Building and piloting a Digital Changing Room

Background

lululemon was exploring how to tap on an insight called, “Digital Reflection”: when lululemon guests visit the retailer’s PDP and see bodies that don’t resemble theirs, it becomes hard for them to see themselves in the brand.

They sought tools to better calibrate how a garment would fit them when online shopping, which sometimes user generated content (e.g., youtube) would solve for. It was apparent current fit tools (e.g., size chart) weren’t enough, and more needed to be done to help them confidently shop online and ultimately convert to the brand.

Early explorations of the solution were called a “Digital Changing Room”.

• Model-led experience: what if lululemon guests could shop for their favourite or new styles by first selecting a model that looks like them (e.g., age, height, weight, identity) to see how those clothes would look on someone similar to them.

• 3D Avatar-led experience: what if lululemon guests could build a digital twin based on their unique proportions and characteristics (e.g., body shape and curves, skin tone, etc.) to understand how an outfit would approximately fit on their body.

As I onboarded as a Lead on this project, my job was to carve the path forward from there on: framing and executing the next round of research, validate or invalidate those experiences use for guests, stress-test their viability and feasibility, and recommended to Senior Leadership Team whether it made sense to productize or not.

All of this was done while navigating a stakeholder-heavy organization where inputs from different teams (UX, Product, Ecomm, Brand, 3D & Product Development) and leaders (SVP, Chief Digital Officer) was necessary for projects eventual buy-in, success, and pilot launch in the brand's app.

Understanding different shopping intentions

To stress-test those two digital changing room concepts, we arrived at the conclusion that the brand wouldn’t make a solution like that in-house knowing there were startups in-market with similar solutions. These aren’t necessarily the capabilities the brand would invest in, but was open to rent/acquire from outside vendors.

After designing demos with these companies (Veesual, Browzwear, BODS, etc.), we aligned on a research study focused on understanding whether these solutions help lululemon guests calibrate their size and fit by getting a sample of guests to demo these experiences on their own.

We used a combination of qualitative and quantitive methods to unpack that major research question:

Besides feedback into the usability and UX elements of each specific demo prepared, research allowed us to better understand the value each kind of Digital Changing Room brings, impact to the brand and perception from consumers, and some of the behavioural frictions around it:

Unconvering shopper's hesitations

We knew it didn’t matter if both versions of the digital changing room were successful at helping shoppers calibrate size or style when online shopping as long as their hesitations weren’t resolved.

To continue this research and stress-test those hesitations, we held a brainstorming session focused on frictions shoppers might be experiencing and potential UX solutions and tactics known to help (e.g. 9 out of 10 shoppers who used a digital changing room were recommended their right size.)

We aligned on testing these problem hypotheses and potential UX solutions through two journey-based prototypes introducing shoppers to a new digital changing room experience on lululemon.com.

• 'Meet them where they are' hypothesis: shoppers will be more likely to engage with a digital changing room if they first encounter it in context of their natural online shopping behaviors.

• 'Onboarding' hypothesis: shoppers will be more likely to engage with a digital changing room if they are well-informed about the experience before first encountering it in their online shopping experience.

• Easing off the tech jargon: content design (CTAs, Nav Bar notifications, etc.) focused on not leading with the tech powering the product (3D-avatar, AI-based tech, etc.) and speaking about it from the benefit it provides instead.

Both prototypes were tested through two rounds of research with participants (20-44 yrs old and XS to XXL sizes) who had shopped for a lululemon item in the past year from that study and claimed to have experienced size and fit problems while online shopping for it:

1. Unmoderated study with 30 shoppers 16 shoppers went through ‘meet them where they are’ hypothesis prototype vs. 16 went through ‘onboarding hypothesis prototype. At the end, both groups answered questions on how experiences would impact their shopping experience and solve for their online size and fit problems.
2. Moderated study with 8 shoppers We discussed shoppers online size and fit problems with the brand, captured their thoughts on both prototypes, and answered questions on how it would impact their shopping experience if implemented in the brand’s site.

Once research concluded we learned:

• By minimizing the tech jargon behind the product, positioning the experience as a ‘Fit Experience’ or through its main value prop (‘Find your fit’) it helped solved for previously found hesitations in product.
• Shoppers would best engage with the digital changing room closer down the funnel at the PDP level, this is a moment where they’re done comparing styles and are now assessing size and fit.
• When surveyed, 78.9% of engaged shoppers said they would try it to find their right fit and 62.2% claimed experience would give them more confidence when making an online purchase.

Testing conversion through app-only pilots

After discovery and validation rounds, we were confident in the product and its value for lululemon guests. Multiple SteerCo presentations to Senior Leadership later, we got the ok to move ahead with testing the experience live in the lululemon app.

The main goals of these pilots were to:

To bring this experience to life, we partnered with French startup Veesual. They had built a digital changing room experience called "mix and match". To build it they used their proprietary LLM to read sizing information detailed in product techpacks, LLM would then size photo-studio shot images of those products and ‘dress’ models according to the meassurements of the clothes in the model's size.

Internally, we partnered with every team involved in the production of the campaigns selected for tests – Brand, Product, UX, Ecomm, 3D Product Development and Size & Fit, Photo Studio & Casting, Sample Ops – with my team leading the charge at making this a truly cross-functional effort.

8+ months of production after, plus some months in-between each campaign, led us to launch 2 different digital change room experiences in the app:

Pilot Results & Product Impact

After sunsetting both versions, model-led changing room shoppers had outpaced non-model-led changing room shoppers – a shopper shopping through our experience or visiting it first when shopping for the same styles converted at a higher rate during the weeks our experience was live.

At high-level, the combined results were:

Based from pilot results, at a modest ATC of 0.5% experience could bring in a bit over USD$ 2MM/year but the cost of producing it (which covered costs like the digitization of samples, model licenses, and SaaS fees for vendor) was sitting at ~USD$ 900K for the first year if brand were to digitize 15% of assortment.

In the end, productization wasn’t recommended unless there was a 1) a way to bring down the costs of shipping and shooting samples through using ghostpacks 2) make a more strategic decision on what to digitize to reap the rewards of product (e.g., core franchises only instead of every yearly seasonal styles).

Although experience wasn't productized, its work was the foundation of current live experiences in-app like the Matching Set Builder and learnings have influenced product sizing development at the brand.

Collaborators & Credits

The project was led while at lululemon as part of their Digital Innovation team, Future Experiences.

Project couldn’t have happened without the support of teams within Product & App (Natasha Kassam), UX (Najla Sekariyanti, Kat Wearing), Ecomm (Alex Williams), Campaigns (Jeane Boni), Global Brand & Sample Ops (Blair Lawton, Taryn Day), Product Development and Size & Fit (Leith Irvine), Tech & Dev (Vincent Cepeda), Analytics (Liona To) and leadership buy-in from Danny Ryder (Chief Digital Officer), Nick Engel (SVP Digital), and Sarah Rogowsky (former VP Digital Strategy & UX).

Within my team at Future Experiences, I was supported by Heather Kaiserani (Program Manager) and Ellisa Calder (Director, Digital Insights & Future Experiences) who provided their guidance while I was leading this work.

Thank you to the Veesual team (Maxime Patte, Célia Bugniot, Damien Meurisse) for being such great partners on this endeavor.