Managing the Category Beyond SKU Rationalization

SKU proliferation has been a fact of life in foodservice much as it has been in other industries in recent years. Proliferation creates considerable pressure throughout the value chain to make tough decisions about SKU assortment across numerous product categories. In foodservice the problem is not shelf space as it is in retail; rather, it is the limited amount of product information that a sales representative can manage in his or her head in order to match the right products with the right customers on a daily basis in real time. As managing assortment has grown more complex, manufacturers and their downstream partners have looked to SKU rationalization to reduce streamline product offerings and manage inventory costs for improved category performance. While SKU rationalization can address these challenges to some extent, it does not get to the core of the problem. The most effective way to improve category performance is to increase demand for products in that category. In turn, the best way to grow demand is to seamlessly match unique customers with the products whose attributes they most highly value. This requires a holistic category management approach, supported by robust data analytics that can take into account the key levers of demand – assortment, promotions, pricing and purchase timing.

The Importance of Collaboration

In foodservice, manufacturers and distributors are the logical partners for a collaborative category management venture. Manufacturers possess deep insights into the product attributes that drive demand for specific customer types, and have a strong understanding of how to manage assortment. On the other side, distributors have the benefit of daily transaction data at a very granular level – what quantities of products in the category are being sold to what locations with what frequency. Combining these insights – ideally through a single integrated data management system able to process inputs from multiple sources and generate insights and actionable recommendations to the relevant decision makers – can create a coherent, unified picture of demand that provides a basis for specific assortment, pricing and promotional activities to grow sales.

Reducing the Guess Factor

A traditional SKU rationalization program may analyze aggregate transaction histories for all the SKUs in a category and mark for elimination some subset of those that occupy the so-called “long tail” – products with sparse data records due to infrequent activity. A typical goal in this regard may be to eliminate 20-25% of all SKUs in the category. The problem with this approach is that without an appropriately detailed level of analytical insight, managers are left to guessing what the resulting effects will be on sales. Transaction frequency is only one variable in presenting a composite picture of demand. For example a certain product may transact on an infrequent basis only, but it may also be a popular niche product with attributes highly valued by major customers. What will the sales impact be of not having this niche product available when a major customer wants to add it to his or her market basket? How can decision makers recognize and differentiate between niche products and other long tail denizens that really deserve to be eliminated from the active product line?

A holistic category management solution, driven by advanced predictive science, can supply answers to these questions. By integrating product attribute knowledge possessed by the manufacturer with quantity and purchase timing data known by the distributor, the system can make recommendations about when to stock the low-frequency but desirable niche items with a higher likelihood of coincidence with the customer’s purchase decision. Techniques such as Hierarchical Bayesian modeling help overcome the analytical challenges typically presented by sparse data. Rather than losing all or part of the customer’s market basket for the sake of an incremental SKU reduction – in most likelihood a losing proposition – the result is retaining a satisfied customer.

Focus on Growing Demand

This approach to category management program shifts attention away from simple cost reduction through inventory rationalization and focuses instead on the revenue side of the equation – growing demand in the category. There are two critical requirements for this to be successful. First, the data management platform must be sufficiently granular to provide meaningful insights at the level of every customer and every product (for example as in the long tail analysis described above). Second, the platform must seamlessly transform into a practical tool which sales representatives can use in the field. This is a particularly important requirement. Foodservice sales & marketing representatives as a rule have very little time for incremental effort above and beyond their existing selling and administrative responsibilities. They certainly do not have sufficient time to juggle multiple sales tools offered by multiple manufacturers acting in the role of category manager. The ideal tool is one with which the representatives have existing familiarity (to avoid time-consuming learning curves for new processes) and which can seamlessly integrate data from multiple input sources.

Manufacturers and distributors need more than just a rationalization program to optimize performance at the category level. A holistic approach, supported by robust analytics delivering actionable real-time guidance to sales professionals in the field, can improve category performance all along the foodservice value chain.

Subscribe   |  

Analytics for Intelligent Category Management

Collaboration between distributors and manufacturers is the cornerstone of category management in foodservice. For a given product category a manufacturer is selected to be category captain, with responsibility for improving category performance. This post addresses some key data and analytical issues with which manufacturers should expect to deal as category captains.

So you have been asked by your most important foodservice distribution partner to be a category captain. What happens next? As captain you are tasked with managing the assigned category for optimal performance. That entails the following:

•    Analyze all products across the category (not just your own brands)
•    Augment the data provided by the distribution partner with your own internally generated insights
•    Provide structured, actionable recommendations based on intelligence obtained from the data

These recommendations relate to product assortment, pricing policies, promotional activities and other important demand levers for driving profitability. At the same time you need to educate your distribution partners, both at corporate headquarters and in the field, about the product characteristics that can help increase demand. This requires an intelligent approach to data analytics.

What insights about products can help drive category sales?

What might a good analytics model for category management look like? Let’s consider the key tasks we identified in the previous paragraph.

Analyze All Products Across Category

Category management is driven by analytics. As category captain you will receive transaction data from your distribution partner to form the basis of your insights and recommendations. The first issue with which you will likely have to deal is the quality and completeness of this information. Bear in mind that foodservice distributors are typically not used to sharing sensitive sales data with their suppliers, and may lack effective internal processes for making it available. Robust data management solutions like Sentrana’s MarketMover™ help collaborative category partners overcome this challenge by providing timely access to clean, customer-level data.

The next order of business is to map out the analytical processes that can best support your distribution partner’s objective to improve category demand. This may be best approached through posing a series of questions. For example:

•    What intelligence can we derive from the data to help identify ways to improve product demand among existing customers?
•    What patterns and associations will provide us insights about products that current customers are not buying from our distribution partner but could be enticed to buy?
•    How can we improve sales turnover by encouraging customers to switch from lower-to higher-velocity SKUs?

Augmenting Data with Internally Generated Insights

In answering those and similar questions one of your most important activities is to augment the data your distribution partner provides with your own unique insights about the products in the category. An important example of this are the product attributes that drive demand among certain customer types. Perhaps you are charged with managing baked goods and you need to figure out what the right use of shelf space will be for muffin products. Your distribution partner’s objective is to increase total muffin sales – for example along the lines of one of those three questions posed in the previous paragraph. As a manufacturer you can provide your own deep knowledge about what features and attributes drive sales among certain customers.

A critical data challenge, therefore, is to have the ability to map specific attributes to specific products. Category managers should be able to access the product database and establish product groupings and categories based on like attributes. Using the above example, for every product you can assign a  quantitative attribute metric. “Butteriness” may be an appropriate attribute for muffins, and you can rank all applicable products along the lines of “very / moderately / not very buttery”. This can facilitate more rational product groupings within the category that better enables you to analyze and evaluate assortment trade-offs, pricing strategies and promotional approaches.

This kind of product administration capability brings up in turn a whole series of issues around how to create standardized attribute definitions for each relevant subcategory and product set. By allowing category mangers to create new product and subcategory groupings, it becomes likely that these categories will not map directly to those of the distributor. A category administration functionality is required that will manage the interface between the distribution taxonomy and the specific product and attribute groupings mapped by category managers at the manufacturer.

Supporting Analytics

As you map out these processes you can get a better sense of what the analytical capabilities may look like. For example, what data exploration functionalities can help you analyze effectively? To orient your own understanding of the structure of subcategories and products to its organization in your distribution partner’s data records you need a mechanism for guided drill-down and drill-up within products, as well as in regard to customers, sales territories and other key information. There should be a filtering mechanism that allows you to view products along a number of descriptors: for example, all the SKUs currently stocked at a particular operating company of your distribution partner, or all products with the item description “frozen”.

Visualization and editing capabilities are also important components of analysis. How do you want to see the information and organize it into compelling formats for your targeted readers? You will probably want to have a variety of formats and the tools to manipulate data into different visual representations to underscore the insights about customers and products you wish to communicate. You will also want to be able to easily access and modify the reports and formats you use most frequently, and to share reporting and editing capabilities with others working on the same projects.

Providing Guidance and Recommendations

Category managers need to consider how best to translate analytical insights into actionable recommendations for their partners. For example, developing strong promotional content around products for the distributor’s sales force can be an important way to execute against category performance targets. A system for uploading, managing and exporting product-related content is thus an important functionality to consider. Another valuable feature could be scenario analysis capabilities to map out alternative approaches to pricing decisions, promotional opportunities and assortment trade-offs. Finally, manufacturers need to consider how to incorporate data from their own market sources: for example sales information at the total market level rather than just the share occupied by their distribution partner.

Collaborative category management can evolve into a long-term relationship that will improve category performance for distributors and improve overall product sales for manufacturers. Over time the scope of a category management program may expand to include enhanced predictive initiatives and a fuller set of demand levers. Building a good foundation with the right data analytics is a good place to start.

Subscribe   |  

Want to Know Your Competitors’ Prices?

You May Already Have the Information

intelligence on competitors' prices may be close at hand

If only you knew what your competitors are charging. How many times on any given day does that phrase get uttered in a corporate boardroom, on a sales call or in a marketing strategy meeting? Knowing what your competition is charging would take so much of the guesswork out of your daily pricing and marketing decisions. It may come as a surprise, then, that critical information capable of revealing competitors’ prices may be very close at hand – in your own purchase history.

Is there a “Market Price” for COGS?

Since you do not have direct access to your customers’ prices, the challenge is to model likely competitor activity based on incomplete information. A good place to start is with costs – specifically cost of goods sold (COGS). This is typically a key input in pricing models. Knowing a competitor’s COGS would provide critical intelligence in determining what prices they are offering in the market.  The trick is to accurately infer the “market” price a competitor pays for their inputs (i.e. their COGS) from the information contained in your own transaction data. Read the rest of this entry »

Subscribe   |  

Avoiding the “Irrelevant Elegance” Trap: Modeling for Practical Business Outcomes

Quantitative modeling is a creative process. There is as much art to modeling as there is science – choices about what relationships you want to express and how to express them. And just as with anything creative, the authors of quantitative models can take pride in the beauty of their creations. In the words of my colleague Ali Mahani, Sentrana’s senior quantitative modeler, models can be truly elegant – they can be things of beauty. But he adds that they can also be irrelevant – irrelevant to the particular business goals they are intended to serve. That presents a problem for enterprises seeking to elevate the role of quantitative insights in their decision making processes. Data and analytical methods are important tools in the arsenal of a modern enterprise. But decision makers would be wise to heed my colleague Ali’s advice: in using these tools, make sure to avoid the trap of “irrelevant elegance”.

Elegance does not always lead to the best outcomes

Elegance in modeling is expressed in the appearance of simplicity – rendering sprawlingly complex interrelationships in the real word into the clarity of precise mathematical formulae. Simplicity and elegance are all well and good, unless in the quest for this holy grail you wind up dramatically misrepresenting how things actually work in the environment you are trying to model. This can result in not only failing to solve the business problem at hand, but actually making matters worse than status quo ante by facilitating decisions based on incorrect assumptions. We have a real world example of just how much worse this can be in the financial markets debacle of 2008, when the elegant models crafted by the best and brightest quantitative experts Wall Street had to offer proved to be fatally flawed in the assumptions and heuristics they used to express the variables affecting housing prices, interest rates and mortgage payment trends. Perhaps modelers need to live by something like the Hippocratic oath taken by medical doctors: first of all, do no harm. Read the rest of this entry »

Subscribe   |  

Physics Envy: Pervasive, But Not Incurable

Everywhere you look, it seems, people are talking about “physics envy”.  This derisive term mocks the attempt of economists and other social sciences practitioners to imbue their disciplines with the equations and mathematical rigor of physics – a rigor that many believe fails when applied to the messy environments of disciplines like sociology or economics.  It’s not a new term – economist Philip Mirowski contributed to the Finnish Economic Papers series way back in 1992 with a piece entitled “Do Economists Suffer from Physics Envy?”

kinetic energy, not supply & demand

Eighteen years later the answer from many observation posts along the byways of public discourse appears to be: yes, they most certainly do, and so do their fellow travelers, business and financial markets experts.  After all, we just barely survived the most devastating economic event of our times, deeper and more far-reaching than any downturn since the Great Depression, and all the high priests of the field can do is shake their heads and say “wow, I sure didn’t see that coming.”  Distrust of fancy math is rampant in all walks of business life.  That presents a real problem for enterprise decision-makers at a time when they need smart quantitative tools – yes, fancy math and all – more than ever.  Markets are more complex than at any time in human history.  Giant waves of transactional data inundate marketing managers with new information every day.  Managers need science to help them gain valuable insights into the markets for their products and services – but how do they know that the growing number and variety of scientific marketing tools out there aren’t infected with the nasty symptoms of physics envy? Read the rest of this entry »

Subscribe   |  

Quantitative Intuition II: The Bayesian Brain’s Achilles Heel

In a previous posting (“Quantitative Intuition: It’s Not Counterintuitive”) I described some of the advancements that have been made in bringing together the disparate worlds of quantitative methods and human intuition, ending on the rather happy note that advanced scientific micromarketing models today are capable of introducing qualitative human judgment and experience into quantitative models, such that the models are able to “learn” from humans about important factors such as competitive threats, nuanced negotiation strategies and even meteorological vagaries – factors that traditionally have been difficult to crunch into the binary 1s and 0s of machine language.  The human brain works in a hierarchical manner, embedding propositions within propositions to think a potentially infinite number of thoughts.   In the example I used in the last posting, a sales rep who reads about a national wholesaler coming to town to open a discount distribution center can nearly instantaneously form a series of mental propositions to evaluate the importance of that news and the probability of potential outcomes that may (or may not) require decisive competitive action from the sales rep’s firm. Read the rest of this entry »

Subscribe   |  

Quantitative Intuition: It’s Not Counterintuitive (Nor an Oxymoron)

Think of the best salesperson you know: if you’re fortunate, perhaps someone in your company or, less happily, in a competitor’s firm.  What are the qualities that make this person excel at the job of sales?  In a classic Harvard Business Review article “What Makes a Great Salesperson” (July-August 1964) David Mayer and Herbert Greenberg likened a star salesperson to a heat-seeking missile: “Sensing what customers are feeling, they [the sales stars] are able to change pace, double back on the track, and make whatever creative modifications might be necessary to home in on the target and close the sale.”   Whereas most of us have intuitive abilities to a greater or lesser extent, excellent salespeople lever this intuition with strong empathy skills (sensing what the customer’s needs are) and the relentless personal drive necessary to cross the finish line.  If they could, managers would bottle this elusive elixir of talents and have all their salespeople drink it, every morning of every day. Read the rest of this entry »

Subscribe   |  

Models Didn’t Bring Down Wall Street; People Brought Down Wall Street

“Burn the mathematics” wrote economist Alfred Marshall in a letter to a friend, musing about the proper role of mathematics and scientific inquiry in the field of economics.  That 19th century cogitation would seem to be a prêt-a-porter soundbite for these latter days of the 21st century’s first decade – a time in which the mathematical infrastructure that underpins longstanding economic and financial theories stands accused of all manner of malfeasance, particularly given its presumed role in the decade’s signature economic event – the financial market meltdown of 2008.  The logic behind the accusation goes roughly thus: More complex (but not necessarily more “accurate”) models allow for more complex instruments to be created. Increased complexity means it takes more time to process and then fully comprehend what the numbers may be telling you. At the same time, though, technology allows buy and sell orders to be executed almost instantaneously through electronic trading systems. Time is of the essence, and ponderously complex computations simply won’t do.  A seemingly elegant (and fast, and commercially viable) shortcut is discovered and becomes the currency of the day. The models’ outputs come to be trusted blindly simply because there is no time to question them (and too much money to be made by using them). The impenetrable Greek letters obfuscate the sensitivity of the models to changes in important assumptions – which is fine for a few years because those assumptions (e.g. rising housing prices) don’t change – but then all of a sudden they do. The models start losing more money than they make. Then the chasm widens further as the high levels of leverage in the system make themselves felt. The losses accelerate dramatically, wiping out years of profits in just a few months. Burn the mathematics, indeed.

But let’s take a different look at this apparent tight coupling of mathematics and dire outcomes. Our recent correspondence with an author who has been widely published on the subject of Wall Street’s use of mathematical models recently offered to us an interesting opinion. His point was that the problem with the models was not so much their complexity, but rather that they were models in the first place. His argument was that you can’t ever perfectly hedge model risk.  Now, I agree with that observation: a model by definition selects some aspects of reality to represent and omits others, and the choice of what to include and what to omit is subject to human error, therefore fallible and not perfectly hedgable.  But I take issue with the idea that the fault lies in the existence of the models themselves.  Models can be misused – I think that much is clear. But the notion that models are all doomed to failure obscures a deeper truth about the goals of predictive modeling; namely that you can seek either to reduce the world or truly explain it. By trying to elegantly reduce the world to as few predictor variables as possible, you are more likely to be sowing the seeds of future failure, because complexity and actual drivers of outcomes are taken out of the equations to make them more solvable (or perhaps sellable, as in the case of the Gaussian copula function that was behind Wall Street’s demise, as we discussed in a previous posting “You Can’t Punt Away the Dimensionality Curse”). Predictive modelers don’t have to go down that road, however: they can also set out with the goal not of reducing an entire system to a single neat, tractable equation, but to quantify and explain all of the relationships that dictate outcomes to the absolute fullest extent possible. Tractability and computability are things to address later in the process, through technological means, but they should not dictate the fundamental mathematical approach at the outset. Read the rest of this entry »

Subscribe   |  

In Economic Modeling, Can Hindsight Lead to Foresight?

In thinking more about my last posting here on failed Wall Street quant models and the dimensionality curse I started to wonder whether we could ever be more than the archetypal Monday morning quarterbacks:  commenting brilliantly on all the reasons why X should never have happened, after X has already happened and done its damage.  Can the mistakes of hindsight lead to foresight?  In other words, can we apply foresight to develop “good” economic models that won’t blow up in our faces?

In trying to answer this postulation we must go back to examine the eternal challenge of good modeling: how to create a simplified representation of reality that in ignoring many real-world features still manages to convey an inherently robust facsimile of the real thing.  For example, one of those maps of New England you buy at Exxon gas stations can serve as a good model for getting you from Hartford, CT to Boston, MA even if it ignores most of the streets and alleyways and other real-world detail that exist along the route.  In their book “Complex Adaptive Systems: An Introduction to Computational Models of Social Life” John H. Miller and Scott E. Page observe that the “ability to ignore is a crucial component of scientific progress”, using the image of a parent’s being able to respond to the incessant “why” questions of a three year old child by saying “just because”.  The trick, as the authors point out, is knowing when (and perhaps more importantly when not) to say “just because”.

While I wholeheartedly agree with that assertion I don’t think that it quite gets us to an adequate level of comfort in applying foresight to the creation of good models.  In his fascinating book “The Origin of Wealth” Eric D. Beinhocker points out that economic modeling took what many consider to be a wrong turn back in the latter years of the 19th century when leading thinkers of the day like Leon Walras and William Stanley Jevons borrowed heavily from the referential context of physics to create models for explaining economic activity, including such notable concepts as a mathematically representable state of equilibrium that continue to serve as the conceptual foundations of modern economics textbooks.  As Beinhocker elaborates, the problem with these models was that some of their fundamental assumptions – like the perfect, robot-like rationality of human beings in making economic choices – didn’t seem to simplify reality as much as contradict reality.  Thus we find ourselves in the present ruminating over the precise, mathematically elegant language of physics and the complex, evolutionary language of biology and debating whether a choice of the wrong science by the founding fathers of economics back in the 19th century led to the failure of models to adequately explain much of what is going on in the economy today and in particular the string of boom-bust upheavals that have become part and parcel of the last 20-odd years of economic activity.

I still don’t think we are there yet in getting closure on the foresight question, but we may be getting closer.  To tie in the strands of thought presented by Miller & Page and Beinhocker, when we get to those basic defining assumptions, Read the rest of this entry »

Subscribe   |  

You Can’t Punt Away the Dimensionality Curse

A single mathematical formula brought ruin to the global financial markets.  What happened was not a failure of quantitative methods per se but rather a lesson in the perils of ignoring real-world complexities in favor of deceptively elegant shortcuts.

The fault, dear investor, lies not in the head of AIG’s Financial Products Group or members of the Bear Stearns Investment Committee or any other anthropomorphic entity: rather it was a single mathematical formula that apparently felled the pillars of global finance.  That’s the gist of a recent article in the 3.17 edition of Wired magazine entitled “Recipe for Disaster: The Formula that Killed Wall Street” by Felix Salmon.  The formula, known as a Gaussian copula function (when is the last time that term became a fixture of the public discourse?), purported to solve the mother of all securitization problems: establishing default correlation factors between the many constituents of the pools of mortgages and other credit obligations whose cash flows served as the underpinning for the complex derivative securities known as collateralized debt obligations (CDOs).  Awareness of the potential in this arcane formula helped power the CDO market to some $4.7 trillion in volume over the course of the housing bubble years of this decade.  As the Wired article explains, the explosive commercial viability of this formula can be explained by its use of a simple sleight of hand.  Rather than modeling out the default correlation implications of pools of thousands upon thousands of individual mortgage obligations – an extremely complex undertaking requiring powerful algorithms and massively robust computational processing technology – the CDO market’s Wall Street practitioners used a shortcut that appeared elegant but proved deadly: using the market price of credit default swaps (CDSs) as a proxy for the actual historical data.

What happened in essence was that the CDO market ran up against one of the most challenging of quantitative modeling problems: the dimensionality curse.  This refers to what happens in complex environments where numerous variables interact with each other and all of the resulting combinatorial possibilities influence the economic value.  The addition of an incremental variable to the pool exerts an exponential effect on the number of possible outcomes.  Think of a simple case: if you have a pool of two variables then the number of potential outcomes is four: add a third dimension (variable) to the mix and the potential outcomes expand to nine, and so on.  In an environment like pools of thousands of mortgage obligations or credit card receivables influenced by a bevy of macro- and micro-economic, behavioral, seasonal and other random factors there are literally billions of combinatorial outcomes that could affect the incidence, magnitude  and frequency of default events and hence the price of the CDOs whose economic value derives from those pools.   Getting to the right answers – and doing so with enough speed to satisfy the blistering pace of 24-7 investment markets every day – is a daunting challenge to say the least.  So when Daniel X. Li, a quantitative analyst at JPMorgan Chase, posited the use of CDS prices as a proxy for historical data in a 2000 paper published in the Journal of Fixed Income Securities, the CDO market rejoiced and basically punted away the dimensionality curse by using this shortcut.  The reasoning and the assumptions employed proved to be flawed and the disastrous results are entirely visible to the naked eye in all their graphic detail.

In quantitative methods as in life there are no free lunches.  You can’t simply punt away the dimensionality curse – you have to embrace it and try to achieve mastery over it using all the knowledge and technology tools at your disposal.  At Sentrana we deal with dimensionality curse problems every day – the demand markets for the products and services our clients sell are highly complex environments: tens of thousands of products for thousands of customers in hundreds of locations reachable by any number of marketing vehicles and sales channels.  Modeling these environments is not for the faint-hearted: but the problems are not impossible.  The computational technology does exist, as does the modeling science.  The critical ingredient is the will and determination of those who practice quantitative methods in business to forego the easy outs and stay focused on solving the real problems, however daunting.

Perhaps the field of quantitative methods needs a variation of the medical profession’s Hippocratic Oath: First of all, do no harm.  Clearly the Wall Street experiment egregiously failed that standard.  Let’s hope that the next time some arcane mathematical formula figures into the cultural Zeitgeist it will be for better, not for worse.

Subscribe   |