Home Featured Conversations: Kurt Kimmerling, Dressometry

Conversations: Kurt Kimmerling, Dressometry


In the first of a small collection of interviews with innovative entrepreneurs and thought leaders to be published throughout the year, here WhichPLM chats with Kurt Kimmerling, Co-founder & CEO of Dressometry.  With undergraduate and graduate degrese in Math from the University of Chicago, Kurt is a self-taught developer with over 10 years’ experience in data science.Kurt chats to us about the motivation behind Dressometry, machine learning, the transformation of the business, and future plans.

Name: Kurt Kimmerling

Occupation: Co-founder & CEO, Dressometry

Likes: Eating Frosted Flakes without milk, math

Dislikes: Carbonated beverages, loud crowded places, dancing

Words to live by: “Fool me once, shame on you, but teach a man to fool me, and I’ll be fooled for the rest of my life”

WhichPLM: To begin, can you tell our community a little more about Dressometry? We see that you offer customers the opportunity to select certain criteria – colour, occasion, price for example – in order to find the right dress.

Kurt: Dressometry is a growing B2B machine learning start-up based in New York City, and we build customer-centric search solutions for fashion retailers.

We apply computer vision (which is a field of machine learning where you teach computers to see like humans see) to fashion images – dresses, shoes, bags, tops and beyond – to help our clients understand a product’s detailed attributes – for a dress this might be the color, sleeve length, the neckline, etc. – and how to use them to build solutions to everything from the online shopping experience and outfit recommendations to trend forecasting and marketing.

WhichPLM: So, where did Dressometry come from? What prompted you to create this kind of experience for the customer? Did you see a gap in the marketplace? And a question that I’m assuming to be related to that: why dresses in particular?

Kurt: Good question. Well, I watched my wife spend an entire afternoon shopping for dresses online, and by the third hour, I decided to step in and learn more about why it was such a frustrating process for her.  I noticed that she spent most of her time jumping from brand to brand, scrolling or searching across dozens of tabs.  Even when she found her way to a dress she liked, she said she’d buy it if only she could change one thing about it.

At that point I saw a problem worth solving, and after hundreds of interviews with other shoppers, I learned it would be relatively simple to solve if only the attribute data were put to better use by retailers.

Insufficient product data is a problem for every retailer, largely because the way they collect it is manual – slow, expensive and inefficient.  While retailers don’t have a great handle on their product data, they do have an incredible stock of product images.  That’s when I knew that the solution to obtaining good product data would come from applying computer vision and machine learning to those images.

WhichPLM: You mention machine learning and image; how do you use image searching within your platform? Are your algorithms image-based as opposed to purely text based?

Kurt:  Most competitor technology is text based, and therefore can only compile the information that a retailer has already collected and published.  We are image-based, which means we can help retailers upstream in the process to actually produce the tags.

We think there’s more than enough work to do to connect images with text data, and doing this well will go a long way for fashion retailers.  You’d be surprised at how big of a challenge this is for companies regardless of their size.

Search-by-image is a neat, emerging technology, but shoppers typically don’t have an image for the item they have in their mind.  Search by image is also notoriously inaccurate.   Our platform provides a more flexible and more accurate solution.

WhichPLM: You speak about the difference between Dressometry and its competitors. From a purely customer point of view, how does the experience at Dressometry differ from that with another e-tailer? Correct me if I’m wrong, but other online retailers also offer filters when searching for the perfect item, be it colour, style, or price; what differentiates Dressometry from the pack?

Kurt: We offer an interactive shopping experience by giving customers the ability to ‘click to change’ an attribute to find similar, better items.  For example, when a shopper is viewing a dress online, and loves everything about it except for its short sleeves, we allow them to find the most similar items with long sleeves.

Each time a customer ‘clicks to change’ an attribute, they’re helping the e-tailer find them the perfect item.  It’s similar to interacting with an in-store associate, and now we’ve built it online.   This all but eliminates needless scrolling, bookmarking, and toggling between browser tabs, and it is only possible with the kind of highly detailed product data we produce for our clients.

Our customers tell us all the time how much they enjoy the shopping experience and how much they wish they could replicate it across other fashion products they buy online – from bags to shoes and beyond.

WhichPLM: And how many businesses do you currently work with? And how do you manage to keep up-to-the-minute information around each of them and their latest product offerings?

Kurt: Our clients range from a small e-commerce startup, not unlike our own, all the way to major household names.

To be able to serve clients of that range, we knew we’d need to build the capability to tag tens of thousands of images per day, across all product categories, in a highly accurate way, and to ensure the process is smooth for exchanging data (pictures and tags) with our clients. We built a platform to allow us to do just that, and we work closely with our clients – like additional members of their teams, if you like – to make sure we’re the first to know about new products rolling out online.

WhichPLM: Let’s talk a bit about transformation. Dressometry began as a consumer-facing platform; you’ve now found that you can expand this and target more areas. Tell us about that, and where you see it going?

Kurt: The experience we’ve built on Dressometry is something we’re working on providing to retailers across the market. Instead of competing to be the one-stop home for fashion shopping, we’re seeing more success partnering with retailers as clients, and building solutions to the many challenges they’re facing.

At our core, we’re data science and fashion experts, and we want to help retailers use their data more effectively.  We think by focusing on generating good product data, we can build solutions to issues like search, trend forecasting, outfit recommendations, marketing, and beyond.

WhichPLM: Interesting. As part of this expansion, do you see an opportunity to work with PLM, using the designers, product attributes, as part of a retailer’s online filtering?

Kurt: Absolutely. We’re always eager to put our technology to use to solve a wide range of challenges.  We think the fashion industry can do some revolutionary things once people begin to think about their data (and its applications) more holistically, and that’s the exact role we hope to play.

But it’s also difficult for designers to provide all of the necessary data for retailers, since every retailer structures their filters and names their attributes differently.

WhichPLM: And what’s your underlining technology? Have you considered APIs to other industry solutions?

Kurt: The first element of our technology is how we’ve used computer vision and machine learning (applying convolutional neural networks) to train our computers to recognize specific aspects of a fashion product image.  So, our computers have learned how to spot a fit-and-flare dress, or a cross-body bag, or kitten-heel shoes.

From there we work with our clients to build easy-to-implement APIs to use that data to help re-create the Dressometry experience on their own product pages.

WhichPLM: Looking at the future, what plans do you have to expand your technology and business even further?

Kurt: Our view of the world is that better product data will power better search, trend forecasting, outfit recommendations, marketing, and beyond.  We want to help fashion retailers align their approach to data with their business strategy.

At our core, we’re a product data company, and we think that our approach positions us nicely to grow alongside our clients.

WhichPLM: To round off this discussion, is there anything else in particular you’d like to share with our community that we haven’t touched upon today?

Kurt: I’d like to finish with a final thought. Despite the incredible growth of fashion e-commerce over the last few decades, we at Dressometry see the industry as still very young, and as only beginning to understand how the revolution that put every fashion product available for sale online has generated new problems many retailers cannot solve alone.  Those who approach product and customer data holistically will be a few steps ahead of their competitors, and in an industry as dynamic this one, that’s a major advantage.

We think that what we’re building at Dressometry will go a long way to helping those retailers take advantage of the changing landscape in a way that ultimately makes every customer better off.

Stay tuned for further ‘Conversations’ throughout the year. 

Lydia Mageean Lydia Mageean has been part of the WhichPLM team for eight years now. She has a creative and media background, and is responsible for maintaining and updating our website content, liaising with advertisers, working on special projects like our PLM Project Pack, or our Annual Publications, and more.Joining mid-2013 as our Online Editor, she has since become WhichPLM’s Editor. In addition to taking on writing and interviewing responsibilities, Lydia has also become the primary point of contact for news, events, features and other aspects of our ever-growing online content library and tools.