If you’re conscientious of health and time, you probably know Huel for its quick, nutritious meals. Or maybe you’ve landed on one of their blogs or health guides, or laughed along with the light-hearted tweets or Instagram posts. The business has done an excellent job of building a serious product with a fun brand and a cult following of customers.
Behind the scenes of this direct-to-consumer (D2C) company exists a host of data that the business uses to understand customers and improve their experience. Hueligans (the company’s term for both customers and employees) are getting a great product, a personalised experience and incredible service thanks, in large part, to the data and the individuals who make sense of it.
Enter Ollie Scheers, Ecommerce Director and Jay Kotecha, Data Scientist at Huel. They work closely together to ensure the company follows the first rule of being a Hueligan: Make Customers Happy. Powered by data and the occasional matching outfit, these two work together to create a holistic understanding of the Huel customer journey and use that knowledge to help propel the business forward. In this interview, we unpack what the customer journey means to Huel and how they’re making sure it’s the best it can be.
Tell us a bit about yourselves and what you do.
Ollie: I joined Huel about two and a half years ago, when we were around 30-35 people, as Ecommerce Director. Since joining, we’ve built out our engineering team and content team, brought UX design in-house and over the past year we have started to build out a data science team covering data engineering and analysis. We’re now starting to look at how we can use various aspects of machine learning to help decision-making.
Jay: I joined Ollie’s team nearly two years ago as a Data Scientist. I’ve helped build the data warehouse from scratch. This included bringing together multiple data sources, getting the data into a nice clean format and finally pushing that into Tableau to give the whole company access to data to make decisions – it’s a continual process and we’re always pushing more data to meet our needs. We are now focusing more on future predictions with machine learning.
Why is the customer journey such an important focus for your business?
Ollie: Every Hueligan receives a metal card with our core values written on it when they join and embossed at the very top is “Make Customers Happy.” We obsess over the customer: what they want, what pain points they have and what issues we can resolve for them. As a D2C brand we have access to a huge amount of data across multiple touch points to help us understand their behavior, ensure they’re happy, and keep them coming back and referring others to expand our community of Hueligans. Many of our customers join Huel after hearing about it from their friends, so we need to make sure every single one of our customers has the best possible experience.
Jay: At Huel we have super fans who adore our product, the merch and the community. We want to pay back that loyalty by giving the best experience possible. Each new Hueligan gets a free t-shirt, shaker and a starter booklet about the company to welcome them in. Our customer experience team interacts with our customers on emails and social channels to ensure we maintain that close community.
What blockers do companies tend to face when trying to analyse a customer holistically? Where are the gaps in insights?
Jay: For us, we had data from so many sources: digital marketing platforms, Shopify, Zendesk, Recharge. Because these sources were siloed, we couldn’t see the whole customer journey. Performance marketing metrics were separate from customer success metrics, etc. This is the classic data silo problem. Recently, we adopted the tools we needed to bring that data together and to visualise the customer journey from start to finish. We’re building this out at the moment and beginning to understand things like: if a customer contacts us, how does this impact future purchases? Or, what customer segments have the best retention? These are the types of questions we couldn’t answer but knew we needed to.
Ollie: The classic data silos problem, really. We had a lot of data but it was housed in the individual systems which made relatively simple analysis, let alone mapping the end-to-end customer journey, difficult. We now have all of our data in Snowflake, which has been great for scalability. Now we can easily analyse combined data from all our systems.
How does your current state of customer data compare to the past? How has that understanding deepened?
Jay: Initially we just looked at high-level business metrics, but we needed to be able to view our customer journey in much more detail. We need to know if our customers are happy. So with all of the data we have currently, we were able to create our customer experience dashboard – the most viewed in our company. This helps us view how our customers engage with us and what queries they have so we can help them in the best way possible. We can also see our key metrics like customer satisfaction, our responsiveness and NPS.
So, how do a data scientist and an ecommerce director work together to understand the customer?
Ollie: Jay sits next to me on purpose. While we have made a concerted effort to make it easy for everyone across the business to analyse our data, it’s still useful to be able to turn to Jay and quickly dig into the data in more detail and answer more complex questions together. It’s invaluable. I lean on Jay to help identify the things that lead us to acquire the highest LTV (lifetime value) customers and to understand what we need to do to move customers into the loyal and VIP categories.
Jay: Huel started out with a single product, our Huel Powder. But we have now expanded our range into Bars, Hot & Savoury and Ready-to-Drink Huel. We’re using data science to make sure we target the right people with the right product for their needs. If a product is doing particularly well or certain flavors are resonating with customers, I work with Ollie to push that on the website more and to make sure that customers are choosing the products best for them.
With a modern data stack, what sources can you analyse from when building your customer journey?
Jay: Fivetran allows us to use a huge array of connectors to help us build our stack. We connect all of our customer experience data from Zendesk, our sales data from Shopify, our marketing data from our digital channels and many more. There are also various tools to allow various custom connections like SFTP and AWS. Using all of these combined we can see a very clear picture of our business.
Have you learned anything about your customers since implementing your data stack and building a more holistic view?
Ollie: I think the most obvious impact has been around where we spend our marketing budget and focus our efforts to improve customer experience. We’re investing more in the sources that bring in the highest value customers. With a past ETL solution, we had integrity issues with our Facebook data. It's critical that we trust the data to make educated decisions. Having an understanding of customer pain points in near-real-time is also beneficial. With quick sync times, we can quickly identify and fix any customer experience issues.
Jay: The stack plays a role in everything we do at the moment. A recent example is centralising and analysing our subscription data. We identify VIP customers who pause or cancel their subscription, determine why and reach out to offer a specialised service or determine what we can do to keep them on board. That data updates hourly through transformations in Fivetran so they can do this as often as they want.
Can you share examples of dashboards that help give you a 360-degree view of the customer?
Ollie: The LTV dashboard is probably my favourite. It allows us to filter, slice and dice different customer behaviours and products to see the impact on LTV. If a customer buys a particular product, interacts with a touch point or takes a particular action, I can determine if that makes them higher value than another customer. I can easily see who our highest value customers are and then discuss with the wider team how we can attract more of these customers.
So what’s next? What are your big data aspirations?
Jay: What I am focusing on now is machine learning and future predictions, such as predicting LTV and churn of our customers. We want to get to a state where, as soon as we get a customer in, we can predict how valuable the customer will be to us and the likelihood of churn so we can direct our efforts accordingly. By using machine learning, we’ll be able to offer our customers a better service and more exceptional customer experience as soon as they come through the door.
Ollie: Exactly what Jay said: predictive modeling, but also determining how we can use natural language processing (NLP) to gain more rapid insights. If a customer says anything about Huel, how can we understand the sentiment behind that and act on it quickly? As a larger company goal, we want to continue to improve data literacy across the organisation. Jay runs various drop-in sessions for employees to improve or upskill in data. This way, everyone can help answer business questions and solve issues to make sure we meet that mission to make customers happy.
Want to learn more from ecommerce businesses that are using data to make decisions? Download the ebook.