Technically Speaking

The Official Bigstep Blog

Reacting Faster in Retail with Real-Time Big Data

Consider the following scenario: You are a supermarket store manager. Popcorn, pretzels, and the like seem to be suddenly disappearing from your shelfs, but you have no idea why. You know you need to leverage this popcorn crisis somehow. You can order more, and have it shipped quickly, but how much stock should you get for popcorn? Why is this happening now? Will it last? To understand the cause of such episodes, one needs to be able to quickly compare the current situation with the same day last year, last months, last weeks and consider events around the shop. Consider festivals, movies, political events and weather. Maybe a new Game of Thrones episode was released.

When top line is no longer the most important growth factor, the bottom line becomes crucial. A relatively untapped source of growth is the time dimension. Data loses part of its value when decisions are not made in the right time window. And these time windows can be very small. Reacting faster, better, offering faster services or well stocked stores all add up.

Optimizing checkout wait times

Kroger uses real-time predictive analytics to estimate when shoppers are likely to reach the checkout lanes. This helps them optimize the number of opened lanes to keep wait times low. Since the technology was implemented, average wait time dropped from 4 minutes to 26 seconds.

The way the system works is by tracking entries into the store using infrared cameras and then using that information as well as historical data to estimate the likely time the new shoppers will reach the checkout lanes.

Making decisions in real-time

One of the largest European hypermarket chains has setup a central Data Lake and is pushing all point of sale transactions immediately after they have been scanned by the cashier across all the stores.

The store managers can watch a dashboard the sale happening and identify which cashier performs best right now, which one struggles and might need help and can redirect help.

fig 1. Realtime cashier shows top performer and strugglers

This kind of tactical awareness was not available in the field until statistical databases and means to query hundreds of billions of records with 98% accuracy.

Fig 2. Data is streamed in real time from POS into a Data Lake where executives can slice and dice it as it flows in versus historical data. The power comes from being able to compare with past events.

Decision makers need to be able to tell if a new trend is an anomaly or has happened before around the same date. This allows them to improve stock and employee efficiency.

Helping customers faster

71% of consumers who indicated that they prefer going to physical store locations for their shopping needs ranked the ability to experience the merchandise in person as one of the top three reasons for doing so. 41% selected it as the number one reason according to Forrester. 1*

Macy’s uses real time technology to help accelerate the sales process. iPad equipped “smart fitting rooms” allow customers to request additional items or different sizes, which sales associates then deliver to them directly in the fitting room.

Real-time, slice-and dice analytics on point of sale data (POS) integrated with historic transactions, data warehouse and ERP systems provide a lot more than just graphs to executives. This real time data can be fed back to iPads of sales associates so that they know what’s hot right now and help customers proactively with that specific product.

Conclusion

Kroger’s score on customer-satisfaction surveys has improved 42% on the speed of checkout. Earnings were up 9.6% in the three months ending May 25, 2013, compared to the same time a year ago, much of it attributed to this initiative.

An important hidden aspect here is that Kroger’s growth during that period was not driven by new stores but rather by operational efficiency. Only 1.5% new stores were opened during that period. There is a point in any retailer’s growth where adding new stores will in fact cannibalize sales in the existing stores. Optimizations are essential if the business is to maximize shareholder returns as well as win loyal customers from other retailers.

The difference between conscripts and the navy seals is not superhuman abilities but their decision-making abilities and well-tuned reactions to what happens around them.

Reacting faster is not just about automation but especially about business agility and the retailer’s staff ability to adapt quickly, to changing circumstances. Automation must augment the human intuition and help it make better, smarter decisions, faster.

1*) Real - Time Data Drives The Future Of Retail, January 2016, Forrester

For more details, please visit bigstep.com

Got a question? Need advice? We're just one click away.
Sharing is caring:TwitterFacebookGoogle+PinterestEmail

Readers also enjoyed:

3 Predictions for Hadoop & Big Data Between Now and 2021

The past few years have brought some growing pains for big data and Hadoop. They had to fight the reputation that they were just buzzwords and had nothing…

3 Real-Live Big Data Success Stories That Prove You Can Make It Happen

Have you been sitting on the sidelines, waiting to see whether others achieved that promised success with big data, or instead fell flat on their faces?…

Leave a Reply

Your email address will not be published.