There is a better way to create plans using business analytics (BA) and 80/20. Organizations of all sizes should take steps to explore the wealth of information available and commit to data-driven planning and decision-making. They should also rethink the yearly budgeting process, to make it more adaptive and analytical. The goal is to have a leaner and more robust process, delivering greater accuracy and agility. Companies create budgets or annual operating plans (AOP), with a large focus on financial targets. For the most part, they aim to reach strategic milestones and close gaps. But typically, they end up with plans that emphasize negotiated objectives and boundaries, with a lot less focus on optimization and transformation. Budgets by and large, become a financial exercise in compromising. Jack Welch (former Chairman of GE) calls the budgeting process “the most ineffective practice in management”. He goes on to say that the budget process is an “exercise in minimization. You’re always getting the lowest out of people because everyone is negotiating to get to the lowest number”. Accuracy, relevance and execution ability are the greatest weaknesses of typical annual plans. During preparation, budgets consume too much of management’s attention and can become irrelevant very quickly as conditions change (and they always do). Budgets are also known to be fixed and inflexible. To make it worse, most companies tie executive compensation to performance against the budget and not against previous years or industry benchmarks. Not to mention that a huge amount of effort and time is placed on budget preparation and data gathering. In fact, more data gathering than analyses. Plan iterations are usually the outcome of negotiating sessions amongst managers, as opposed to teamwork and analysis of contextual information than leads to foresight. Managers spend too much time bargaining and setting boundaries, and not enough time planning for execution. No wonder many companies fail due to lack of effective operating and strategic plans. The solution is to create an adaptive and analytical performance management process. Instead of cramming the entire exercise into annual planning seasons and quarterly or semi-annual reviews, companies should uncouple forecasting and financial analyses from planning itself, and make them ongoing activities of the organization. Once you uncouple them, what is left for the planning season is work that is more prescriptive (how to make it happen), as opposed to data gathering and number crunching. The year-round activities (reporting, analyses and forecasting) receive a large dose of intelligence using BA, while the once-a-year activity (planning) is redefined to focus on more prescriptive planning. The benefits are more adaptability, accuracy and relevance to the organization. Once you uncouple the planning elements, you end up with three parts:
Step two above (rolling forecasts based on analytics) is key to create an ongoing cycle of planning, execution and evaluation. It will also add intelligence to the process by tying the forecast to key business drivers. These two elements will make the company more agile and dynamic in the planning process. Rolling forecasts coupled with analytics, reduce uncertainty in planning and forecasting while encouraging constant assessment of the plan throughout the organization. A good example of a company that benefits from this approach is Southwest Airlines. Given the volatility in the airline industry, they’ve opted for a 12-month rolling forecast and quarterly planning. This approach has worked very well for the company, which has been consistently profitable for over 30 years, in a very tough industry. It’s also important to blend actual performance with analytics, to succeed. Forecasts need to be created rapidly and with less complexity and cost. Therefore it’s important to structure the data and the analyses, providing a decent level of automation. Bear in mind that analytics are not the sole responsibility of the IT department, just as budgeting is not an exclusive responsibility of the finance department. BA is a capability that needs to be developed by all functional areas and business units, while IT provides access to data and analytical tools. Typical areas to apply predictive analytics are as follows: Predictive analytics is by no means exclusive to e-businesses and is already penetrating brick and mortar manufacturing and distribution companies. Data is easy to collect and to access. The millennial generation, or the tech savvy generation, is taking over. New and more cost effective tools are available and companies are evolving quickly to more advanced levels. A few leading edge firms are already using prescriptive analytics. So allow me to take a moment to explain the different types of analytics. The first type is known as descriptive, and is used by the majority of companies, in more or less structured ways. It is done with standard and ad hoc reports and financial analysis. They look at history and tell you what happened to the business, how often it happens and derive actions to fix problems. Companies have different ways to extract insight and to diagnose the reasons why something happened. Management uses KPIs and other indicators that point to root causes and provide clues about potential actions to stop a problem from happening, and to improve the results. But keep in mind, this is all looking at the rear-view mirror and explaining what happened. While this approach can provide reasonable insight, the knowledge gained is all in hindsight and the future depends solely on critical analysis and management experience. The second category of BA is known as predictive analytics. It uses data mining techniques to look for hidden patterns and correlations that may exist in contextual data, in order to derive predictors that will point to the future. Today, e-commerce companies are the primary users of these tools. However, more companies are waking up to the power of predictive analytics. On top of the performance improvement potential by this method, there is an incredible opportunity for manufacturers to improve the longevity and reliability of products and production lines, for example, using predictors for maintenance and early product failure detection. The most advanced type is prescriptive analytics. It’s used to find the best courses of action for a given situation, feeding from both descriptive and predictive types. It’s a natural evolution for companies that are consistently mining and analyzing data, and trying to establish correlations and trends. It presents you with actionable options to develop data based optimization plans. Prescriptive analytics is an evolving field and advanced techniques are being tested, including the use of artificial intelligence software, to use both structured and unstructured data elements. You can start applying the prescriptive mindset right away, through mining the data that is already in your servers. In fact, 80/20 analytics is based on simple ways to evaluate correlated sets of data, such as products and customers, to arrive at optimization strategies. The picture below shows the three areas required to develop an adaptive planning process using business analytics. Many predictors are similar to KPIs used by managers. The difference is that most KPIs are set based on historical analysis and goals, while predictors are set based on data analytics, for relevance and insight. They allow you to set smart goals, while keeping the organization focused on fewer issues. Although Google Analytics is a tool for online businesses, it’s a good example of how to use prescriptive analytics. Google offers users a Prediction API (application program interface) that automatically provides pattern matching and machine learning capabilities. After it learns from a set of training data, the API predicts future trends, detect changes in usual patterns and recommend actions. It does amazing things with remarkable accuracy, like predicting spending patterns by online shoppers, customer attitude towards a certain brand, based on positive or negative feedback, and so much more. I believe such tools will become mainstream, at all types of industries, in the very near future.
While both predictive and prescriptive analytics are often associated with big-data and sophisticated algorithms, the basic principles behind these tools are really down-to-earth and useful to managers trying to focus on what is important (the “eighty”), mitigating risks and knowing which levers are most effective, in case they have a problem during execution. Planning with 80/20 analytics requires that we spend time upfront analyzing contextual data and identifying the few key pressure points and optimization actions. Let me give you a simplified practical application how a distribution business can improve its profitability outlook by applying adaptive planning and analytics. Like manufacturing companies, distribution companies want to improve the return on invested capital (ROIC) or, at a minimum, the return on assets (ROA). ROIC is a profitability ratio that tells us how good a company is at turning capital into profits. Companies want ROIC percentage to be attractive, so investors can continue to allocate capital. ROA is an indicator of how profitable a company is relative to its total assets, and a subset of ROIC, but mainly useful for distribution businesses, since most of the investment sits in the inventory part of the assets. In order to improve these metrics, there are two strategies available for the business: 1) to increase return on sales in terms of EBIT (earnings before interest and taxes) or 2) to turn its assets faster (or reduce the assets applied). Based on our experience, a reduction in assets (in terms of inventory and accounts receivables) or invested capital (CAPEX), has to be way beyond normal (or even healthy) levels in order to significantly impact ROA and ROIC. They can impact free cash flow positively in the short run, but will not necessarily make the business better. The other way to do this is to improve the fitness level of the company, by increasing the efficiency reflected in the Productivity Index (PI), which was discussed on a previous article. PI is the ratio between contribution margin dollars and overhead dollars. The crux of the issue is to know which levers are most effective to increase ROIC (pricing, overhead cuts, inventory reduction, etc.), while continuing to march towards profitable growth and value creation goals. In other words: what is the art of the possible without damaging the business? Through the application of business analytics that is connected with a rolling forecast, companies can arrive at an optimization plan that makes these apparently conflictive goals work together. In most cases, distributors find the following contextual data sets to be the most useful: markets and sales, customers and products, product lines and suppliers, contribution margins and overhead, inventory and contribution margins. Without a dose of foresight on how certain actions impact the P&L, managers can destroy value. It becomes a costly trial and error process. One of the common mistakes, for example, is to go after additional contribution margin dollars by lowering prices across the board, in order to increase sales (making it up with volume). The problem is that you need sales to increase significantly more to offset the margin reduction from price cuts. It is not a realistic scenario in most cases – it’s a “fools game”. Another common mistake is to underprice the tail end of the product line. In other words, distributors tend to be under compensated for investing in inventory or not to get paid enough for managing complexity on customer’s behalf. Sales analytics can help you avoid these errors, by recommending and predicting the actions’ impact on margins and ROIC. It tells you how much portfolio momentum you have, so you can take appropriate pricing actions. It also performs inventory investment analysis (turns and earns) and recommends a pricing strategy that is different for different groups of products, depending on how much contribution margin they generate and how fast they turn. Based on predictive models, managers can develop viable optimization plans, to improve the business and avoid deadly mistakes. Examples of such actions could include: a) increase prices and contribution margins selectively, using 80/20 analytics to define the different pricing levels for different products; b) scaling down overhead in selective areas of the business, allowing SG&A expenses to increase at a lower rate than sales (for example, below 2% sales increase); c) increase sales in certain market segments, which are projected to have above industry growth rates; d) recommend the optimal parts mix in inventory to maximize turns and earns, based on demand patterns and trends. At the end, the plan is still about setting priorities, improving contribution margins, reducing overhead, optimizing investment in inventory and looking for the best market segments to grow. What’s new here is how the actions can be combined, with the right amount of intensity, to produce maximum results. Could we thrive without a better planning process and business analytics? We could probably survive, but given the need for agility and the pace of change, the ability to react faster and apply analytics are key differentiators in most industries.
42 Comments
In previous writings I’ve talked about the importance of 80/20 data analytics. Here I want to give you an application example and hopefully explain how effective this process can be, when you apply sales analytics in conjunction with 80/20 thinking and tolls in search of profitable growth. Sales analytics tools are used to compile data from different sources, but mainly related to customers and products, transforming the records into useful information and insight that sales leaders use to understand their business and develop improvement actions. Data can come from databanks and pipelines such as CRM (customer relationship management), commercial transaction files, product cost databases and so on. This multiple interfacing capability is called MDS for Multi-Data Source integration. Tools with MDS capability literally mine the data from different archives, cluster it together and enable visualization through dashboards and reports. The level of refinement is growing rapidly and most tools now offer predictive analysis, which looks for leading indicators. At a first glance, these indicators may not look like they are directly connected to sales events, however through further correlational analysis and multiple associations in the data, a connection can be established and the indicator becomes a viable predictor of sales. As an example, the number of visitors to certain websites might be a good predictor of future sales of products related to those specific websites. Marketing organizations are constantly mining data to look for hidden relationships and trends that can provide clues about customer’s wants and sales trends. Perhaps the most widely used sales analytics tool for ecommerce is Google Analytics. It is a service from Google that provides statistics and basic analytical tools for search engine optimization (SEO) and marketing purposes. Geared towards small and medium-sized retail websites, Google Analytics has many features such as data visualization (dashboards and scorecards), motion charts, segmentation analysis, as well as custom reports. In spite of this growing level of sophistication, the vast majority of these tools are multi-purpose, or designed to organize the output in generic ways that are easy to understand and track. Then it’s up to managers to create strategies and extract conclusions from the many dashboards and predictors. And since these are primarily off-the-shelf tools and every business is unique, customization not only costs a lot of money, but customization also puts a cast around the company’s analytical capability. Having said that, almost all basic software products offer interesting tools such as sales dashboards, order pipeline and territorial management, sales planning and productivity metrics. The more advanced products offer additional features like predictive analysis, sales forecasting, customer data segmentation and product profitability analysis, Off-the-shelf analytic tools are great to track the performance of the sales team, but they are not meant to be transformational tools on their own. Managers obtain useful information and clues about productivity issues and market penetration, for example, but they do not provide a roadmap to growth or increased profitability. It’s when you couple these tools with 80/20 analytics and mindset that you can develop a strategic plan to optimize the portfolio, increase profitability and grow. Below I make an attempt to stack these tools from the tactical or day-to-day dimension (level 1) to the most strategic dimension (level 5). The first two levels are what I call “performance” levels and the other three can be considered “transformational” levels. The higher you go, the more impactful and lasting the results will be. Level 3 or 80/20 analytics, allow us to use the data from levels 1 and 2 to execute on the notion that “some customers and products are more equal than others”, and apply a super high focus on a select group of customers within the existing market segment. The customers in this select group are the ones that are likely to pay more for your unique value proposition (UVP), compared to the trivial many, or the “twenty” customers. Tools such as quad analysis help you fine-tune the portfolio by reducing overhead and SKU count, better aligning them with the needs of the “eighty” customers. The misaligned overhead and the low-volume and low-margin products are only creating complexity (read cost) and these are the tools that help you simplify your sales processes and optimize your offering. But once you are happy (for a while) with the optimization and the simplification then comes growth.
Top-line growth is essential for the long-term viability of any business. Segmentation (level 4) is the best and the safest way to generate new revenue streams. You start by looking inside your current markets to find new growth avenues, before you branch out into unknown areas. You should first exhaust adjacent or incremental revenues, near the core business, and then look beyond the core for new markets and selective acquisitions as growth catalysts. Preferably focusing on new product lines that will enhance your UVP to existing and prospective “eighty” customers. Tools such as cluster analysis and marketing mix planning or 5 P’s (price, product, promotion, place and people) help validate the ideas or segmentation hypotheses that might have surfaced through sales analytics tools in level 2. But segmentation is not just about finding niches, but it’s mainly about the way you run the business, redirecting and specializing your sales and marketing organizations and eventually splitting them into new, expert business units. By using segment-focused business units, the expansion costs are kept closely coupled with the unit’s ability to deliver growth and to charge for the value of their UVP. Segmentation will lead to de-commoditization of your business, while empowering your best people to succeed in the refocused business units. But how do you sustain growth and profitability overtime? Innovation (level 5) is how businesses sustain market share and create moats, to defend against competitive attacks. You need to defend your market share and explore incremental opportunities, but you also need to be capable of developing new strategies that can lead to transformational growth. The key to new strategies is innovation, which is the driving force behind adaptability and the skill that yields direct change. Peter Drucker has a simple, elegant definition for innovation: “change that creates a new dimension of performance.” Sales innovation needs to go beyond products. It needs to solve problems for customers and markets. It starts with the analytics and requires an understanding of the problems or pain points faced by the “eighty” customers, within a specific market segment. Using ideation tools and techniques and problem solving methodologies, the pain points are converted into ideas and solutions. Market-segment-focused business units are better prepared than any other type of organization to find the pain points and deliver innovative solutions to known and unknown customer problems. Specialized sales organizations are closer to end users, hence they are best positioned to create a unique view of the market segment, using analytics as a microscope to slice and dice the customer base. This is where the majority of product and service innovation comes from—discovering pain points that customers might not even know they have and figuring out how to solve them in a collaborative way. Companies like General Electric and ITW have used this formula over and over to deliver outstanding performance. In summary, it takes more than sales analytics to grow and sustain profitable revenues. The data needs to be part of a framework that contains other strategic elements, such as those found in the 80/20 Business Process. An off-the-shelf software package will help you manage the effectiveness of your sales force and provide you with ideas, but it will not create a higher level of performance alone. To attain transformational growth, you need to combine sales analytics with a selective mindset and a sharp niche focus, provided by specialized sales teams and business units that are constantly trying to innovate beyond products. This combination represents a virtuous cycle that will create a strong moat and differentiate your company in the market over time. A very interesting book by MIT Media Lab Professor Cesar Hidalgo, titled Why information grows: the evolution of order, from atoms to economies[i] talks about why and how information shapes up our world and its economies, in a lot more subtle ways than we can imagine. People, companies and economies form complex networks that exchange information and use knowledge and know-how to create our business world. It offers a fresh way to look at how information gets disseminated and how it helps economies grow.
Information is inherently embedded into every product, process and customer interaction kept by the company. Products and processes contain “information crystals” that determine how they work and how they are put together, intentionally (through knowledge and know-how) or unintentionally (modified by unintended complexity). Information grows as knowledge and complexity grows and complexity grows as the business grows and matures. Inevitably, as the business evolves, complexity modifies some (or most) of the original purpose or intent of the company and gradually rewires the portfolio, processes and other critical functions of the enterprise. In other words, complexity transforms the foundational information base that constitutes product portfolios and customer interactions. This is the very reason why complexity leads to non-intentional activities, non-value-added tasks, unprofitable products and low-value customers. Complexity rearranges the information crystals in unpredictable and most of the time, in undesirable ways. Data analytics is a very powerful tool to read and interpret the information crystals and uncover clues on how complexity has rewired the business, changing the way the company operates versus its purposeful intent. I find data analytics equivalent to taking an X-ray of you portfolio. 80/20 data analytics (8020dA) is a tool for examining interrelated sets of data to uncover hidden patterns, unknown correlations, market trends, customer preferences and other useful information that can lead to improvement actions. If nothing else, it unveils the very “center of gravity” of the business, i.e. the vital few customers and products. In almost every analytics exercise you will deal with large sets of raw data (big data), in order to draw useful conclusions. Fortunately, with the ever-increasing power of information-technology resources, 8020dA has become highly practical and desirable. The abundance and the easy access to data, the processing speeds and the analytical tools make it easier to mine for meaningful information nuggets to exhaustion. The most sought after 8020dA is related to data at the intersection of customers and products, as managers search for new insights into buying patterns and marketing strategies. However, it’s important to note that 80/20 data analytics is not exclusive to customers and products; it can be applied to any two sets of interconnected data to understand how activities are wired in order to reach actionable conclusions. Different sets of data are linked throughout the company, such as suppliers and purchasing transactions, people and fixed-costs, procedures and reports, and so on. There is great value in analyzing different data sets for simplification of business processes. Data analytics for customers and products contains the two most valuable interrelated sets of data about the business and is a powerful way to bring your company’s most important activities into focus. It allows you to visualize complexity and extract essential conclusions that are not so intuitive or easy to interpret when you look at individual bits and pieces of the puzzle. The outcome is much more meaningful and revealing than the raw picture we draw in our minds when we start the process. It also gives you the ability to begin adjusting your portfolio right away. 8020dA can be divided in three different types of analyses: correlational, descriptive and inferential. Correlational analysis describes how the raw data is interrelated, using tools such as quartile, customer-products matrix and quadrant analysis. You are trying to answer the question of how much complexity and distortion has entered the business. The sheer shape and size of the data set is very revealing in itself. How dense or sparse is your offering? Do you have patches, uneven distribution or gradients in the data? All of these characteristics can help the experienced observer understand how complexity has rewired your company’s portfolio. The descriptive analysis depicts threshold properties and unique characteristics about the data, using common statistical metrics (mean, median, sparsity factor, etc.) as well as experienced observation of unique characteristics amongst customer and products. Once you characterize the data, then you turn to the inferential portion to derive actionable meaning from the information and optimize the portfolio. In order to do that, we first need to transform raw data into information. In summary, 8020dA is an essential tool to uncover the intentional and the unintentional “crystals of information” in your portfolio. To use a medical analogy, before you decide to take on surgery to rearrange your portfolio, it is advisable that you use all the X-ray and imaging capabilities available to you, to fully understand the situation and avoid surprises when you cut open the patient. [i] Why information grows: the evolution of order, from atoms to economies / Cesar Hidalgo – Published by Basic Books, 2015 |
AuthorPedro Ferro Archives
August 2016
Categories
All
|