3 Steps to Building a Retail Big Data Program by Next Holiday Season

by   |   January 27, 2014 6:59 pm   |   2 Comments

 3 Steps to Building a Retail Big Data Program by Next Holiday Season

Joshua Siegel of EMC Professional Services

Retailers who didn’t invest in big data capabilities during the last holiday season missed out on understanding a lot about their customer base: buyer motivation, the kind of data shoppers share, targeted mobile couponing and analytics-based loss reduction, among other things.  But is it too late to have something in place by the next season?

For the many retailers that are not Wal-Mart or Amazon, it’s not too late to implement an analytics program that captures this kind of data. The key to meeting the Black Friday of 2014 deadline hinges on three steps: Define what big data means to you. Be specific about the benefits your company should expect. And identify implementation priorities.

If done right, retailers of any size can redefine their marketing efforts by capturing and analyzing data from shoppers as they use mobile devices to compare prices, get directions with geo-location apps, share pictures of merchandise with friends and pay for their purchases.

Related Stories

Podcast: How analytics can help retailers put goods where buyers want, and when.
Read the story »

Redbox finds new kiosk sites by tracking DVD rental behavior data.
Read the story »

McDonald’s moves to franchise-level analytics to elevate performance.
Read the story »

Focus On: Retail Analytics
Read the story »

Black Friday mobile sales in 2013 grew 43 percent year-over-year to reach 21.8 percent of the approximately $1.2 billion in total online sales, and close to 40 percent of all online traffic on Black Friday came from smartphones or tablets, according to studies cited by TechCrunch.  These online and mobile purchases produced a rich and vastly different amount of data for retailers than an in-store purchase.  Even if consumers are not buying online, they are searching, comparing, gathering opinions and finding their way to the store online – all data that can be helpful to the offline retailer.

Many retailers, especially small and mid-size retailers (less than $21 million in revenue), understand the new reality surrounding Big Data, but are unable or unwilling to respond to it.  They may be overwhelmed by the prospect of capturing and tracking data or not convinced that doing so will make a measurable impact.

However, any size retailer still has time to design and deploy targeted data analysis programs and reap the rewards next holiday season.  Gaining insight into customer behavior and buying patterns need not be complex or costly – and the benefits are extensive.

1. Define what Big Data means to You

Retailers typically have access to reporting generated from point-of-sale terminals, website traffic and inventory management systems, among other sources.  Big data involves getting access to a deeper level of transactional or operational level data, that has been previously unavailable (such as mobile application data or social media) or unusable (due to system limitations or data incompatibility).  The “big” in “big data” can in fact be misleading.  Useful retail data can come in four forms:

  • Deep data: offering more detail on customer transactions such as time of day, SKUs purchased and previous similar purchases
  • Fast data: offering quicker insights and analytics from customer expressions of interest and actual purchases so that more precise and timely offers may be sent to customers
  • Integrated data: allowing a more complete view of your customers’ cross-channel behavior, such as when a customer views a product in a showroom and then buys online.
  • Traditional big data: incorporating more data points from sources such as mobile devices, mobile application, social media and advertising partners.

Defining the term big data is not as important as defining an objective for getting access to the available data.  Smaller retailers especially worried about the scope or cost of a data analysis program should remember that the objective determines the data you need.

2. Be Specific About What Benefit to Expect

Retailers’ interactions with customers can be in-person, through the Web, mobile phones or through social media.  Pulling data from these interactions can help you understand the desired benefit:

  • Customer behavior such as how a customer gets advice before they purchase, how a customer learns about a retailer’s products, the purchase cycle – how long it takes, how consumers buy.
  • Customer motivation to determine if customers are looking for the best price or a specific set of requirements. Are they a value shopper? Are they buying a gift? Are they responding to a life event?
  • Customer psychology to understand the consumer’s mindset.  What is the product’s perceived vs. actual value, how should the product be priced?
  • Dynamic customer segmentation to learn what variables affect willingness to buy, such as the operating system or browser if they’re shopping online, type of mobile devices used, the time of day they shopped, and whether or not consumers need advice or are they self-directed during the buying process.

Retailers who build this understanding can use data to address any number of goals, such as:

  • Improving promotions and marketing
  • Increasing the speed of product development (for example, by creating virtual games to engage customers) and product placement
  • Pricing products and product bundles
  • Reducing fraud and return rates
  • Improving word of mouth and reputation
  • Improving internal operations such as inventory and supply chain management.



3. Set Specific Implementation Priorities

While there are many use cases for retail analytics, it’s vitally important for a company new to implementing the technologies to pick a priority and show a timely return on investment.

Choosing a priority should be a function of potential revenue gain or loss avoidance, or some other suitably beneficial financial benchmark.  But also take into consideration how available the data is, how easy implementation will be and how ready your organization is to accept the change that accompanies the solution. The chart below crystallizes the choices a typical enterprise faces as it sets priorities for quick action and identifies strategic initiatives.

Joshua Siegel of EMC square exhibit 380x356 3 Steps to Building a Retail Big Data Program by Next Holiday Season

Moving forward requires retailers—those that are still offline, as well as those with limited online exposure—to enter the world of big data.  With the right focus they can execute and benefit from a data analysis strategy by next holiday season.

Here are five steps to help prioritize:

• Clearly define goals and get alignment. This is especially true for larger institutions, but critical for smaller ones as well: while big data can be hugely effective, it is not a panacea.  Identify a specific business goal such as increasing online cross-sell conversion rates, reducing fraud losses, or improving customer experience.  While Big Data can help with all of these, focusing on just one initially allows a program to gain traction and show results.

• Start with owned data. Look for the data that will help achieve your goals.  Start by taking an inventory of your own data; retailers may find unformatted or siloed data already available that hasn’t been previously leveraged, such as sales from other locations, household data or partner data from credit card companies or banks. Prioritize available data sources based on effectiveness and cost – not on sexiness or cool factor.

• Take risks. Think strategically and realize the data available to those in the retail field has grown and will continue to grow exponentially.  Marketing, fraud loss, co-branding, and sales techniques not available three years ago have peaked and retreated in lieu of still newer ones such as geo-location based marketing offers, social media marketing campaigns and mobile-based biometric authentication.

• Build capabilities incrementally. Once goals and initial data sources are identified, deploy the analytical models to analyze the data for trends and suggest actions.  Off the shelf modeling tools such as Microsoft Excel, Access or SAS are typically sufficient.  Test, learn and develop a way to learn from the data coming in (a “learning loop”).  Add data sources incrementally and monitor success metrics.

• Tackle mobile computing. Regardless of industry, channel of sale, country or customer segment, mobile computing is by far the most disruptive channel retailers have ever known.  It has fundamentally changed prospecting, marketing, servicing and fulfillment.  Retailers, even those to-date exclusively offline, can benefit greatly from this trend and enterprises ignore it at their peril.

An Example of a Prioritized Plan

Step Related actions
Clearly define goals and get alignment A sports equipment retailer with a dozen locations and a website decides it wants to improve its ability to close abandoned online sales.
Start with owned data The retailer inventories its data and determines it can link non-filled shopping cart orders with the orders placed when those customers use their reward card at the point of sale.
Take Risks The retailer decides to test a program to push a notification to its cashiers with a sale price on the item of interest. The order can be placed at the point of sale and shipped free to the customer.
Build capabilities incrementally The retailer tracks success of the offers over a six-month period and determines the program is worthwhile.
Tackle mobile Lessons from the program indicate that linking loyalty program numbers can open the door to further data and data-based marketing efforts.

Holiday season shopping continues to evolve: it’s starting earlier and continuing longer.  More and more shopping is being done online.  It is no longer sufficient for retailers to rely on legacy, offline focused data collection to gain insight into the consumers of today.  In order to truly understand the evolving needs of consumers, those online and cross-channel experiences need to be captured, analyzed and improved.  If retailers can do that, 2014’s holiday season can bring in more black ink than ever.

Joshua Siegel, an EMC Professional Services practice leader, has more than 18 years of consulting experience in the financial services and technology industries.  His practice within EMC focuses on finding big data solutions for EMC’s enterprise customers.    He holds a Bachelors of Science degree from the Wharton School at the University of Pennsylvania, and an MBA from Yale University.



46f42d0e 1554 48a8 bf42 8f2338eedc44 3 Steps to Building a Retail Big Data Program by Next Holiday Season



Tags: ,

2 Comments

  1. Big Data Queen
    Posted January 28, 2014 at 3:50 pm | Permalink

    Joshua, very nice article on Big Data. When considering a big data strategy, I think it’s worth mentioning HPCC Systems from LexisNexis. Designed by data scientists, HPCC Systems is an open source data-intensive supercomputing platform to process and solve Big Data analytical problems and can help companies derive actionable insights from their data.
    HPCC Systems provides proven solutions to handle what are now called Big Data problems, and have been doing so for more than a decade. The main advantages over other alternatives are the real-time delivery of data queries and the extremely powerful ECL language programming model. More info at http://hpccsystems.com

  2. J C Green
    Posted January 29, 2014 at 9:40 am | Permalink

    This article very concisely describes practical and rapid approaches to gain value from the fastest-growing part of retailing today, online sales. Rarely do experts describe a rational approach to begin gaining value from Big Data rather than searching for the Holy Grail. Thank you for sharing important insights.

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>