Just a few decades ago shopping was done by physically going into retail stores or shopping through mail order magazines. Then, the shopping experience was revolutionized with the internet, and always connected mobile phones. Artificial Intelligence is now making significant changes in the way people buy and sell online, from creating more personalized experiences to targeted marketing, crafting tailored messages to be delivered at the right time and through the right channel or AI enabled chatbots to interact with customers at any time of the day.
Ella Hilal, Head of Data Science, Engineering, Revenue and Growth at Shopify
At an upcoming Enterprise Data and AI event on January 6, 2022 Ella Hilal, Head of Data Science, Engineering, Revenue and Growth at Shopify shares how AI and ML are being used to enhance offerings and create new experiences for their merchants as well as share key tips on how you can apply machine learning for anomaly detection and forecasting at scale. In this interview, Ella explains innovative ways Shopify uses AI, ML, and advanced data analytics, the large role that data plays at Shopify, as well as some of the opportunities and challenges AI can present.
What are some innovative ways you’re leveraging advanced data analytics to benefit Shopify?
Ella Hilal: At Shopify, we have over 1.7 million merchants across over 175 countries, with hundreds of millions of consumers shopping at their stores. We’re focused on leveraging the scale of our data to not only empower Shopify, but to create new experiences for our merchants that are impossible without data.
Entrepreneurship can be challenging and many entrepreneurs struggle to get their businesses effectively off the ground as quickly as they hope to. We’re really focused on putting the power of data back into the hands of our merchants. My team specifically builds data-informed products that enable merchants to start and grow their businesses on Shopify. Experiences like personalized onboarding for the merchants as they are starting their business on Shopify, or building an intelligent ad tech platform to help them effectively join Shopify, or even a data-informed business name generator to name their business.
In our daily operations at Shopify, we are also highly data-informed. Some of the ways we’re leveraging advanced analytics is by building an anomaly detection engine that allows us to process over 300,000 metric/segment combinations, while focusing attention on what is important in less than 30 seconds. We are also empowering our teams in all the markets around the world to support merchants through a localized data-informed understanding of the market’s needs.
MORE FOR YOU
How do you identify which problem area(s) to start with for your data analytics and cognitive technology projects?
Ella Hilal: At Shopify, we take a merchant-first approach to identifying problem areas. The problems we’re trying to solve aren’t about how we can build more deep learning models (although we have them in production). The problems we’re focused on solving are removing the barriers to success for our merchants. So first we identify the merchant problem, then we identify how we can use data to solve that problem. We definitely prioritize by analyzing the scale of the problem and the size of the merchant base that it will help. For example, one barrier to success for a lot of entrepreneurs is funding, so we created a product called Shopify Capital that uses machine learning to provide funding to merchants. To date, we’ve extended over $2 billion to merchants through Shopify Capital.
What are some of the unique opportunities you have when it comes to data and AI?
Ella Hilal: The scale of our data provides us with a deep view into the commerce landscape, which enables us to create new experiences for our merchants. Experiences like offering funding to merchants without them having to apply, or using machine learning to categorize billions of products to ensure better product discovery for our merchants. In 2020, we saw the unique opportunity data provided during the pandemic. By tracking, learning from, and putting our insights into action, we were able to steer not only our company, but our 1.7 million merchants through an unprecedented time. We were able to use data to pivot our product strategy and offer experiences like buy online, pick up instore and extended Shopify Capital funding that enabled our merchants to not only survive, but thrive.
Can you share some of the challenges when it comes to AI and ML adoption?
Ella Hilal: When it comes to implementing and scaling AI/ML at Shopify, we take several approaches to ensure easy adoption and avoid challenges from the get-go.
First, we ensure what we’re building is solving a merchant problem and that we have enough data to create a solution.
Second, we need to build a deep understanding of the data available related to the problem space, understand its statistical distribution and its mix over time.
Third, we start simple. If a regression model will solve our merchant’s problem, that’s where we start. This doesn’t mean we avoid building complex models, it just means that we first prove how a baseline algorithm can solve the problem and such models allow for high explainability of the model’s performance. Then, we iterate by building complex models.
These three steps are key in demonstrating the impact that AI/ML can have, which ensures we get stakeholder buy-in. We’ve got more tips that you can check out here.
How do analytics, automation, and AI work together at your company?
Ella Hilal: I think all of them are parts of the same coin. I strongly believe that no one can build an effective AI model without doing a deep analysis on the data to understand the behaviour, the trends, and their changes over time. And as I mentioned, it all comes back to our merchants. Our data team is not focused on using the fanciest technology, but on solving the merchant problem. That’s why we hire full-stack data scientists. We don’t hire specific machine learning specialists or analysts. Our data scientists own solutions from inception to production, and are empowered to use data to find the best possible solution. A lot of times that means using a combination of analytics, automation, and AI. So what that might look like is our data scientists working to investigate the data to truly understand the merchant problem, figuring out how we can use AI or automation to solve that problem, and identifying what our success metrics look like.
How are you navigating privacy, trust, and security concerns around the use of your data? Ella Ella Hilal: Prioritizing the privacy and security of our merchant data is central to how we work and develop at Shopify. We are dedicated to designing all of our products with privacy in mind as a first class citizen and not an afterthought, being transparent about how we collect and use data, and returning the value derived from that data through improvements to our platform’s features and functionality to benefit all merchants.
What are you doing to develop a data literate and AI ready workforce?
Ella Hilal: At Shopify, we like to say we’re data-informed, not data-driven. This means that everyone at Shopify should feel empowered to make decisions based on data. We do this in three ways:
First, we’ve embedded our various data teams within the different areas in the company. We’ve done this so that our product, commercial, and service teams can easily incorporate data in their decision making process, and our data teams have the context to help make those decisions.
Second, we have a centralized and searchable data portal that allows anyone in Shopify to search for any data dashboard or data report to learn or investigate any data question.
Third, we’ve created data literacy programs that teach our non-data employees how to work with data, like how to run a simple SQL query or work in mode.
What AI technologies are you most looking forward to in the coming years?
Ella Hilal: I’m excited about the future democratization of AI. Specifically how we can put these advanced technologies into more hands, whether it be small businesses, philanthropies, etc.. The more people who have access to these technologies, the more innovations we’ll see in how we use AI to improve our experiences in the world.
Ella Hilal digs deeper into these topics at the upcoming Enterprise Data and AI event on January 6, 2022 with a special focus on how Shopify makes data driven decisions and applies machine learning for anomaly detection and forecasting at scale.
