Pump Up the Profit

With it’s recent close, the last major asset protection conference of the year, NRF LP 2013, highlighted the rapid evolution of this area in retail.  Asset protection (AP) as a field is not only witnessing tremendous technological advancement, but also...

Retailers are constantly analyzing the shopping experience to help consumers, generating extensive amounts of data in the process, but many have not translated this data into actionable intelligence. The key to big data is taking action from the information. Retailers need to make the transition from data repository to business solution. The focus should be on how the data can help improve their business and better serve their customers. The difference is how they leverage the information. Traditionally, the data collected is used to help retailers in many ways, including affecting prices, changing product lines, offering coupons, introducing new products, adding loyalty programs and other business strategies. Leveraging this information, retailers can also reduce variable and fixed costs or even increase sales volume by running promotions, etc. All of these tactics are effective in providing a possible increase in profits. Using the data to create actionable intelligence adds an additional layer to the decision-making by providing businesses with the evidence they need, and the necessary guidance to have the desired impact. Companies can now make decisions in a timely manner to increase sales and productivity, but that’s just the beginning of big data analysis. When it comes to market impact, big data is now being talked about in bold language. It creates an omnipresence for executives, and with information being logged and harnessed faster than ever. With the proper solution, retailers will be able to access all the available information and data that they need; send it to the right person at the right time; more importantly recommend best practice actions, based on the culture and policies of the company; and monitor and track the actions. All of this should be able to be done automatically, easily, and of course, quickly.

At this year’s RILA Asset Protection conference I witnessed how the loss prevention / asset protection department has begun to recognize the need to expand beyond the traditional role of alarms, locks, and metal detectors.  More importantly, the role of the next generation asset protection professional is emerging as a hybrid that can bring together various departments, including operations and merchandising.  With a recognized understanding of adding value beyond traditional LP services, next generation AP professionals are leveraging existing and emerging technologies to bring added value to the company, and increase the return on existing investments. As an example, CCTV has been historically used to track and monitor retail criminal behavior.  The obvious benefit of having multiple “sets of eyes” recording various parts of the store mitigated the need for additional on-site personnel.  However, this technology can now be used to assist with data collection, helping keep track of staffing or merchandising needs and execution.   New life can be brought to CCTV as a way to monitor the employees on the sales floor.  Ensuring there is enough personnel to service customers for sales floor assistance, or checkout at the point of sales. Furthermore, this video data can be correlated with the number of customers coming through the doors at any given time to determine the optimal staff distribution throughout the day as well as conversion rate by department. Even more intriguing is the blend of the data scientist and detective.  Only these next generation data AP professionals have an eye on margin, not just malicious behavior.  It was interesting to hear the ways data AP teams are cross correlating information and using analysis, rather than “gut feelings” to identify root causes impacting the financial health of the retailer. With technology that can bridge the gap between information, and execution.  Data AP professionals utilize algorithms coupled with business logic to make cost reduction but more important top-line growth a new tool in their arsenal. A key component of this sophisticated tool is to not only bring forth opportunities, but also resolutions of questions and issues that may have otherwise never been asked or brought to the retailers attention.  Data AP teams can utilize big data and predictive analytics to draw correlations that will be key to implementing effective process change and ultimately increasing sales and margin. The growth of the new Data AP professional is allowing asset protection to become a profit hub and not just a capital expenditure.  Retailers will be able to rely on them for information and not just criminal deterrence, making them invaluable, and not just the “cost of doing business”.

As we approach the RILA Asset Protection conference, I often think about some of the conversations I have with people in retail.  In particular, was a response I received back from a retailer when I asked “what are you doing to protect your profit margins”.  Their response was not unexpected, but it did prompt me to give some serious thought to the answer, “all of our projects are to improve profit margin.” After more than 40 years in retail, I have come to realize that PROFIT is a byproduct of what we do and not the direct motivator.  Yes, we need PROFIT to stay in business and survive.  However that isn’t what is driving implementation of WMS, ERP and many other systems over the years.  The driver is the competitive nature of Retail and either you stay competitive or you go away. PROFIT is part of the margin calculation and we wouldn’t be implementing anything that would knowingly decrease our PROFITS (although there was a company in the early 1970’s that discovered if you price your products to ‘give them away’ that people would actually ‘come and take them’) but I’ve never heard anyone proclaim ‘we’re going to do this purely for PROFIT growth’. When we set our PROFIT margin we do not realistically expect to realize all of it and have created a number of ‘buckets’ into which we calculate/assume where the unrealized PROFIT is going.  Among these buckets are malicious events (fraud, theft, collusion, etc.) which accounts for about 30% of the unrealized value and these events are both internal and external in cause.  Retailers fund this side of unrealized PROFIT margin to the tune of BILLIONS OF DOLLARS ANNUALLY in an effort to control exposure.  Based on available studies and surveys; we’ve reached a stalemate.  With the investment in cameras, honesty testing, background checks, investigative teams, visual analytics, alarms, security tagging, etc., it is a staggering investment.  I can make the point that the needle of control on this side of unrealized PROFIT is barely moving ‘year to year’ so additional investment struggles to be justified. And, we have the non-malicious events (lost sales, failed processes, compliance lapses, mistakes, poor buying decisions, damage, waste, failed credits, unrealized returns, out of season, date compliance lapses, excessive markdowns, etc.) which accounts for the other 70% of unrealized PROFIT.  The pursuit of controlling, recovering, managing these lost profit opportunities is segmented along the functional areas of the business including Store Operations, Finance, Audit, Warehouse Management, Inventory Control, and so forth.  Each of these areas has their own ‘turf’ and appropriate tools and staff BUT no one owns the whole picture of PROFIT on this side of the discussion.  This has been referred to a ‘multiple versions of the truth’ since each area has its own view, message, and analyze their own data. If we were undertaking a project purely focused on PROFIT, we would devote assets to these two areas and while historically the ‘malicious’ side of unrealized PROFIT has garnered billions of dollars annually in staff, services and products—-the investment on the ‘non-malicious’ side has had little investment.  Largely, this area is allocated to the ‘cost of doing business’ bucket when honestly it should be placed in the ‘cost of not knowing what is going on in the business’ bucket!! Significant PROFIT that goes unrealized on the non-malicious side of the discussion and to date I cannot name one single retail company (worldwide) where one individual is charged with owning what happens to PROFITS in this area.  Is the reason is that PROFIT is not really the focus but the byproduct?  Or perhaps because we are functionally ‘silo’d’ and no one wants to take on the turf wars with these powerful elements of the organization (Finance, Asset Protection, Audit, Warehouse/Logistics Management, Store Operations)?  So the duplication of efforts, processes,

The final edition in our “Tell me something I don’t know!” series examines the last of the three main categories of discoveries that retailers are looking for in a solution.  Unknown, unknowns are the things impacting the retailers business and they don’t know, what they don’t know.  When the retailer is unaware that there is an opportunity, nor where to look for it and what questions to ask, is what creates the most uncertainty in the business.  But it also presents a category with the greatest opportunity for interesting insight into the business. There are several examples of this category, and the key in identifying these unknown issues is through the use of pattern analysis.  Automatically detecting the anomalies in the data will shed light on these previously unidentified issues.  People, processes, or systems out of alignment with the benchmark give off pattern signals that become a starting point. A recent example of an unknown, unknown involved a store’s loyalty program. We came across a situation where a retailer built their own loyalty card program.  Customers would receive points for every dollar spent, and these points could then be used towards purchasing other goods within the store chain. However, there was a gap in terms of how they would account for layaway purchases, and when payments are made to purchase something such as a ‘big ticket’ item. Take for example, a customer would buy a dining room table for $1000 retail and they would put down $500 as the layaway. As the program stood, the customer would receive 500 loyalty points at that time. But when the table is shipped to the house after paying the final $500, they would receive the full 1000 loyalty points at that time, bringing the total to 1500 points for a $1000 purchase. It was also discovered that customers were able to keep loyalty points from returned products. In effect, the extra points equate to “free money” for the customers to use in the stores at the expense of profits for the retailer. Pattern technology was able to align customers’ loyalty points with their purchases.  The unequal correlation between the two were identified immediately. The retailer was able to readjust the program to properly account for layaway and returned purchases, saving them from giving away points, and therefore free money, which was an unknown unknown system opportunity. These unknown/unknowns are always eye openers for retailers when found, and they do exist in any complex organization and more common than most retailers realize. Using accumulated data streams by crawling through the data at the lowest level available, highlights the unknown/unknowns through pattern detections and statistics.  Bringing these descriptive insights and guided actions to the right person in a timely manner to act quick and deliver the maximum value eliminates the unknowns.  As the expression goes, knowing is half the battle…

Continuing our series on the three main categories of discoveries that a solution should provide retailers to help minimize the effort necessary to identify controllable factors that can be translated into action, this edition will break down what is known as a “known/unknown”. Recalling from the previous blog, these categories are determined based on whether or not retailers know there is an opportunity, as well as if they know how this opportunity is used to generate value.  The value is generated by finding and acting on opportunities correctly within their systems, processes, or personnel. A “known/unknown” is when the retailer knows there is something going on, but doesn’t know how to find it; or they have a gut feeling about some problem, but can’t put a finger on specifics. This may be the most frustrating type of discovery for a retailer, since they are aware something (a system, process, or personnel) is not working the way it should be, yet the cause, and therefore the solution, is unknown. Leveraging pattern seeking software and appropriate benchmarks helps the retailer find the root cause of a known opportunity and assign appropriate actions to correct it, or myth-bust its existence, eliminating false positives. One simple example of this occurrence comes from perpetual inventory, or PI adjustments. Let’s take a retailer that runs based on retail accounting and therefore calculates shrink based on retail price. Now let’s assume a store manager has 1,000 units of a product in their system at $50 (retail) a unit, and therefore they have $50,000 of inventory at “retail”. After an inventory count is performed, the on hand perpetual inventory shows 800 units, meaning only $40,000 of inventory. With this PI adjustment, the store’s inventory value shrunk by $10,000, which is counted as shrink for that store, if calculated at “retail”. However, since PI adjustments can typically be done whenever the store manager decides to perform them, he may wait until this high-priced item is marked down by headquarters, to lets say $25. With 1,000 units in the system, the inventory count at this time is believed to be $25,000, and that loss of $25,000 ($50,000-$25,000) is not reflected upon the store manager; it is the headquarters P&L, because it was a headquarter markdown in price from their level. If the manager performs a PI adjustment now, the current 800 units are now worth $20,000. So the store manager’s shrink number was reduced from $10,000 to $5,000, simply because they chose to count their inventory at a different time. It is very easy for a store manager to discover this power to “lower” the store’s shrink. Once this is realized, it is fairly simple to manipulate the numbers even more. What is even more worrisome is that this practice could be considered legal (however unethical), since the store manager is not responsible to count inventory every day. Managers are often informed in advance about markdowns or cost reduction (for retailers doing cost accounting), and can plan PI adjustments accordingly, for their own benefit. The incentive is high to manipulate these counts, especially since bonuses are often calculated with these shrink numbers in mind. Retailer are aware the PI adjustments are being made, so the problem is known, but it looks like the loss is stemming from corporate, as a result of poor markdown strategies and merchandising. However, the real cause is related to missing inventory in individual stores. The same amount of money is being lost, but these “wooden dollars” are simply being shifted around, moving the blame from the store manager up to corporate. Therefore hitting the product P&L side, and not the shrink side, of the equation. The retailer knows something is happening, and the perception is that its happening in the field. But they cannot perform the complex analysis of the reports needed to find out what is actually happening.

As a retailer, you have many choices about which software vendor to choose. For users, one of the biggest demands we see in the market is having the software inform you of something you don’t already know about your business. You don’t need another reporting tool giving you another report with the same numbers. You want something different. Something that fills in the missing pieces of the puzzle. Retailers need a solution that can help minimize the effort necessary to identify controllable factors that can be translated into action. There are 3 categories of discoveries that a solution should provide you. These are “known/knowns,” “known/unknowns,” and “unknowns/unknowns.” These categories are determined based off of whether or not your organization knows there is an opportunity. More importantly, it is how these opportunities are used to generate value, by finding and acting on opportunities correctly within your systems, processes, or personnel. In the first of this series we will start with the known/knowns. These are opportunities that you are usually aware of within your organization. Retailers are able to find them, but it takes time and resources to do so. In most cases, the retailer is not being efficient or effective in identifying these opportunities, and this is where improvements can be made in the “known/known” category. Improving efficiency means the retailer does not need to waste time searching for the known opportunity, sifting through multiple systems and reports to try and identify its source. It also means that you will have everything in one place, and the opportunity will be pushed to the right user rather than a user trying to find opportunities. Pattern recognition software boosts efficiency by accumulating occurrences, automatically alerting the retailer to a problem. If you compare pattern recognition to legacy EBR (Exception Business Reporting) methods you find that there is a significant increase in the quantity of alerts created by EBR. Additionally, many of the alerts end up being false positives and the retailer is forced to waste time and resources tracking them down. Real efficiency and effectiveness is having all the necessary data in a single, easy to use system that will provide “true positive” opportunities. To better understand the “known/known” category, we can look at an example of an opportunity identified through pattern recognition. The retailer had a very lenient return policy, which they were proud of because it helped stress their focus on customer service. In the past, they only focussed on their high return vs. sales rates. The average return rate was about 7% of sales for all products. Whereas it was identified that one product was being returned at a rate of 18% across multiple stores within a region. A pattern was created to analyze returns vs. sales, correlated with returns vs. damages, to minimize false positives and properly direct associates. The pattern allowed the retailer to quickly discover that the entire product batch was damaged. Since the high damage rate was across multiple stores, it was determined the issue stemmed from the manufacturer. This entire issue was found and resolved within a week of go-live with the retailer’s data. The quick response allowed the remaining product to be pulled from the shelves before they were sold and returned, assuring increased customer satisfaction by reducing the returns that would have occurred. Moreover a new batch of the product was expedited and sales were in-line with expectation as well.

Analytics are not a new idea. The tools have been used in business since the mid-1950s. To be sure, there has been an explosion of interest in the topic, but for the first half-century of activity, the way analytics were pursued in most organizations didn’t change that much. Let’s call the initial era Analytics 1.0. This period, which stretched 55 years from 1954 (when UPS initiated the first corporate analytics group) to about 2009, was characterized by the following attributes: Data sources were relatively small and structured, and came from internal sources; Data had to be stored in enterprise warehouses or marts before analysis; The great majority of analytical activity was descriptive analytics, or reporting; Creating analytical models was a “batch” process often requiring several months; Quantitative analysts were segregated from business people and decisions in “back rooms”; Very few organizations “competed on analytics”—for most, analytics were marginal to their strategy. It was in 2010 that the world began to take notice of “big data,” and we’ll have to call that the beginning of Analytics 2.0. Big data analytics were quite different from the 1.0 era in many ways. Data was often externally-sourced, and as the big data term suggests, was either very large or unstructured. The fast flow of data meant that it had to be stored and processed rapidly, often with parallel servers running Hadoop. The overall speed of analysis was much faster. Visual analytics—a form of descriptive analytics—still crowded out predictive and prescriptive techniques. The new generation of quantitative analysts was called “data scientists,” and many were not content with working in the back room. Big data and analytics not only informed internal decisions, but also formed the basis for customer-facing products and processes. Big data, of course, is still a popular concept, and one might think that we’re still in the 2.0 period. However, there is considerable evidence that organizations are entering theAnalytics 3.0 world. It’s an environment that combines the best of 1.0 and 2.0—a blend of big data and traditional analytics that yields insights and offerings with speed and impact. Although it’s early days for this new model, the traits of Analytics 3.0 are already apparent: Organizations are combining large and small volumes of data, internal and external sources, and structured and unstructured formats to yield new insights in predictive and prescriptive models; Analytics are supporting both internal decisions and data-based products and services for customers; The Hadoopalooza continues, but often as a way to provide fast and cheap warehousing or persistence and structuring of data before analysis—we’re entering a post-warehousing world; Faster technologies such as in-database and in-memory analytics are being coupled with “agile” analytical methods and machine learning techniques that produce insights at a much faster rate; Many analytical models are being embedded into operational and decision processes, dramatically increasing their speed and impact; Data scientists, who excel at extracting and structuring data, are working with conventional quantitative analysts who excel at modeling it—the combined teams are doing whatever is necessary to get the analytical job done; Companies are beginning to create “Chief Analytics Officer” roles or equivalent titles to oversee the building of analytical capabilities; Tools that support particular decisions are being pushed to the point of decision-making in highly targeted and mobile “analytical apps;” Analytics are now central to many organizations’ strategies; a survey I recently worked on with Deloitte found that 44% of executives feel that analytics are strongly supporting or driving their companies’ strategies. Even though it hasn’t been long since the advent of Big Data, I believe these attributes add up to a new era.

The first and second editions of this blog mini-series discussed common concerns about today’s business intelligence solutions for retailers. In this final edition, we will focus on the area of the profit amplification solution where the most value is derived from: taking action. Most of the questions that retailers are asking, when it comes to analyzing the vast amounts of data, and finding true value from this big data, involve what to do with an identified opportunity: “Who do I send responsibility to, if not myself? If I am the right person to take action, what actions could I take, and how do I know the ‘correct’ course of action? If responsibility is given elsewhere, how do I track it? How do I know that the actions are actually being performed, and the issue has resolved?” Wouldn’t it be nice to have answers instead of questions? With the proper solution, you should be able to access and analyze all the available information and data that you need; send it to the right person, at the right time; recommend best practice actions based on your own culture; and monitor and track the actions. All of this being done automatically, easily, and quickly. The real value of a solution comes from the actions that are being taken, not necessarily the opportunities that are being identified. So what really makes an opportunity actionable? It must be timely, detailed, and guided. Timely: The opportunity needs to be created and communicated in a timely fashion. It needs to be sent as close as possible to the time it occurred and/or was identified. Opportunities need to be evaluated every day. If you wait a month or a quarter to find the opportunity, you will miss the chance to take an action and to have an actual impact. Most large issues that are identified are only large because of the time that passed and the value that was accumulated from not taking action. If only known about sooner, and acted on, the issue could have been dealt with and prevented (with less resources and effort) while still a small issue. Detailed: The opportunity needs to have as much detail as possible; store, sku, day or week level, as this will allow you to focus your actions on the right entities. By having detailed opportunities, your store visits will be much more valuable. When you visit a store, you are able to focus on the areas that require attention, instead of aimlessly visiting all the areas of all stores. We refer to these efficient and effective store visits as “value visits”. Guided: By guided, we mean two basic things. First, the opportunity needs to be sent to the right person that will be able to take the actions. The last thing you want is for everyone to look at everything and try to figure out what is theirs and what is not. Secondly, the solution needs to contain recommended best practices, so when the assigned person receives it, they will be guided on what actions to take and how to perform them. We now know that actionable opportunities need to be timely, detailed, and guided; but how do we make this a sustainable process to ensure we are continuously improving and bringing value to the business? The automated process of identifying new opportunities, based on the patterns that are already set or known, is referred to as pattern matching. We know what we are looking for, so we let the system do it for us automatically. However, it is very important to continuously look for and recognize patterns in new areas, based on your business knowledge, and then add them to the process of pattern matching. This is referred to as pattern recognition. So in the world of the profit amplification solution, the job of the analyst is not to analyze the same thing over and over again every day; their job is to look for new patterns, new opportunities, and automate them. This is the continuous improvement that creates operational excellence.

In our previous blog, Editor In Chief at Integrated Solutions For Retailers, Matt Pillar discussesThe Fine Line Between BI and BS. We will delve deeper into the discussion, and take a look at what you believe is the “best business intelligence system” and why it still fails to deliver the business with the value, in terms of actual dollars, you expected. Assume your organization has the best BI system for creating any conceivable report or view of the data and is coupled together with real-time alerting, event monitoring and task management.  Regardless of whether this BI solution provides dashboards, worksheets, bar charts; it still boils down to a reporting system as the output. This system generates a report, identifies an issue, and the monitoring tool sends the report to personnel. You then must rely on your employees to have the necessary talent to correctly translate and understand the report’s insight – “What is this trying to ‘tell’ me?”. Assuming the employee has the specialized talent necessary to correctly understand the report; then what? There are many different possible ways to harvest the identified opportunity. She can try to solve the issue herself, or assign it to another person, maybe a group of people. This means that you not only need to rely on the employee’s initial understanding of the report, but also how she understands your culture.  The employee must have the proper business acumen to act on it in a way that makes sense for your organization and their role or the ability to instruct others. Even further, whatever action is taken, you need to monitor the opportunity to ensure it gets done and the value is realized. Even with the best BI system, there are several ways this can go wrong.  These include, but are not limited to the following: Relying on employees to understand the report’s insights; Having employees take the right action, based on your culture and their roles and responsibilities; Whatever action is undertaken, ensuring that it is completed correctly will build on the opportunity and generate real dollars from the insights. Because you cannot guarantee that everyone will interpret the insights the same way and act on them in the right way, you can not guarantee that the company is performing in the most efficient and effective manner.  This can limit the potential dollars generated from correcting the identified opportunity. So what needs to be done to fix this issue, or opportunity? You need a system that can identify the story behind the opportunity and translate it to an automated task that can monitor the completion. These tasks are outlined in clear, simple language. Opportunities should be connected to the “best practice” solution bank, removing the need to rely on “talent” to correctly interpret the actionable opportunities. Even if there are multiple ‘correct’ actions that can be taken, they should be ‘spelled out’ within the task management, including which personnel should be associated with each possible solution. Once these automated best practices are acted on, the system can track the actions until they are completed, allowing the retailer to minimize personnel interpretation and therefore diminished return. With this in mind, retailers can realize the full dollar return from every opportunity identified by the patterns crawling the big data schema.