HipVan: Making Designer Furniture the New Normal
This month we talk to Shobhit Datta, the co-founder of HipVan, a curated e-commerce platform that features design inspirations for everyday living. A leading brand in Singapore, HipVan is considered a one-stop shop for anyone interested in furniture and home decor that’s stylish yet affordable. Currently experimenting with a 10,000 square foot retail space to test out an omnichannel strategy, we discuss the role AI and Computer Vision can play in augmenting that strategy and in creating a seamless shopping experience.
Tell us a bit about HipVan. What do you do, and what is a typical shopper on HipVan like?
We started HipVan in 2013 with the primary goal of making designer furniture more accessible and affordable. Which is why our target market is majorly new home-owners and young, newly married couples. These are people who are not looking for run-of-the-mill, cookie-cutter style of furniture – but people who actually want to express themselves through the furniture they buy. I’d say most of our shoppers are urban individuals who wish to pride themselves in the way their house looks, and want to depart from the regular furniture that you find in most homes.
How do you reach out to your target market?
We usually segment our customers based on their position in the buying funnel. For mid-funnel or bottom of the funnel marketing, we follow retargeting and AdWords. These are people who’ve shown search intent or have browsed for products on the site. For the top of the funnel shoppers, we generally target them through inspirational and lifestyle pieces on Facebook and other social media. This has helped us build a community around design, and around people who care about designer furniture. We also leverage a lot of content that’s targeted towards the creative community.
We focus a lot on the end-to-end customer experience. So whether it’s the on-site browsing and buying experience, or managing the after sales experience through NPS surveys, service and post-sales support, we make sure we put a lot of effort in engaging and retaining our customers.
What were some of the technology challenges you faced that made you explore the AI space?
While we had looked at a lot of technology solutions for customer experience in the past, our biggest challenge was the fact that most of these solutions were “rule based” primarily driven by cohorts or the wisdom of the crowd. For example, recommendations such as “Viewed also viewed”, “Bought also bought”, and so on. We at HipVan are in a space where we prefer to curate the offerings based on the shoppers’ personal styles and preferences – making it more individualized.
While we had looked at a lot of technology solutions for customer experience in the past, our biggest challenge was the fact that most of these solutions were “rule based”
With rule-based recommendations, we never really got the kind of results we were looking for. It didn’t make sense for us to put a large bunch of shoppers into pools and provide the same recommendations which are based solely on buying behavior. But with vue.ai, we are able to leverage the visual attributes like preferred colors and styles making the shopping experience a lot more personalized.
How would you describe the transition to AI?
The transition to AI has definitely helped, since it doesn’t just take historical data around buying behavior but also the visual preferences of each shopper on the site. It makes products more discoverable, and makes the exploration of the products on offer much easier for the shoppers so they can find similar or complementary products without any effort. I’d say it falls in line with our company’s mission – the kind of company we are trying to build, and the kind of shopper experience we want to create on our site. And this is especially important when you consider the number of SKUs on our site, with product discovery being key. We don’t really sell the same run of the mill furniture, and provide a whole variety of options to suit different tastes and preferences.
The transition to AI has definitely helped, since it doesn’t just take historical data around buying behavior but also the visual preferences of each shopper on the site.
Also, considering that we’re selling to shoppers who are building a home, and buying furniture for the whole house, it makes it important to recommend products that complete the look for the living room, the bedroom, and so on. The visually similar and cross-product recommendations make it really easy to do that. It increases the probability of a shopper adding a product to cart, while also increasing the average cart size. The most important differentiator for me, is the fact that it isn’t just based on historical data but is more forward looking. I think that could really be revolutionary.
The most important differentiator for me, is the fact that it isn’t just based on historical data but is more forward looking. I think that could really be revolutionary.
If you had to list 3 things that have worked for you with vue.ai, what would those be?
- Real-time personalization based on style: Vue.ai provides live recommendations based on visual elements like color and style apart from the other data signals based on shopper behavior, unlike some of the other recommendations solutions based on static rules. And this has worked really well for us.
- The breadth of the offerings: We work across different product categories and are also getting into physical retail, so vue.ai solutions like image search help us make product discovery easier under a single umbrella that is HipVan.
- Being partners in growth: For us, this is like a partnership where we get to co-create the products. Mad Street Den tailors its offerings based on our business and our business goals instead of following a “one size fits all” approach. We iterate together, understand what works and what doesn’t to optimize the products. That way, we can see the best results.
For us, this is like a partnership where we get to co-create the products.
You’re also getting into offline retail, and opening physical stores. How do you see vue.ai adding value to you with regards to a phygital approach?
Yes, we’re experimenting with a new showroom, and it’s a new space for us. One thing that we see that might make a huge difference is using image search in the physical stores. We’re quite confident about our products, and this would allow them to replicate the experience of browsing multiple stores before making a purchase – as they browse products in different styles and at different price points. It will also bridge the gap between offline and online stores, as our shoppers will be able to look at similar products to the ones they like, and explore more products in a particular category.
What are some of the other technology initiatives that you have planned as a part of your roadmap to enhance the on-site/in-app shopper experience?
One of the things that’s a priority for us is a way to better illustrate the sizing of the products, giving the shoppers an idea of how big a product they’re browsing really is. For example, if they’re looking at a sofa, they’ll be able to visualize not just its size, but how it would fit in their living room.
As for a long-term plan, we want to be able to automate the process of providing style recommendations for a room. This would be something like an in-house interior designer who’ll be able to recommend furniture and decor which is in-line with your aesthetic, and matches with the existing color of the walls, and more. A lot of this is manually curated right now, but we want it to reach a point where shoppers can just come to the site and get automated design advice for their home interiors. Being able to provide an AI-assisted style guide might be really impactful, and could be a potential game changer. It would be great to work closely with Mad Street Den to see how we can co-create something like that.
Being able to provide an AI-assisted style guide might be really impactful, and could be a potential game changer. It would be great to work closely with Mad Street Den to see how we can co-create something like that.