Maryam Mahdaviani thinks the on-line shopping experience leaves much to be desired. So, with its staff of four, her Vancouver-based Internet start-up is developing an intelligent platform to improve e-commerce for retailers and their customers.
Born in Tehran, Iran, and now living in North Vancouver, Mahdaviani earned a master’s degree in computer science from the University of British Columbia in 2007. During the 27-year-old entrepreneur’s time at UBC, she studied machine learning and artificial intelligence, and served as president of the Computer Science Graduate Student Association for one year. It’s also where she met Jan Ulrich, with whom she founded Optemo Technologies in 2008.
In May 2009, Optemo launched LaserPrinterHub.com, a Web site that helps users find, compare, and buy printers. Then, on October 19, Optemo began a commercial trial of its Discovery Browser with Best Buy Canada. Shoppers in the digital-cameras section of the consumer-electronics retailer’s site are now using the platform to narrow down their options.
Mahdaviani stopped by the Georgia Straight offices for an interview.
How did Optemo Technologies come about?
Jan and I both were on-line shoppers, and we realized that things are not working the way they should be. So, there are problems. There’s a lot of information available on-line. However, it’s hard to access them.
And then we were doing sort of similar research. We were working on multi-document summarization—that’s Jan—and I was working on understanding the user behaviour and preferences automatically. So, we thought that with the state of the research that we were doing—we also published papers—with this research, we could make a contribution here. That’s how we started the work.
What does your Discovery Browser do to on-line shopping?
Discovery Browser is software that helps retailers to manage their catalogue better and organize things better, and helps customers to navigate catalogues better, save time, make their on-line visits more efficient, and find what they’re actually looking for. They’re telling us already that customers feel more in control when they’re using Discovery Browser. We’re also helping retailers understand their customers better and their needs, because we’re providing reports and analytics to them.
How does using your Discovery Browser on an on-line shopping site differ from using, say, Amazon’s recommendations?
Amazon’s recommendation is mainly based on categorizing an individual shopper with a large group of shoppers. That sometimes is good but sometimes becomes a bit ridiculous. For example, I bought a present for my niece when she was an infant, but right now I still am getting recommendations about, you know, infant presents—that sort of stuff. However, I don’t have a baby; my niece is older.
Also, Discovery Browser is about personalizing the entire shopping process and not just giving recommendations. So, what we are working on currently is how to adapt the navigation process based on users’ preferences. For example—the way your preferences are and the type of product you are looking for—we help you find the products faster and easier through the navigation process and the management of information. Whereas Amazon and other retailers just use the personalization to recommend products once you’re down to one single product. So, it is a bigger picture.
How does your Discovery Browser incorporate machine learning?
Machine learning, in one way, is used to automatically group and organize products. So, the products are not just organized now based on their single specification—for example, for the case of digital cameras, resolution or optical zoom. They are organized in a more natural way. Different types of cameras, it’s more natural to the shopper. Rather than just saying “high resolution” or “low resolution”, we have sort of “medium range”, so based on their lifestyle or different functionality that each shopper is looking for. Also, generating descriptions, like easy-to-understand terms that are describing products, that’s a more natural way of describing products.
The other way the machine learning is used is automatically getting users’ preferences—customers’ preferences—and using these preferences to change the navigation and adapt the navigation process. So, basically, this is an intelligent software that learns what you’re looking for, what your needs are, and helps you find the product and information easier.
What are your future plans for Optemo?
Optemo is still developing the Discovery Browser, making it smarter, more scalable. Besides development, we are talking with our future customers. We are planning a meeting with Target and eBay. So, our plans are basically to approach customers in Canada, in North America, and hopefully later internationally. The plan is basically to help all these retailers to reach their customers better and provide a better way of navigation and on-line shopping search for their customers.
Will the Semantic Web factor into your work in any way?
The way that we are doing the sort of search by look and feel, it is sort of a Semantic Web. Also the way that we are describing products based on easy-to-understand terms and generating these natural languages for describing the products, it’s also part of the Semantic Web. So, it is as next-generation as some other start-ups that are involved in Semantic Web. This is the next generation of on-line shopping.
Every Friday, Geek Speak catches up with someone in Vancouver’s technology sector, video-game industry, or social-media scene. Who should we interview next? Tell Stephen Hui on Twitter at twitter.com/stephenhui.