Precision Forecasting

Have you ever had an experience at a meeting that you will never forget? Here’s one of mine.

A few years ago, I was advising a major company in restructuring its supply chain and customer relations. I visited one of the company’s most important suppliers. This supplier was very well run, with a well-regarded sales and operations planning process. In the course of our frank and confidential discussion, a senior VP of the supplier leaned over and said to me, “We always forecast their forecasts.”

The supplier executive was referring to the difficulty of forecasting sales. In fact, my client was quite well run, and had state of the art forecasting and replenishment systems. Nevertheless, the forecasts were often inaccurate, and the supplier simply formed its own judgment about the customer’s upcoming purchases.

Why is forecasting so difficult in so many companies? I thought about this question a few days ago as I prepared for an interview with a writer penning an article on forecasting branch sales and expenses.

Traditional approaches

The traditional approach to forecasting is to analyze historical trends, and project them into the future, often by product or product line. Of course, this assumes that the future is like the past. It also implicitly assumes that the company does not have the ability to dramatically change sales. In a rapidly changing world, this approach leaves much to be desired.

A second approach is to correlate sales with a broad economic or industry indicator, such as an industry purchasing or inventory index. The difficulty here is that, even if you understand the correlation and relevant time lags, you still have to forecast the indicator.

Here’s a third traditional approach. In many companies, the sales forecast is a compilation of each sales rep’s individual forecast. How does a sales rep forecast sales? For a start, each rep has a sales quota for the upcoming year, and it’s always larger than the previous year’s – even in a declining economy. Think about this: have you ever seen a sales rep forecast sales below his or her quota?

The underlying problem

The underlying problem with the first two traditional forecasting approaches is that they are too aggregated and passive to be consistently accurate. In fact, a company’s overall sales is a blend of a set of identifiable revenue streams – difficult to forecast in aggregate, but individually much more amenable to accurate prediction. Importantly, you need a different predictive technique for each major revenue stream. These can be summed into a very precise forecast.

The problem with the third approach, of course, is that not all sales reps meet their quota – often leading management to develop a “secret forecast” using the first two approaches. This is a real problem in a declining economy.

A precise approach

The starting point for precision forecasting is to disaggregate a company’s revenues into a set of major revenue streams, each of which has similar characteristics. Here’s an old Byrnes family recipe.

Divide your company’s accounts into core and non-core categories. Core accounts are major accounts where you have a strong relationship and ongoing sales, where you are a dominant supplier. Non-core accounts are lower-volume accounts that either purchase steadily or occasionally. Core products are products that have high aggregate sales volumes, while non-core products are slower movers.

You can display this categorization in a simple 2 x 2 matrix, like the one below.

Core products in core accounts. This is a critical part of your business. It represents your most important, high volume products in your most important customers. Note that there will be a relatively small number of products and customers in this category.

Here, there are two key elements to a good forecast.

First, it is critical to have ongoing intensive discussions and collaboration with these customers, especially with the customer’s sales and marketing team, so that you develop a deep understanding of the factors that are driving their sales and purchases of your products. Often, you will have as much knowledge as the customer has, and surprisingly often the customer’s replenishment and purchasing team does not have a close relationship with their sales and marketing counterparts, or even a deep understanding of the complexities of their business.

Because there are relatively few key products and customers, this is not an onerous task. Besides, you will develop a very productive set of relationships throughout the customer’s buying center that naturally drive higher sales along with operating cost reductions.

Second, new and lost business reports make all the difference here. Many companies require their sales reps to file reports outlining new and lost accounts. Because sales reps are busy and have many accounts, they often neglect this. By focusing the process on core products in core accounts, and tracking this carefully, you can make the reports pragmatic and effective. This is one of the most critical elements in precision forecasting.

Non-core products in core accounts. Once you have developed a strong understanding of the core products in your core accounts, this knowledge will strongly inform your forecast of their non-core purchases. In the course of your discussions with these major customers, you can find out the big changes they envision, and develop a sense of how these will affect the lower-volume products they purchase.

For example, you might learn that the customer is phasing out a product line, or going after a new set of accounts. Both would have an impact on your sale of non-core products to the customer. Note that this important information would not emerge in an aggregate forecast.

Core products in non-core accounts. This category has two major components.

First, many non-core customers are simply small businesses that do not have a lot of purchase volume, or they may even be larger companies that do not make large purchases in your category. Here, the first two traditional forecasting approaches are adequate.

Second, some non-core customers are really very important, high-potential underpenetrated accounts. This is a critical group because this is where a great sales rep can obtain rapid major sales increases. In fact, in my research and consulting, I’ve found that the top-performing President’s Club winning sales reps in company after company are experts at turning around high-potential underpenetrated accounts.

The key to success in this group is to distill your company’s best practice from your President’s Club winners and teach it to your average performing reps. When your average performers develop account plans to turn around and grow their high-potential underpenetrated accounts, you can simply use the account plans as the basis for forecasting this group. Of course, not all reps will meet their account plans, but with a little experience, you can develop a very good forecast of expected sales of important core products to this group.

Non-core products in non-core accounts. This category is comprised of slow moving products bought both by small businesses, and by high-potential underpenetrated accounts. Here, it makes sense to use the first two traditional forecasting approaches. Because the high-potential underpenetrated accounts are in growth mode, their purchases of non-core products probably will be relatively minor.

Forecasting paradigm shift

Precision forecasting involves a paradigm shift in your fundamental approach. Traditional forecasting is largely based on historical, aggregated, passive information.

Precision forecasting, in contrast, involves disaggregating your revenues into critical components, and forecasting each component appropriately. It also involves inserting a major element of control – both intensive discussions and collaboration with major customers, and purposeful penetration of high-potential underpenetrated accounts.

This new paradigm gives you both much better forecasts, and much better account and sales performance. What could be better than controlling your own destiny?

Strange Finance Problem: Too Much Accuracy

One of the biggest problems in effective profitability management is the instinctive desire on the part of many CFOs for too much accuracy. How can this be?

In my experience, there is good accuracy and bad accuracy. Let me explain.

Good Accuracy

Accuracy is essential for financial reporting, a prime CFO job.

I’m on the Audit Committee of a NYSE company, and we spend a lot of time making sure that the company’s financial results are reported extremely accurately. Like all public companies, we have controls and audits to ensure that even inadvertent errors are uncovered and rectified. The SEC and other regulatory bodies take strong measures to enforce this requirement.

This need for strict accuracy in financial reporting shapes the culture of all Finance Departments. Most CFOs and other top finance managers have strong accounting backgrounds, and many have managed internal auditing departments.

This strong cultural bias causes two big problems in profitability management: (1) the overly rigid use of financial reporting information for granular profitability analysis, precisely because it is so accurate; and (2) an unfortunate insistence on a very high degree of accuracy in profitability management. Both get in the way of effectiveness.

Bad Accuracy

The first problem, using financial information for granular profitability management, seems counterintuitive. After all, great financial information is available and the company has worked hard to ensure its accuracy – why not use it?

The problem with simply using existing financial information is that it is compiled into categories that are much too broad for effectively managing profitability. Here’s the acid test: if you select five accounts at random, and select five products at random in each of those accounts, would you know (or could you find out quickly) the profitability of each product in each account? In virtually all companies, the answer would be no.

This leads to the problematic situation of virtually all companies: 30-40% of the business is unprofitable by any measure, and 20-30% provides all the reported profits and subsidizes the losses. This results in a huge amount of embedded unprofitability in virtually every company – unmeasured, unseen, and untapped.

Why does this occur? Because revenues and costs are compiled into categories that work well for reporting the company’s overall financial results, but are not granular enough to match revenues and costs on an invoice line basis. As a result, finance managers embark on a problematic exercise in allocating these accurately measured costs and revenues into broad categories of account segments and product families – usually by using imprecise allocation functions.

All too often this becomes an exercise in trying to measure accurately things that can’t be accurately measured. The objective implicitly becomes measurement itself, rather than clearly and relentlessly focusing the organization on discovering and acting on the most important profit opportunities.

Instead, you need a completely different process to develop effective granular profitability information. I call this process profit mapping. It involves essentially developing a 70% accurate “income statement” on every invoice line, then putting these into a database and sorting them by product, account, and other essential dimensions. I describe how to do profit mapping, and provide a number of concrete examples, in my book, Islands of Profit in a Sea of Red Ink.

The process takes an insightful manager and analyst about a month or two, using standard desktop tools.

It is critical to understand the important difference between information developed for financial reporting and information developed for management control. This is one of the basic distinctions drawn in introductory accounting and finance courses, yet it is deceptively easy to blend these together when the finance managers have already developed a huge amount of very accurate financial reporting information.

What is 70% accuracy?

Many top financial managers, particularly CFOs, have difficulty understanding the need to work with 70% accurate information for profitability management. They are so culturally accustomed to producing extremely accurate financial reporting information, that they tend to view 70% accurate information as merely the result of inadequately careful work.

Nothing could be further from the truth.

In order to understand the needed degree of accuracy for granular profitability management information (as opposed to accurate financial reporting profit information), you have to start by thinking carefully about what managers throughout the company will do with the information. Then you work back to the nature of the information that will be most helpful to them.

The first set of actions that managers will take with a set of granular profitability information is to quickly see where the business is making money, where it is losing it, and why – its islands of profit and sea of red ink.

In practice, most companies have perhaps 10-20 clusters of product/account segments, each with its own internally similar characteristics and profitability – but each very different from the others. These will be very obvious in your profit map, even with 70% accurate information.

The action question for each product/account segment is: what one or two things can you do that will have the biggest impact on profitability: Is it pricing? Order pattern? Product portfolio? Visibility into big orders, etc?  I call these profit levers, and the big payoff comes from finding the smallest number of actions that produce the biggest results.

The most effective way to manage this process is to start with a rough profit map, using readily available information and estimates. The outlines, composition, and profit profiles of the product/account segments will literally jump off the page. By examining a few typical accounts in each segment, you will quickly see the most effective profit levers. But remember that these almost always are different for each segment.

At this point, it is helpful to sharpen your pencil where the additional accuracy will make a real difference:  measuring the exact problem that the profit lever will fix, and developing a precise program to remedy it. This requires a higher degree of accuracy in order to specify highly focused initiatives in your target areas.

This process will enable you to match your degree of accuracy with the likely use of the information. Where better information is clearly not needed for a high-priority action, why spend valuable time and bandwidth on developing it? Instead, find out early where enhanced accuracy will really make a difference, and focus your time and resources where they will produce the biggest payoff.

Maximum profit impact

The more subtle problem with bad accuracy is that profitability management is a moving target. Companies change all the time, especially in a complex economic situation.

Even more importantly, when your managers have worked with an initial set of profit levers to produce radically improved profits, a new set of opportunities for profit generation will naturally arise.

This means that your profit map, along with your segment and profit lever identification, must be very dynamic. If you make extremely accurate measurement the implicit objective, it is very hard to move as quickly as necessary in order to maximize your profitability.

Another critical problem arises. When you focus on measuring everything extremely accurately (implicitly believing that everything is important enough to warrant it), you are in danger of flooding your managers and sales reps with so much information that they will not be able to act effectively. They will lose the critical understanding of the smallest number of realistic actions that will produce the most powerful set of results.

Instead, by aligning your information accuracy with the nature of the information use, you will have the biggest impact on your company’s profitability. And for all CFOs and other top finance managers, maximum profit impact is the ultimate management objective.

Change Management: Paving the Cowpaths

Change management is one of the most difficult problems facing managers at all levels. All too often, managers focus primarily on defining the best end-state, and deal with the change process almost as an afterthought. If the end-state really is better, the logic goes, then people will find the vision compelling and migrate to it. Often this almost has the feeling of a “lay-down” hand in bridge.

This situation reminds me of the story of the math graduate student who woke up in the middle of the night and smelled smoke. He walked into the kitchen and saw that his stove was on fire. He looked at the water tap, looked at a bucket on the floor, said “a solution exists,” and went back to sleep.


“Paving the cowpaths” is an expression often used to express disdain for weak change management programs.

The streets of downtown Boston are characterized by byzantine windings that seemingly defy logic – until one realizes that they trace the original cowpaths that cattle trod to skirt fields when ambling home from pasture in early colonial times. In downtown Boston, the streets are, indeed, paved cowpaths. In the newer section, Back Bay, planners had the luxury of laying out streets in a grid pattern on landfill.

The essence of “paving the cowpaths” is to point out the seemingly obvious folly of managing change by simply changing nothing of importance. Much later, the “reengineering” movement echoed this idea, noting that change managers should reengineer processes rather than simply automate them.

Yet managing programs of fundamental, sweeping change requires a unique process that is almost counter-intuitive. The reason is that change management most often takes place in a well-established organizational context.

Cowpaths revisited

I thought about the metaphor of “paving the cowpaths” this weekend, as I was reading a very interesting book, Wired For War, by P.W. Singer. This book is about the history of robotics, and the use of robots in warfare. The really interesting parts, however, are about the process of new technology adoption. This process seemingly defies logic.

For a number of years after robots were shown to be superior for certain uses, military leaders refused to embrace the new technology. For example, drone aircraft were demonstrably better at particular missions than manned aircraft. Yet because most top Air Force officers were former pilots, they rejected the new innovations and even used the new drone aircraft for target practice in training pilots.

It was not until relatively recently that external events created a situation in which the new robot technology began to be accepted. In the early 1990s, the tide began to turn. Singer notes that in the prior period, “It isn’t that the systems weren’t getting better, but that the interest, energy, and proven success stories necessary for them to take off just weren’t there.”  The situation began to change dramatically in the wake of the “Black Hawk Down” disaster, which generated a widespread public unwillingness to commit ground troops to the Balkans and Rwanda.

Not long after, the high injury and mortality rates experienced by our military personnel in Iraq caused a public outcry. The military responded in part by accelerating the deployment of robots in high-risk military missions, which enabled the new technology to break through the traditional military attitudes toward robots. This reversal led military leaders to convert the role of robots in warfare from one of “paving the cowpaths” to a much more sweeping role accomplishing powerful new activities –  like staying aloft for days doing surveillance – that simply could not have been done with manned technology.

A similar set of situations characterize the whole history of technology innovation. The moral of the story? Paving the cowpaths is often an effective first step in sweeping, paradigmatic change. The real questions are: (1) whether the initial formulation is flexible enough to support migration to a new system, and (2) whether the change managers have the vision and capability to create the conditions favorable for change, and to manage a multi-stage change process.

Business change management

Paving the cowpaths, in this context, is an important first step in major change. In fact, many successful IT system implementation projects start by mechanizing existing processes, but keep latent capabilities to support deeper change. Over time, as the organization experiences the benefits of the new system and as it becomes part of the company’s “plumbing,” change managers can roll out the more sweeping new capabilities. Because the system users experience the early benefits of the new system, they become much more willing and motivated to change their practices to take advantage of the powerful new possibilities.

The business school teaching case literature has great case examples of companies that instituted surprisingly sweeping change, by starting with piggybacking on current organizational processes and relationships. In my course at MIT, I teach one such case juxtaposed against another that features a sweeping change program implemented as a “big bang” all-at-once change. Not surprisingly, paving the cowpaths turned out to be the secret of success – enabling much more rapid and broad change when properly deployed and managed.

Profitability and change

One of the biggest problems –  and opportunities –  in profitability management is the scope and magnitude of the potential improvements. Virtually every company is 30-40% unprofitable by any measure, and 20-30% provides all the reported profits and subsidizes the losses.

Profit mapping, the core analytical tool of profitability management, shows the profitability of every product in every customer in a company. (I describe how to construct a profit map in my book, Islands of Profit in a Sea of Red Ink.) Profit maps show exactly where profit is flowing and where is it lost.  The problem is that the sheer quantity of information can be overwhelming, and often leads managers to wonder where to start.

The most effective profitability management programs start with a degree of paving the cowpaths. The key to success is to prioritize the profit opportunities against the difficulty of change, and to develop smart ways to accelerate the organization’s pace of change. I think of this as the “view-to-effort” ratio.

View-to-effort ratio

When our kids were young, we used to take them hiking in New Hampshire. They complained on the way up the mountain, but they loved the view from the top. This led us to rate hikes by their “view-to-effort ratio. By starting with small hikes with nice views, we were able over time to sharpen their interest in great mountain views, and this in turn allowed us to increase the size of the hike.

In short, we started with a “pave the cowpath” approach, and moved forward from there. Through this process, our children learned to love hiking and enjoy nature. Today, they are hiking on trails that would leave us breathless. Had we not started incrementally, we would not have experienced this result.

Smart change management

One of the most important keys to success in profitability management is smart change management. It is also the biggest stumbling block.

The well-known inventor and entrepreneur Ray Kurzweil notes, “About thirty years ago, I realized that timing was the key to success….Most inventions and predictions tend to fail because timing is wrong.” Singer observes, “Kurzweil has found that the challenge isn’t just inventing something new, but doing so at just the right moment that both technology and the marketplace are ready to support it.”

The key to success is to craft a change program that starts with “paving the cowpaths.” This essential first step allows the organization to see the power and feasibility of the new approach, which in turn creates the essential conditions for more sweeping change.

Successful managers draw on their reservoir of wisdom and experience to craft the right formulation. This critical element makes all the difference between success and failure.

Great Summer Reading

Here are four terrific books that you might enjoy this summer. One is an enduring classic of nature writing. Two are great management books that are not technically about management. And the fourth is a very interesting brief overview of the history of mankind.

The Outermost House: A Year of Life on the Great Beach of Cape Cod, by Henry Beston.

In 1927, Henry Beston spent a year living in a one-room house on the beach at Cape Cod. He wrote this wonderful book in longhand at his kitchen table. In it, he describes his observations about the changing moods of the beach and water, and the joy of living in solitude in a little room overlooking the North Atlantic and dunes. Beston originally planned to spend just two weeks in the house, but was drawn to stay for a whole year. Some great chapters: “Autumn, Ocean, and Birds”, “Night on the Great Beach”, and “Orion Rises on the Dunes”. How good is this book? First published in 1928, it is still on the shelves of most bookstores.

The Accidental Billionaires: The Founding of Facebook, by Ben Mezrich.

This terrific book tells the story of Mark Zuckerberg’s founding of Facebook. Ben Mezrich’s fast-paced book narrates the engaging story of how Zuckerberg translated a relatively simple idea into one of the world’s most popular websites – carving out a dominant competitive position in less than a year.  Look for the essential elements of Facebook’s positioning success: focusing on something everyone does all the time, offering a slightly more efficient solution, and network effects that spread like wildfire. Something to think about for the inner entrepreneur in all of us. This book is very well written, and much better than the movie.

Game Change: Obama and the Clintons, McCain and Palin, and the Race of a Lifetime, by John Heilemann and Mark Halperin.

This fascinating history of the 2008 U.S. Presidential election is both a great read and an insightful management study of one of the most incredible turnaround electoral victories in U.S. history. The book tells the inside story of how Barack Obama vaulted from relative obscurity to beat the clear frontrunner, Hillary Clinton. This is a story of his relentless, thoughtful management, juxtaposed against Hillary Clinton’s dithering approach. It is a study in the critical importance of management style and process. Also compelling – the story of John Edwards’s implosion, and of John McCain’s selection of Sarah Palin to be his running mate. Although Game Change is not technically a management book, it is one of the best management books written in recent years.

The Origin of Humankind, by Richard Leakey.

In this short volume, renowned paleontologist Richard Leakey systematically answers the question, “What made humans human?” Leakey surveys our knowledge of the prehistory of humankind over the past three million years. This story is fascinating. But more importantly, the book is a wonderful example of concise, insightful analytical writing: in 157 pages, Leakey frames the essential arguments over humankind’s prehistory, impartially weighs the evidence, and offers well-synthesized conclusions. It is a great example to managers of how one can systematically investigate and deeply understand a complex field. Besides, think about this: modern humans only emerged 35,000 years ago, and began to settle in villages only 10,000 years ago. Look at the technological progress all around us – it’s truly astonishing.

A Short Bonus

Check out David Brooks’s recent New York Times op-ed column, “The Unexamined Society”.

It is a great explanation of the importance of clear thinking about things that managers and policy makers all too often take for granted. Here is his lead-in: “Over the past 50 years, we’ve seen a number of gigantic policies produce disappointing results — policies to reduce poverty, homelessness, dropout rates, single-parenting and drug addiction. Many of these policies failed because they were based on an overly simplistic view of human nature.”

Marsha joins me in sending our best wishes to you and yours for an enjoyable, relaxing summer.