Cloud retail analytics and Big Data: redefining the art of the possible

  • Richard Thelwell
  • December 1, 2014

Cloud retail analytics big data It’s exactly twenty years since one of the most famous experiments in retail analytics. A three-month experiment which in due course led to British supermarket chain Tesco doubling its market share—proof positive that retail analytics delivered value. Ever since, retail analytics has been hot.

And today, retail analytics is getting hotter still—because by leveraging the power of cloud computing, Cloud retail analytics brings retail analytics within the reach of every retailer, and without compromise.

Simply put, just as twenty years ago, Cloud retail analytics is today pushing the boundaries even further.

Cloud retail analytics: customer insights, delivered

The three-month experiment in question was carried out by retail analysts Dunnhumby, which was then a small husband-and-wife business. It directly led to the setting up of Tesco’s well-known loyalty card—the Tesco Clubcard, which the retailer uses to capture retail data linked to the demographic data that it captures on every customer.

It also led to one of the best-known quotes in retail analytics, as a stunned Tesco board of directors listened to Dunnhumby’s presentation.

“What scares me about this is that you know more about my customers after three months than I know after 30 years,” Tesco’s then chairman, Lord MacLaurin, reportedly told the husband and wife team.

The rest, as they say, is history.

Cloud retail analytics: scoping the challenge

What’s less well-known about the Dunnhumby story is that its great insight was to use statistical sampling techniques to study just a tiny proportion of overall sales, and yet still be able to claim high rates of accuracy. Trying to analyse all the data was too expensive—even for a giant such as Tesco.

Which remains a problem today. Point-of-sales (POS) data is simply too voluminous for many retailers to meaningfully analyse.

And while samples can deliver startlingly accurate projections, not every business—especially small to mid-sized retailers—possesses employees with Ph.Ds in data science or statistics. Which, frankly, is what’s required in order to construct samples that deliver the accuracy required.

For most retailers, the safer option—even if it takes them into the realms of Big Data—is to try to work with all the POS data they have, rather than to try and carve out statistically valid samples from within it.

Cloud retail analytics: Goes here.

The problem, of course, is the practicality of this.

Try storing data in the vast Big Data volumes required, for instance. Try extracting it from the multiple legacy POS systems that many retailers seem to have. Try analysing it within meaningful timescales, on the available hardware that is running in-house.

Simply put, it’s not easy.

Enter Cloud retail analytics. Which at a stroke leverages the Cloud to deal with the scale and complexity of the data extraction and storage challenge. And also uses the Cloud to throw enough computer horsepower at the data in order to deliver meaningful results in a meaningful timescale.

And does so not by requiring a significant upfront capital investment in data warehousing and analytics software and hardware, but on ‘pay as you go’ subscription basis.

In fact, with Matillion Cloud retail analytics, the only upfront expenditure is a one off fixed-price setup fee, which delivers a full enterprise-class Cloud retail analytics solution, paid for on a monthly subscription basis, with no further hardware or software required.

And implementation time? Well, a typical implementation of Matillion BI can take as little as four to eight weeks elapsed, from signed order to users being trained.

Cloud retail analytics in action

Where’s the proof point, you might well ask. As, doubtless, Tesco’s board did twenty years ago.

Well, consider this. Matillion’s most recent Cloud retail analytics customer is firmly in the realms of Big Data.

600 million rows of data, growing at a million rows a week. 350 stores. Five years of sales history.

And the implementation time? Six weeks.

To find out more about benefits of Cloud retail analytics, download our free E-book today