengths – some with eight questions and others with just three.
In alignment with standard industry thinking, we assumed that shoppers wouldn’t want to take a long survey to receive a size recommendation. Interestingly, we found that when the questionnaire was too short, it caused customers to doubt our suggestion. Through our user testing, we discovered that consumers were willing to put in the effort to find their best size online, even if it meant a slightly longer user journey. Contrary to conventional wisdom, we have found that when it comes to a size and fit solution – a medium-length sequence of questions tailored to the shopper outperforms a shorter one. A/B variant testing shows higher conversion rates and comparable completion rates for a medium-length sequence across both desktop and mobile.
We also identified the great value in asking customers for their fit preference – what could have seemed like a simple question to some, actually had a big impact on the final size recommendation. If shoppers preferred a tighter fit, they could go down one full size in an item – and this had a direct effect on the probability of returns.
Bodies, Brands, and Social Proof
It is essential to use body reference images in addition to a brand comparison item when generating an accurate size recommendation. While brand comparisons are useful, they often don’t provide all of the necessary information to identify the right size. Shoppers can indeed select an item from a similar brand, but the details around the type of product and fit of it are not provided, which compromises the recommendation accuracy. That’s why we decided to use both body modeling and brand comparison. The combination of the two allowed for recommendations that were much more accurate than solely relying on brand comparisons.
Another interesting finding that came to light during user testing was what we call Social Proof. When shoppers reach the final screen of Fit Finder, they are shown their best fitting size based on the answered questions, as well as characteristics of other similar customers, and sales and returns records. The size recommendation is presented as a percentage showing how likely the two sizes are to fit the shopper. Knowing that a size medium, for instance, is 92% likely to fit well, while a large is only 40% likely to fit, proved to increase shopper confidence and led to higher conversions.
Fit Indicators show how a garment is likely to fit along certain physical dimensions (hips, waist, chest, etc.). Fit Analytics used Fit Indicators to recommend sizes until 2013, when we made an important discovery by testing Fit Indicators against Social Proof.
Through this method of testing, we found that Social Proof performs significantly better in terms of generating trust and converting browsers to buyers. Fit Analytics is currently the only solution that uses Social Proof to recommend sizes.
The Launch of Fit Finder Features
As we improved Fit Finder and applied our learnings from user testing, we seized the opportunity to launch additional features to further refine our premiere solution. Some of these include Multiple Size Alert, Product Suggestions, and more. Multiple Size Alert is used to help retailers reduce returns and multiple size orders. When enabled, shoppers are notified when they attempt to add a second size (of the same item) to their cart. At that moment, customers are prompted to use Fit Finder to find their best size and, therefore, order just one size instead of two or more.
Product Suggestions is another feature of Fit Finder which recommends additional styles that are in stock in shoppers’ size and fit preferences. This feature is activated when an item is sold out. Product Suggestions uses machine learning algorithms to recommend similar in-stock products, leading to an increased average order value.
These features – and several others – are available to retailers with the implementation of Fit Finder. Our Client Partner team ensures that you enable as many features of Fit Finder as needed to fully support your business, customers, and overall goals.
Expanding Our Product Suite
As we expanded the offerings of Fit Finder, we also broadened our overall product suite. We realized that just providing a size recommendation wasn’t enough for today’s consumers. Customers are looking for a personalized shopping experience no matter where they shop – online, in stores, or on mobile. In addition, retailers are hoping to better understand their customers and provide them with the best products and experience.
With that in mind, we created three new products: Fit Connect, Fit Intelligence, and Fit Consult. Fit Connect is our API-driven personalization platform. It gives retailers access to sizing and style recommendations from shopper and product data to create highly personalized shopping experiences and compelling promotions. Shoppers are able to search by size and fit to find items they know they’ll love and that will fit well.
Fit Intelligence provides retailers with a range of data analytics tools, including several key reports, that allows retailers to make informed business decisions. It gathers insights that drive impact to apparel retail businesses in many ways – notably to supply chain, forecasting, marketing, and merchandising. Some of the reports include: customer analysis reports (like age and gender), product improvement reports (like fit consistency and missing sizes), and product and sales reports (like sales trends and out of stock).
Fit Consult helps retailers plan and produce apparel more profitably while optimizing their supply chain. It assists companies with enhancing production size charts to reduce manufacturing inefficiencies and more tightly match real-world customer demand. Fit Consults also offers expert assistance and pre-launch consultations to ensure a successful go-to-market.
The combination of these four products ensures that retailers meet their customers at every step of the shopping journey and learn from the data collected to make better decisions.
Cheers to Many More Years of Fit Analytics
Fit Analytics continues to evolve every year. We’ve come a long way since our camera solution and predict many more exciting changes and innovations. We take great pride in our user testing and formulate new products and product updates based on our learnings. We are looking forward to the next 10 years!
To learn more about how we can help you and support your specific needs – get in touch today.