Big data saves lives. A study in Ontario, Canada, applied real-time big data analysis to all the variables associated with a premature baby (around 1,250 data points per second). The system was able to alert staff to the onset of potentially deadly infections before experienced doctors were able to detect any symptoms.
So much of the hype around the potential of big data analysis is deserved. But as anyone familiar with the Gartner Hype Cycle will know, before we can reach the Slope of Enlightenment or even the Plateau of Productivity we must first deal with the Peak of Inflated Expectations and the Trough of Disillusionment. A recent study by Accenture found that only 22 per cent of companies are very satisfied with their analytics programmes, and 34 per cent are dissatisfied. So while companies have accumulated large data sets and invested in some analytics capability, they’re not getting the results they want. Perhaps the investments were made under inflated expectations and we are now wedged in the bottom of the Trough of Disillusionment.
To start, let’s frame the conversation. Big data analysis means using the entire data set, or multiple data sets. Most data analysis today uses sampling techniques to deliberately limit the sample size, and then the results are extrapolated to the entire population or data set. Now that technology enables us to look at the whole data set, and we have discovered big data analysis yields many benefits.
The data set we will discuss here is mobile network data. Once anonymised and aggregated to protect individual privacy, this data contains three super-variables: demographics, location, and behaviour. For consumer out-of-home behaviour trends and patterns, this is an empirical data source that is second to none. It has the potential to yield incredible insights for marketers, market researchers, transport companies, town planners, government, and many more.
The benefits for market research and marketing planning are clear. Mobile network data will enable media planners to say “we know” rather than the current “we think” when pitching a media strategy. For example, a car manufacturer may be surprised to learn that while their target demographic for a high-end model worked in central London, they actually live in middle-income commuter areas. So their media agency can deploy outdoor advertising in those areas, avoiding paying the London rates, and raising the effectiveness of the campaign.
So, information on location dwell time and motion maps, overlaid with demographic segment and behaviour variables, will enable outdoor advertising companies and their clients to improve their targeting, refine their rate-cards, and improve ROI.
Today, the organizations responsible for anticipating and understanding changes in consumer behaviour are still working with traditional methods, such as daily/weekly diaries and installing a monitoring app on consumers’ phones. But diary entries may be unreliable and monitoring apps can be switched off.
In contrast, the data that mobile operators’ have within their networks is a remarkable source of empirical consumer insights and market intelligence. Now quantitative big data analysis can augment and in some cases replace the traditional techniques, yielding better insights at lower costs.
To conclude, let’s look at a very current example. Big data analysis on consumer out-of-home behaviour could also help save the ailing high-street shop. Mobile network data can provide greater intelligence into who is coming to the high street and when; how far and how long they've travelled; dwell time, traffic patterns and blockages. By delving into the data, retailers and town planners may well find new ideas and the evidence they need to make investments.
On the flip side, the data can also help online retailers capture more market share from the traditional retailers by improving their understanding of consumers’ show-rooming behaviours.
So big data analysis can definitely save premature babies; could it save the high street retailer?