Marks & Spencer is a popular place to work – too popular, in fact. Each year the company receives 225,000 applications for around 35,000 vacancies. The human resources team can’t sift through this volume by hand. Since they refuse to dump half in the bin – once practised by a bank on the grounds that “No one wants to work with unlucky people” – the only way to do the job is use data automation.
Candidates fill out an online form, inputting qualifications and answering basic interview questions. Then a data-analytics programme filters candidates into tiers of quality. The best candidates are identified using a wide combination of metrics depending on the job. Human resources can then interview the very best. The data-based method is fast – an interview can be booked within 35 minutes of starting the application.
Data analytics allows M&S to have only two human touch-points during an entire recruitment process, saving time and money
The software used by M&S for the task is run by WCN, a specialist in e-recruitment software. WCN head of global sales Samir Khelil says: “The bespoke WCN system allows M&S to have only two human touch-points during the entire process, posting the vacancy and meeting management for an assessment, saving both time and money. Enabling M&S to keep the candidate engaged throughout has proved successful as its candidate experience is rated at an incredible 98 per cent.”
The M&S story is just one of the extraordinary ways in which data is now dominating business processes. Scarcely a single area of business is untouched.
Data analytics is used to stock supermarket shelves. Once upon a time a store manager would look at cabbage sales and make a guess at the number needed for tomorrow. Today, algorithms examine historic cabbage sales data going back years. They factor in sales of other products, promotions, the weather, the day of the week and dozens of other data sources. The result is a daily prediction for cabbages based on facts, not gut instinct.
Blue Yonder does the predictions for the Otto supermarket chain in Germany. It makes five billion forecasts a year for Otto, using 300 million records a week and factoring in 200 variables each time. The forecasts are fed into an automatic ordering system, representing an incredible labour-saving approach. Prediction accuracy rose 40 per cent with Blue Yonder’s system, cutting overstocking by 20 per cent. It’s a tough job to do well. It should come as no surprise that Blue Yonder was founded by the same academics who do the software at the CERN Large Hadron Collider and employs more than 80 PhDs, many of whom worked on particle accelerators such as CERN and Fermilab in Chicago.
The legal profession has been utterly transformed by data analysis. A lucrative stream of income was sifting through documents during litigation. A team of paralegals could take months to examine tens of thousands of bits of paper. The job might include listening to thousands of hours of recorded telephone conversations for snippets of dialogue. Agony for the poor wretch assigned the task.
Today voice recognition software listens to the text, automatically identifying different speakers, highlighting key sentences and even moments of anger. The written documents are scanned by e-disclosure software, which blasts through the task in hours. The data-analytics approach saves cash, time and the sanity of litigation lawyers.
It is simple to find equally productive examples in other fields. There’s a great case study, which commuters will love, on how Southern Railways is using telemetry data crunched by analytics firm Tessella to improve train performance, reducing delays by 63 per cent and cancellations by 66 per cent.
What is perhaps more worthwhile is asking why data analytics is still so underused. A YouGov survey, commissioned by First Data Merchant Solutions, found that 42 per cent of small and medium-sized companies go by “feel” rather than using data analytics to find out the preferences, requirements and dislikes of their shoppers. Almost one fifth of those questioned said they “do their best” to use data to research customers, “but find it difficult to keep track”.
All areas of commerce will hinge on intelligent use of data – for those who get stuck in, the rewards are immense
There are two explanations. The first is that many companies don’t know what software is on the market. This is a pretty reasonable excuse. It’s almost impossible to gauge what services exist or what the products do. For example, did you know that Ordnance Survey now gets only 5 per cent of its £144-million revenue from selling maps? The rest comes from location-based services data provision to the private sector. The OS invested £200 million in boosting its processing power. Not bad for an institution founded in 1791 to help guard the nation from Napoleon. Even the OS admits its obsession with data takes people by surprise.
The other factor is shyness. It’s hard to admit that you’ve fallen behind the times and aren’t comfortable talking about predictive analytics. This is absurd. As the data analytics firms will tell you, it is their job to guide clients through the tricky bits. Ask for help and you’ll get it.
Data is clearly the future of business. All areas of commerce will hinge on intelligent use of data – for those who get stuck in, the rewards are immense.
As Google’s chief economist Hal Varian says: “I keep saying that the sexy job in the next ten years will be statisticians. And I’m not kidding.” He’s right – and it’s vital British businesses are in the vanguard.
