Data, experimentation and innovation

All businesses face the challenge of managing hugely increasing volumes of data, ranging from usage data derived from their systems and devices, to secondary data sources such as social media or customer reviews. Most organisations recognise that leveraging data will help them run their operations more effectively and productively and respond faster and smarter to their customers’ needs and preferences. But data driven experiments can also support innovation, and I read with interest a recent Harvard Business Review article which explains how businesses are using this approach to pilot new products and to innovate more effectively.

The article explains how data-driven decision making is about developing algorithms that will support smart, timely decisions based on data analysis.  Uber used this approach to launch its Express Pool ride-share service, where passengers are asked to walk short distances to meet their rides. Express Pool was launched in six randomly selected cities, and the resulting metrics were compared with the metrics for a synthetic “control group”, which was a weighted combination of metrics from other cities that were not part of the pilot group. This methodology, which is not limited to tech companies or online services, measures the true impact of a new product or service. In addition to producing and comparing usage figures, it highlights whether the new offering is crowding out existing services or attracting new customers, or more business from existing customers. These insights can help indicate whether products are succeeding or failing and whether a new product adds value to the overall portfolio.

This type of data-driven experimentation can also help manage some of the risks associated with innovation, as it validates decisions – for example if an algorithm has analysed several years of purchasing data together with customer usage patterns and feedback, it will represent genuine, real-time “know your customer” insights. As ever, there are challenges, among them: ensuring data is complete and relevant (and therefore not misleading), hiring/training people with data skills and, perhaps most challenging of all, developing an organisational culture that embraces the opportunities that data analysis and experimentation can bring to improve processes and product offerings.