‘Operationalizing’ Data at Retail
By Brad W. Smith, special to RetailWire
Retailers are looking to make better and timelier decisions. For a growing number that has meant looking to provide business users with near real-time data analysis through the use of an enterprise data warehouse.
“Retail wants us to help them to be more affective in mining the large amounts of data and to move it to more exception in dashboard capabilities with the data,” Darryl McDonald, chief marketing officer at data warehouse vendor Teradata, said at the company’s annual user conference in Las Vegas last month. “They still need all the data to do the analysis. But what they are saying is we need to operationalize it, make it more usable by more associates within their enterprise.
“Retail customers are wondering how to handle the impact of RFID, and with some of the pilots, there has been a 10-20x explosion of a single item as it is tracked through the enterprise. We are helping them to understand how much data to keep, how much history to keep, how to then aggregate that (data) and then how to better use that information to run their business.”
McDonald added, “There was a big push in retail to understand their customers, and then there was a big push to have more effective marketing with them; now they are looking for more real-time ways to communicate with their customers. Especially with their best customers, the top 20 percent, how do they treat them differently? Companies are being innovative with those customers.”
As an example, he said Teradata is working with a company now that can push offers when customers enter the store; customers swipe their card at a kiosk and the offers are available right away. Also if a customer is in their car or at home, and is planning a visit to a store, they can enter a number in their cell phones and any valid offers are made available to them.
Discussion Questions: What impact will the move to near real-time data analysis have on retailers, both at the store level and at headquarters? What areas (merchandising, customer marketing, replenishment, labor management, etc.) do you believe will be most greatly impacted by improvements in the analysis of near real-time data?
As retailers focus on becoming more “customer-centric,” they inevitably begin to discuss ways to store, process, analyze and share their data. Data yields insights, there’s no doubt about it, but too many retailers have a hard time translating insights into the decisions that drive sales and loyalty. The retailers leading this charge are implementing Business Rules Management Systems.
Operationalizing data means a major culture shift and I find that in the zeal to provide actionable information closer to the front line, this often gets forgotten or under-estimated.
When I worked for Viewlocity, we got pretty far down the road with a large department store retailer. We showed them all of the alerting capabilities of the solution, and how the exception management created an opportunity for the company to really change how they did business–instead of creating huge reports that merchandisers had to shift through every day, they would have a list of exceptions to deal with. The retailer said, “I get it, I get it. But where are the reports?”
It’s all well and good to want to provide this information, but this is definitely NOT a case of “build it and they will come.” You must spend a lot of time and effort trying to make this information consumable–and actionable–by someone who does not deal with a huge amount of data as a natural part of their job. This particularly applies when you want to bring it into the store.
The mammoth technology forward step that real time data on product availability combined with individual shopper loyalty and buying patterns ironically get retailers–especially grocery–back to the level of personal customer service on which the industry was originally founded. A time when Joe the corner grocer saw Mrs. Smith coming down the street on her regular Wednesday morning visit, bagged up her favorite staple products that he had already put aside and threw in a extra doggie treat because he heard that she just brought a cute terrier puppy. Mrs. Smith was immensely loyal to Joe the Grocer.
With times much more complicated, both in terms of the number of skus and the volume and variety of shoppers and shopping habits, that level of personal service can only be fulfilled through smart, real-time and automated queries of enormous databases. But–if executed the result will be the same–Mrs. Smith will be immensely loyal to Joe the Grocer.
It’s not the ability to have real-time analysis, it’s the ability (or the knowledge) to act on that data that provides the advantage. An obvious and oft-mentioned advantage is the ability to reduce out-of-stocks by automatically “knowing” when stocking levels are too low. This assumes the item had been ordered from the warehouse, delivered to the store, and there’s a stocking person available to act on the need. Without the full chain working correctly, real-time is irrelevant.
Many retail chains don’t plan their store-level tasks very well. Buyers at headquarters sometimes want to change prices and displays giving the stores very short notice, without allowing enough lead time or payroll hours. Real-time data can speed up decision making, but everyone needs to plan out the time needed to implement the decisions.
The retail world is beset with data. The trick is to transform the data into actionable information. Loading real time data into the data warehouse will make the folks at Teradata happy. But it will profoundly confound the transformation of data to action.
As it is, the retail user community is generally underserved with business intelligence. Before getting all absorbed in tracking real-time data, retailers should ask themselves these questions:
1. Are merchants automatically informed about recurring problems such as stock outs, inventory buildups, late shipments, broken assortments? Can, and do, they take action to resolve these problems?
2. Are store operations managers informed about zones of poor productivity, staff shortages? Do they respond powerfully with corrective action?
3. Are planners notified with the operational plan diverges significantly from the actual results? Does this insight stimulate replanning?
4. Are BI users across the enterprise given reasonable response time to queries? Or are they wasting countless hours looking for needles in haystacks?
Agreed, that real time time will expand the opportunities for retailers to be responsive to consumers, resolve problems in supply chain, and staff their stores properly. But until the fundamentals are mastered, real time data will only clog the arteries of decision making.
That’s a real nice sentiment but much harder to do than this technology offers. With the increase in “quick trip” for same day/next day usage, retailers will need to get a little lucky on promoting exactly what the consumer has in their consideration set for the day.
If Ms. Smith is going to the store to make ham and macaroni and cheese for her family that night, and gets a text message on her favorite peanut butter, so what? The next step will need to be more interactive and very flexible on the retail side to give cents off when needed by the consumer, not just weekly specials.
Personal marketing, personal shopping and a personal supply chain. Moving data use closer to the time immediately before and during a transaction certainly has potential benefits to increase retail sales. Retailers need to spend time not only on the technology side, but also on the customer side. Without proper customer education, and even permission from the customer (e.g., swipe the card at the kiosk), customers may become more resistant than receptive to such marketing techniques.
Unfortunately I agree completely with the Teradata analysis but couldn’t be more convinced than ever that the answer is not a data warehouse. The way I always express the retailer’s data needs is that “Retailers don’t ask a thousand different questions, they ask the same questions thousands of times.”
Retailers are not conducting basic research where they are trying to pull knowledge from millions of data points whose relationship is unknown. Retailers understand they need to know how an item fits in its category and who is buying it. Retailers understand they need to know which consumers are their consistent spenders and who are the cherry pickers. And retailers understand that some stores will use fewer labor hours than others or experience greater shrink.
So they do need the answers to all these questions and it is nice to have them in a timely fashion so they can take corrective action. But they don’t need to run inquiries against standalone data warehouses to get them. The answers need to be located in the operating applications that allocate inventory, schedule transportation, and manage labor. Those applications need to include the answers in the decision processes as they operate the business.
There is a huge opportunity to create loyal consumers by customizing offers to them. However, making that possible requires the ability to analyze data in a timely fashion as well as the ability to make decisions using that data in a timely fashion.
One hurdle has been the analytical one: gathering, absorbing, and analyzing the mountain of data to determine critical pieces of information has been a daunting task. No two companies identify the critical pieces of information in the same fashion, making the analytical task a challenge both in terms of doing the analysis and in terms of having people in the company understand the information they are given.
The second hurdle is creating advantage for consumers using that information. Managers need to understand the information, its implications, and its relevance to consumers before making decisions that provide value for consumers. Everyone is moving in the right direction but attention needs to be give to both hurdles before the information is used in a consistent way to create value for consumers and profit for both manufacturers and retailers.
There is a direct correlation between the number of SKUs that a store carries and the need for data intelligence.
Pricing models carry the greatest potential to impact the bottom line.
Outsourcing product and category management to CPM reps leaves money on the table.
The majority of mass merchandisers need to upgrade the caliber of their merchandisers. They do not compare well with the MBAs that the CPMs hire.
With the increased profits that private label can produce the mass merchandiser should be able to justify the expense of competing with the CPMs for the best and the brightest.
Real-time data analysis, particularly at store level, will take some time to have an impact for most retailers. They are barely able to handle the data that they have available to them now.
But, assuming that the enterprise system makes the output easy to understand, assuming that store managers and staff are properly trained on how to make decisions based on what the data says, and assuming that the retailer is already established in the core Category Management principles, this data will be invaluable to the retailer at headquarters and in stores.
In the future, retailers need to link Category Management with local marketing. They need to understand how to compete in each local market (at a lower aggregate than metro areas). This real-time store level data will allow them to get there, but not without a lot of work.
While retail technology has continued to advance in the realm of data capture, storage and dissemination, the sad truth is the logistic constraints of retail grocery will not be able to fully leverage these advances anytime soon.
In order to have more effective replenishment, the whole concept of case pack and pallet-based replenishment would have to be overhauled. The efficiencies which have provided the bulk of the margins for product sales, are also those which will hinder the ability to move to a “scan and replace” model of replenishment.
As replenishment goes, it is still a forecast based activity, with the emphasis being placed on anticipating consumer demand by accurately tracking past sales behavior. The analysis and processing of daily POS information must be done very rapidly, but not necessarily in real time in order to gain full advantage of the value of the information.
I believe that consumer marketing will benefit, especially in the area of item affiliation. For example, if a customer were to buy hot dogs and buns, and the system could determine that they have not purchased mustard for the past 5 weeks, a reminder might be welcomed.
There is much to learn, but we must remember that true change will have to involve the entire logistics ecosystem.
I think Replenishment has the biggest opportunity for improvement, of the list provided.
Merchandising information, in the form of MIX and MAX numbers by SKU by store is rarely, if ever, updated in replenishment systems. Even updating your MIN/MAX data on a weekly basis would be a huge improvement over the current state, and may be enough to consider it “real-time.”