Retailers are under pressure from many areas: consumers demand more choice and flexibility, and buying habits change quickly, plus increasing cost bases and competition from new entrants. To ensure long-term sustainability, innovation is key, but the question is how to innovate in an effective and sustainable way?
Over the past few years, we have seen a number of retailers responding to this challenge; take Sainsbury’s , Marks & Spencer , John Lewis , Kohl’s, Sears and Sephora, who are investing in “digital innovation labs”. This comes as no surprise: retail’s obsession with seamlessly connecting brick and mortar stores with web and digital to improve the customer experience reflects the data-driven innovation revolution.
It’s an imperative today for retailers to innovate through experimentation to remain relevant in the eyes of the shopper. Yet for many, bringing data-driven innovation into businesses can be easier said than done. The Telegraph mentioned that “The future belongs to data-driven retailers” but estimated only 20% of data coming from retailers is even considered for analysis. So, the challenge is turning more of that data into real-time actionable insights for decision makers across the business.
Today the multi-channel shopper is in charge! They are empowered to interact with retailers using any channel at any time of their choosing and are demanding greater levels of service and personalisation than ever before. If an experience is unsatisfactory, they will broadcast their feelings on social media for all to read and will switch allegiance mercilessly to get the best deal.
So, what can retailers do to survive in this brave new world? Well, one thing that the modern shopper leaves behind is a vast data footprint. The insights gained from mining this Big Data can be used to understand, predict and target customers, thereby delivering a major competitive advantage in the battleground for winning market share.
Every few months the retail industry uncovers a new cool technology, Oracle’s chatbot innovation reaching beyond text-based bots (for example, on Facebook Messenger) to incorporate intelligent responsiveness on voice-driven systems like Amazon Alexaor, meaning businesses need to constantly explore opportunities and alternatives to keep ahead of the competition. This can create tremendous time and cost pressures, along with risks to business-as-usual operations, which often leads to innovation taking a back-seat.
The solution to these problems is to use some battle-tested principles around experimentation:
- Create an ecosystem which enables experimentation with data in a non-disruptive and cost-effective manner. This requires the ecosystem to be flexible, in order to integrate additional data points, new algorithms, and scale-up to manage more extensive computations (think cloud!).
- Use principles of agile to design and run experiments. While data science is clearly distinct from product/software development, there are quite a few effective principles which can be carried over – concepts of sprints, minimal viable solutions, and market testing – which save a lot of pain in the experimentation process. Apart from being efficient, agile also enables early validation with limited budgets, before making any large commercial decisions on the initiative.
- Embrace open-source like Python and R (and contribute back!). It’s now widely acknowledged that to meaningfully tackle a field such as AI, the data science community needs to come together and break down barriers to collaboration. This has resulted in the big power-houses (Facebook, Google, Microsoft, and Amazon) and academia flooding the internet with research, codes, and frameworks for the latest breakthroughs in machine learning. Retailers should look to leverage this in their experiments, reducing the need for heavy in-house investment in proprietary technologies, or development from scratch.
- Time- and cost -box your experiments. While some initiatives will certainly be green-field with longer time horizons, it’s important to segregate R&D efforts from rapid experimentation. Identifying areas for innovation which address immediate business needs will respond to management looking for quick wins. And to build momentum for innovation, your initiatives must show ROI: building measurement in at the planning stage will help fuel more ambitious and longer terms goals.
Being unsure how to leverage data to solve a complex business problem shouldn’t be a barrier to experimentation. By taking advantage of external specialists with advanced analytical assets, open source technology, and sector experts, risks, time and costs are minimised, and delivers value through innovation at scale.
Ian Jarvis is head of retail at The Smart Cube.