Building Better Business Forecasts

Develop Key Metrics .

Nonfinancial metrics can greatly impact a company’s ability to create shareholder value. However, developing key metrics is no slam dunk. Here’s how finance can apply business intelligence to the mother lode of available data, improve the forecasting process and create performance reports that drive profitability.


Wall Street analysts and prospective investors no longer rely solely on past financial performance to assess prospects for future success. A study by Ernst & Young LLP revealed that research analysts make extensive use of nonfinancial data when evaluating companies and base about 35 percent of their buying and selling decisions on nonfinancial performance areas such as employee satisfaction, customer satisfaction, and operational efficiency. The study concluded that strong and well-communicated nonfinancial performance could improve a company’s ability to raise capital. The message is clear: Nonfinancial performance indicators can be used as leading indicators to help predict future financial performance.


The need to expand performance measurement beyond financial metrics has turned what used to be a relatively simple reporting exercise into a complex process. Companies are trying to make sense of a bewildering amount of data and a variety of technology tools that access, organize, analyze and utilize data to achieve strategic business goals. What types of data should be looked at, and what kinds of technology do you need?

The answers lie in knowing what to measure. The metrics you define will determine what types of business intelligence to use; which types of data to focus on; and what kinds of software tools will best support your performance reporting.


Determining the right metrics for your company begins with accurate goal setting. Find a company that’s successful in setting goals and targets, and you’re likely to find accurate forecasting techniques.


Driver-Based Forecasting

Frequently, companies base their budgets, planning, and forecasts solely on their previous year’s performance. However, some other factors — such as competitors’ performance, the performance of benchmark companies in similar businesses, the company’s capabilities, and input from suppliers and employees — form the foundation for forecasts that reflect the realities of business.


Many CEOs can attest, poor forecasting can ruin a company. For instance, if a company underestimates its demand, then production lines are understaffed, which forces labor to work overtime at a wage premium. The scarcity of raw materials on the shop floor forces buyers to accept higher prices as they’re rushing late to market, and added costs put a stranglehold on cash flows. The overestimation of demand has just as many penalties.


Having massive amounts of data available at your fingertips has not improved forecasting. In some cases, the deluge has blinded managers because they can’t decipher what’s important. By limiting the amount of information and by capitalizing on technology’s speed, you can make better forecasts faster, using driver-based forecasting.


Driver-based forecasting identifies factors that drive revenue, cost, and profit. These factors can be internal or external, such as seasonal changes’ effect on raw materials (such as heating oil) or a strong economy’s impact on the availability of labor. The driver must have materiality and volatility to qualify for inclusion in the forecasting model.


“Materiality” refers to financial line items that have the greatest impact on overall costs for an organization. “Volatility” means how much that material factor varies from period to period. Is it material? Is it volatile? If both answers are yes, some measurement of it needs to show up in a timely report to management, according to Eyermann.


When General Motors Corp. (GM) implemented a pilot driver-based forecasting program at one of its divisions, GM Powertrain Group, Pontiac, Mich. Individual plants no longer forecast hundreds of financial line items; instead, the division’s central office generates forecasts. Financial managers now focus on what’s important to the company by looking at a few critical drivers that have the greatest impact on its costs.


Forecasting Tools

Refining focus is one step toward improving forecasting. The second step is knowing how to enhance forecasting accuracy. That’s where technology can help.


Financial managers can find key data for forecasting through a variety of Internet-based software tools. These include tools such as “neural learning agents” that correlate disparate sources of information, and tools that analyze workflow inside the core company and throughout the enterprise.


Neural learning agents refer to software technology that can help organizations predict performance. Let’s say, that you wanted to focus on cost-efficiency metrics. The software can detect unnecessary costs such as fraud propagated by employees who may create a dummy supplier or take payments from clients and randomly deposit those checks into separate accounts. Without a detection mechanism like that in place, it might take a few months before anyone realized that a customer’s account was delinquent, and by then the employee responsible for the theft would be long gone. This type of software can nip a problem like that in the bud.


The software also can correlate disparate and multiple sources of information to develop forecasting models. It’s Internet-based software that can be trained to look for certain patterns of behavior that may affect a performance metric. For example, a performance metric might focus on improving suppliers’ efficiency. The software can uncover information from disparate sources about a particular supplier that has, for example, been late with deliveries or lax on product quality. Then it can look at the supplier’s financial condition, including its manufacturing processes, credit rating or Dun & Bradstreet Corp. rating. The data can be organized into a report that predicts the likelihood of future problems.


These Internet-based tools also can reveal a customer’s financial situation and its potential ability to pay for products purchased. It can give you data that allows you to decide whether you want to do more business or less business with a particular customer in the future. For example, the data may tell you a customer is trying to penetrate a new market, and you can take advantage of that knowledge by telling the customer that you can offer products and capabilities to meet their needs within this new area of interest.

The internet-based software can locate lost, missing or inaccessible data, the kind of data that reveals how to do more business per customer. Suppose you are in the airline maintenance industry and you want to generate greater revenues. You can use software tools to access key bits of data such as information on spare parts, the cost of labor and pricing, which can help you compete more effectively.


Electronic workflow supports a metric such as customer retention rates by giving customers more timely account information. It provides customers and vendors with Internet access to information on things like the status of their invoice or the status of an order by placing those things online. That enhances customer-service performance by improving response time to their needs.


Determining Key Numbers

To make more accurate predictions, finance executives need reliable performance measures that reveal where true strengths and weaknesses lie. Which key numbers drive the success of each business unit and support company goals?


If a companywide objective is to earn 10 percent more revenue, the key numbers within the sales department might be sales volume and gross margin. If a company goal is to increase customer-satisfaction ratios, then a key number that the customer-service area might monitor and improve is the number of customer calls that are answered within three minutes.


A metric may appear to be a good fit for a company’s needs but could adversely affect another important metric. It’s the difference between looking at what’s called multivariant performance metrics vs. a univariant metric. People tend to look too narrowly at the effect just one metric will have. Metrics tend to interlink and cluster.


For instance, looking at the error rate may not be enough. At one plant where McInnes worked, the financial department tracked the correlation between the error rate and employee turnover. As turnover increased, the number of errors increased as well. It showed us where to direct our efforts for improving the error rate.


Dissecting Metrics

A metric such as turnover rate or research and development (R&D) spending may not give an accurate picture. Underneath that single key number, there may be other, more revealing figures that explain why a key number is what it is.


Suppose one company goal is to reduce costs by 10 percent across the board. In that case, one key number might be the annual cost of research and development. You may find that the cost of R&D has gone up 10 percent each year over the past three years, which seems exorbitant. However, go a step further: Ask what those higher costs have produced. If they’ve led to three new products that have all become sales revenue leaders, the increased cost of R&D may be justified.


The traditional accounting system looks at a $500 credit to a customer and says, ‘What’s going on here?’ However, if you look deeper, you may see that the $500 was an investment that strengthened your company’s relationship with the customer.


Getting Buy-In

The most important aspect of all could be buy-in, not only among the top brass but among those who will use the forecasts and metrics.


If employees don’t understand the value of the metric, they may devise ways of appearing to support it and upset customers in the process. At accounting firms I’ve worked, where a prime performance goal is a higher number of billable hours, accountants were notorious for padding their billable hours. The trouble was, the firm inevitably ended up receiving payment on only 60 percent or so of the total number of hours billed because clients refused to pay the rest.


Workshops, board games, and computer-based business simulation games can all be used to educate employees about metrics. Players learn the terminology of the performance system and how to focus on areas that need improvement. Some games are customizable and relate to an individual company’s key numbers and objectives.


Effective performance measurement programs based on sound forecasting techniques can improve a company’s share price and its ability to compete for capital. On the other hand, if performance reporting consists of only a record of past financial performance, you may be handicapping your company’s ability to attract investors and to grow.