r marketing activities in real time, but that’s really only possible because we are a 100 per cent digital business and because of our investment in the data platform. We spent six months building a fully integrated data and reporting environment from the ground up, and that has given us the ability to make tactical and strategic decisions based on what customers are doing right now. Where previously a Business Intelligence report with yesterday’s performance was sufficient to inform our business stakeholders, we now have a true net position including consolidated digital marketing metrics, sales, and detailed business costs accurate to within 15 minutes. Where we see customer behaviour or engagement changing, our market team alters content and targeted cohorts to optimise the customer experience. IR: Can you give an example of how data has informed the launch of a successful product at HiSmile? WM: Cohort analysis and marketing efficiency are crucial elements of our operation. When we prepare for a major sale, pre-sale activity including teasers and other marketing activity is driven by what customers have responded to in the past, what types of customers have engaged (or not), and also by what their behaviour is right now.  Innovative campaigns and trying new concepts are key parts of the HiSmile ethos, and we measure the effect of new ideas immediately and in detail in order to tweak and tune our approach on the fly. In back-of-house, we measure and respond to emerging trends in customer support interactions to alert us to issues like delivery delays in different geographies. IR:What do you think is the most reliable source of customer data? WM: By far the most reliable source of data is your first party data. Changes in browser security and data sharing restrictions are rapidly rendering second and third party information invalid, or at best questionable. Some providers are aiming to mitigate privacy changes by using server to server integration, but that too has limitations for depth of data.  Our experience is that your best opportunity to form a solid understanding of your customer through data is to invest in a 1:1 relationship with them. Targeted and measured campaigns, accounts and subscriptions, support interactions, and analysis of buying behaviour at an individual level allow you to form a mutually beneficial relationship where your offers are relevant and timely, and where your customer doesn’t suffer campaign fatigue or over-exposure. IR: Are customers still wary about sharing data, do you think? WM: I think media coverage related to privacy breaches is slowly educating people about the risks of being too forthcoming with information about their lives, but most people are yet to join the dots in terms of where their data goes. Arguably, you no longer even have a choice about sharing your data. What your smartphone (and therefore its manufacturer) knows about you is genuinely concerning. With card and contactless payment becoming the norm, your bank knows your life in excruciating detail. Even your single sign on with Facebook or Google or Microsoft builds a deep pool of information about you. Collectively those things mean that just by going about your daily routine, you knowingly or otherwise, fill someone’s pocket with data. Despite all that, few people bother to read long and detailed privacy statements when they sign up to a service and so the burden of protection falls upon regulatory frameworks. IR: Can you tell me a bit about your journey to this career? WM: Data science as a business function is relatively new but the core concepts used in data science have been around and applied to business problems for many years. Data science involves software programming, mathematical analysis, business/Business Intelligence analysis, and effective communication of information to business stakeholders. It’s the effective combination of all four of those skills that make true data scientists unique and sought after.  I’ve spent most of my career in the software and technology space, and as my roles have grown I’ve become increasingly focused on – and conscious of – how my team’s performance influences the trajectory of a business. That business and leadership experience and an educational background with a strong bias towards mathematical analysis (engineering) has allowed me to move into the data science space. That said, I don’t consider myself a data scientist. I just really enjoy analysing and understanding what makes our customers behave the way they do. There is nothing in this world more interesting than the people around you! Dulcie de Koning, PepsiCo commercial strategy director Dulcie de Koning, PepsiCo Inside Retail: How important is data in shaping the customer experience at PepsiCo? Dulcie de Koning: Understanding the shopper should be at the centre of any strategy as data is essential to influence the consumer experience. There is often a discrepancy between what shoppers say they do, and what they actually do in-store when you compare the results of focus groups against data-led insights. Therefore, having access to correct shopper behaviour will inform the strategy to enhance the shopper experience. IR: Is data informing the launch of new products at PepsiCo? DdK:Over the last few years, we have sought to use data, both quantitative and qualitative, to understand what the changing behaviours of our consumers are and how our snacking portfolio can best meet their needs. There is a segment of the population that loves snacking but are looking for a permission point to feel better about their snacking choice and are on the search for snacks that are ‘better for you’. An example of how we innovated to meet growing demand is the launch of Smith’s Poppables in February this year. Smith’s Poppables have the potato taste and crunch but are light and airy providing a different snacking experience to our regular potato chips.   Additionally, we have recently introduced a product that has seen great success in other markets to Australia – Smith’s Baked. It still has the same great taste and flavour that Smith’s is known for, but 50 per cent less fat than our regular crinkle cut potato chips – to offer even more choice to our consumers. IR: What do you think is the most reliable source of customer data?  DdK: The primary default for a reliable data source is Scanned Till. However, as we know, this only explains the ‘what’ and not the ‘why’. For a truly holistic understanding of their shoppers, organisations need to have programs in place to understand the shopper: the path to purchase, who they are, propensity to spend, and what was the trigger point for purchasing at that particular moment in time. IR: Are retailers still wary about sharing data with suppliers? DdK: Retailers recognise the value of sharing a balance of scan data and shopper data where available, as they know these insights are beneficial to develop category growth strategies for FMCG companies.