KPIs (Key Performance Indicators) Help Firms Monitor and Improve Productivity

September 18th, 2014 by

What do we need to do to improve our productivity and as a result our profitability? This is the key question all firms need to ask and answer. If a business is going to achieve sustainable profitability, it has to first convince customers to place orders for goods and services. Second, the business has to deliver these goods and services effectively and efficiently. Effectiveness implies that the goods and services are acceptable to customers and will result in repeat orders. Efficiently implies that the cost of providing these goods and services (direct costs + overhead costs) are substantially less than the revenue generated from the sale of these goods and services.

In the past Income Statements and their associated row/column budget comparison reports were the only means whereby managers and executives could determine whether economic goals were being achieved. Unfortunately there is a significant flaw in this rudimentary financial/operational control system. Firms might spend hundreds or thousands of hours creating detailed budgets, but these budgets are instantly out-of-date. If the mix of products and services actually sold differs from the budget (and they will), the budget will no longer be accurate. You could constantly revise the budget, but that’s an expensive proposition that does not really enhance business or operational control.

The delivery of goods and services (direct operational costs) and the process of managing a business (indirect business management costs) can be viewed as the sum of many activities, each of which can be associated with some form of cost. These costs then flow into the Income Statement. Rather than trying to control these dollar costs which have no lasting budget values, firms should identify, budget, control and improve these key business activities. If these activities can be controlled, the Income Statement will take care of itself. Key Performance Indicators (KPIs) are the measure by which these critical business activities are measured, controlled and improved.

KPIs are graphical representations of key business drivers. Rather than presenting this key data in a tabular format as you would find in an Income Statement, KPIs display the same data in a graphical format that makes it easier for users to interpret. That’s all well and good, but the KPI needs to represent key business metrics and be displayed in a format that leads to some form of action. In most instances KPIs display key data (dollars or percentages), but the user has no way to understand whether the information presented is good, bad or indifferent. As a result the user does not understand whether they need to take action or not.

This concept of “taking action” is one of the keys to the successful implementation of a metrics based business evaluation and improvement system. Accounting and ERP systems can generate directly or indirectly vast amounts of data and metrics (e.g. inventory turns or labor as a percentage of revenue). Managers do not have time to review lengthy reports and identify specific business processes or conditions that require their attention. Instead managers need to be able to quickly identify where they need to concentrate their attention. This objective is the underlying reason why KPIs were created in the first place. Unfortunately executives and managers don’t fully understand how KPIs should be utilized to assist them identify specific aspects of their business that require improvement. In addition many KPI graphs support a control system that reacts to negative issues after they have already created a problem rather than proactively identifying issues before they create a problem.

Virtually every accounting or ERP system offers some form of user dashboard with various Key Performance Indicators (KPI) displays. These KPIs can be displayed in any number of formats (Pie Charts, Bar Charts, Line Charts, gauges, etc.) and most look very professional. Unfortunately in many cases these graphical KPI displays really accomplish very little and that’s the issue we will discuss in this article.

Pie Charts do not support decision makingPie Charts, Bar Charts and Income Statement snippets do not help users determine where they need to concentrate their attention.

Pie Charts are what can be labeled as single point KPI displays. They are constantly associated with the marketing of KPI solutions, but fail to present any context by which users can determine if some form of remedial action needs to take place.

Pie Charts display a specific business condition (e.g. Top Ten Customers) at a specific point in time, but like all snapshots fail to assist users take any form of action.

An effective KPI based business evaluation and control system must let users “see” where the business condition has been, where it is today and where it appears to be heading. Pie Charts present a single piece of information and it is therefore impossible to base any decision on just such a snapshot.

Most Bar Charts do not support the decision making processBar Charts and Income Statement snippets look pretty, but do not help users understand whether they need to take action.

Pie Charts and Bar Charts are the darling of firms selling dashboards and KPI displays. They certainly look nice when displayed on a user’s dashboard, but do they really help users take any form of action?

A Pie Chart or Bar Chart of a firm’s top ten customers looks nice when displayed, but does it really tell a story or precipitate an action? A table that displays specific data extracted from an accounting or ERP system or a Bar Chart or Pie Chart is no more than a picture of a specific condition at a specific point in time.

What does a Pie Chart displaying a firm’s top ten customers accomplish? It’s no more than a picture taken at the end of the month selected. Does it precipitate an action based on the information presented? Does the Income Statement snippet pictured below provide any information other than the past month’s results? The summary values for the Income Statement for last month are displayed well, but tell no story that indicates whether the key values over time are increasing or decreasing. In fact there is no context for any of these KPIs that allow the reviewer to do anything! It’s just data that provides no direction. The reviewer learns nothing and therefore the Pie Chart, Bar Chart or Income Statement snippet is useless.

Bar Charts suffer the same potential weakness as Pie Charts. There is a great deal of information that can be displayed in a Bar Chart, but for the most part this information reflects a specific business condition at a specific point in time. The display may look interesting, but it doesn’t precipitate an action. Effective KPIs must help users immediately “see” whether a business condition needs to be improved. That requires a time-phased picture of the condition (e.g. each of the last 12 months). Most Bar Charts do not present time-phased information and therefore are not effective business improvement tools.

Income Statement SnippetMany ERP dashboards contain snippets of Income Statement information or some other row and column GL based display. While the information may be interesting, again there is no context to determine whether improvements need to be made.  Although you could say that a negative variance of a line item is sufficient proof, previous months may have generated a significant positive variance that far outweighs the negative variance of the displayed data. As we will discuss, a single data or KPI value means nothing. What’s more important is the time phased pattern that demonstrates a trend that requires a user’s attention. Effective KPI displays must concentrate on this pattern recognition.

Time Phased Bar ChartTime-phased Bar Charts help, but the benefit is marginal.

Look at the Bar Charts to the right. They contain information that utilizes the context of time and therefore makes it easier for users to understand whether the information displayed “tells a story”.

This is really the key to the effective utilization of KPIs. In the case of the first Bar Chart to the right the KPI appears to be increasing over the past 12 months and the rate of increase also appears to be decreasing toward the end of the 12 month period. This may give the user sufficient information to determine if some form of action needs to take place.

Variable Bar ChartNow let’s examine the second Bar Chart to the right. While the first Bar Chart seems to indicate a fairly even history, the second Bar Chart does not appear to present a pattern that can be interpreted easily. In addition there is one critical piece of information that is missing. What should the value be at the end of each month? If KPIs are going to be utilized effectively, users need to create targets and measure performance against these targets. That gives users the full context they require to understand whether action is warranted. As we have been discussing, users need detailed information to determine whether they need to take action. If a KPI image does not give them all the information they need, it is not possible to take action.

Bar Charts with target linesBar Charts with targets are better, but are still difficult to interpret.

Notice in the example to the right that we have added a straight line to indicate the target value for this KPI. Unfortunately interpreting the highly variable Bar Chart is somewhat difficult and the imposition of a target value doesn’t really help users understand whether the KPI implies that some form of action is required.  The target line does make it easier for users to relate actuals to targets, but only if the bar chart is relatively static over time.  In this case the the Bar Chart itself is variable to the point that it’s not really possible to draw a conclusion regarding the behavior or potential behavior of the KPI.

Line Graph with Target and ActualLine Charts are much easier to interpret.

Some Bar Charts do paint an effective scene, but they can be difficult to interpret. A Line Chart gives users a clear picture of conditions over an extended length of time as you can see in the graph to the right. You can immediately see the entire picture and determine fairly easily whether action needs to be taken. If the actual results are below the target, the KPI seems to indicate whether action is warranted. In this case though the actual results over time seem to line up with the target and therefore no action should be taken even though the actual results for some time periods are below target.

Auto Alerts may not work properlyTake care when setting up and utilizing system generated alerts.

The graph to the right is a variation of the previous graph and once again the long-term difference may not warrant action. This is one of the issues precipitated by KPI displays or exception management systems that flag a KPI display if the actual results differ from the target. Other KPI displays may indicate that an issue requires action, but that can be deceiving. If the KPI in the example to the right measured inventory turns, there are three separate instances where an exception event would be created when in fact the long-term position does not appear to warrant action.

Line Graph with Confidence IntervalsIn some instances long-term KPIs need to be studied before any action is warranted. As we will discuss later, KPI deficiency notices should not be based on a single point-to-point comparison. Instead the target should encompass a range of acceptable values (target value ± defined %) as in the example to the immediate right. Statisticians would refer to this use of ranges as confidence intervals. If the actual results fall outside the confidence interval (target range), then the system would indicate that action may need to be taken.
KPIs are not values that automatically create an alert. To some extent this can be thought of as a knee-jerk reaction whereby the system operates with no oversight from users. If users are going to adopt a proactive approach productivity management, this reactive approach actually reduces the effectiveness of the entire system. Managers and executives must set the tone, not the systems.

Extreme data variabilityExtreme data variability may require some form of data smoothing.

In some instances the actual results of data representing specific KPIs may display a highly volatile history as the first graph on the right indicates. While target values may exist, the volatility of the actual results makes it difficult to determine if any action needs to be taken. In this case the data needs to be “smoothed” by utilizing some form of trend analysis adjustment that will translate the display into a format that is more easily understood. While it may seem that we are moving away from traditional dashboards and KPI displays, if the information as represented by the KPI is of significant interest to the firm’s objective of improving productivity, then we must adopt a more scientific approach so we can achieve the objective.

Data SmoothingThere are many different ways data can be smoothed. Although the raw data in the image to the right is volatile (the green data points), the smoothed line (the solid green line) is much easier to read.

A linear trend line is a best-fit straight line that is used with simple linear data sets. Your data is linear if the pattern in its data points resembles a straight line. A linear trend line usually shows that something is increasing or decreasing at a steady rate.

A logarithmic trend line is a best-fit curved line that is most useful when the rate of change in the data increases or decreases quickly and then levels out. For example, when you start a business, sales might increase quickly. As the business matures, increases in sales are likely to be more gradual.

A moving average trend line works well for data that fluctuates higher and lower. It smoothes out fluctuations in data to show patterns or trends more clearly. You select the number of data points that you want the trend line to average and use as a point in the line.

Although the creation of trend lines makes KPI displays easier to visualize, two cautions must be taken into consideration. First, you have to study the pattern of each KPI data display so that the most appropriate trend line can be created. In some cases a linear trend line would be appropriate while in other cases some other smoothing formula would be most appropriate. Second, trend lines tend to discount the most current period’s data and therefore you have to create target values that are appropriate for the trend lines you are going to create.

Line Graph using predictions of possible futuresAdopt a proactive KPI based exception management system.

As we all know business success is based on taking action before you are forced to do so. If you can anticipate change rather than reacting to change, the likelihood of success will be improved. The same argument holds true for KPI based exception management systems. Rather than waiting for a KPI to become unacceptable, users should extend actual data into the future as the image to the right indicates. If the trend line is extended into the future and breaks through a confidence interval, users then have the ability to examine the trend and take action before action is actually required.

Define KPIs that effectively represent key business activities.

As we discussed earlier, GL based budget comparisons measure cost as the driving influence. That’s fine, but cost doesn’t really represent how you are running your business. In essence you have to measure how well you are “doing the work”. In many instances time is the key business driver. How long does it take to create a sales order and ship the merchandise? What is your inventory turns? How long does it take to receive goods once a purchase order has been issued? How long does it take to issue a purchase order? What is the average days late for overdue customer invoices? All of these questions are associated with specific business activities that you define as being important. Each of these questions can then be associated with a KPI. If you manage your business activities or processes effectively, the costs associated with those activities with take care of themselves and therefore the Income Statement will take care of itself.

Take action to improve underperforming KPIs.

Identifying business processes or activities whose KPIs are below assigned targets is but the first step in the process of managing your business operations effectively. The second, and perhaps most important step, is taking action that will improve the KPI. This isn’t usually a single action, but a series of actions or decisions taken by multiple people that collectively have a positive influence on one specific KPI.

Many accounting or ERP systems support the notion of a dashboard that displays KPIs, but there’s something missing. Once an underperforming KPI has been identified, the steps people take to improve the KPI are almost entirely manual. If you are going to effectively “manage” your KPIs, there should be some form of system that gives you the ability to analyze the business activity, collaborate with other people in your organization to determine what needs to be done to improve the business activity (as represents by the KPI) and track the results of the changes you make.

The key concept here is the notion of collaboration. This is particularly true for larger organizations where multiple people are interested in a single business activity. Each of these people should contribute their thoughts regarding what needs to be done to improve an underperforming business activity and that “discussion” needs to be supported by some form of software application. There are literally hundreds of collaboration applications that meet this requirement. Some are project management oriented with all of the appropriate bells and whistles while other collaboration applications are much less complex. Microsoft Office 365 supports document sharing and that may be sufficient as the document being shared is designed to give people the ability to contribute to a discussion regarding the improvement of a single business activity. The key here isn’t just contributing to a single document. People also need to be able to attach files (e.g. spreadsheets, etc.) that support the business activity analysis.

The collaborative business activity analysis and improvement isn’t just a single step. While it’s certainly important to analyze the business activity and determine what specific steps need to be taken, the KPI representing the business activity needs to be monitored over time to determine if the changes made are having a significant positive result. That’s why it’s critical to measure the KPI over time (usually monthly), determine if the trend in the KPI is acceptable, return to the improvement document if required, continue the collaborative analysis and make additional changes.

Improving KPIs isn’t a single step taken by one person. It’s a continuous improvement process that requires the collaboration of multiple people, all of whom are jointly responsible for efficiency and productivity. While KPIs help users identify business processes and activities that require improvement, it’s what happens after that that leads to success. KPI owners lead the analysis and improvement process, but all of the people who are connected to a business activity must participate in the process, lending their experience and analytical skills and supported by collaborative software applications that help people focus their attention on achieving success.

Conclusion

Key Performance Indicators are graphical representations of business activities that indicate whether these business activities are meeting pre established targets. While users can create any number of graphs and charts that represent the status of a business condition, those displays accomplish nothing if there is no context by which the KPI can be interpreted. Pie Charts and many Bar Charts present but a single snapshot of a business condition (e.g. Top 10 Customers). Users cannot take action based on a single data point and therefore such charts are useless (unless you are making some form of presentation).

The efficiency and effectiveness of business activities can only be judged if that business activity is viewed over time. This time-phased view gives users the ability to determine if the business activity as represented by their KPI is improving or slipping. If the KPI falls below its targeted value or the trend graph indicates that the KPI is moving in a negative direction, users can then take proactive action, hopefully before the KPIs value actually moves into a negative position.

If the KPI needs to be improved, the owner of the KPI can analyze the business activity, take corrective actions and then track the KPI to determine if the steps being taken appear to be having a positive influence. If several people are connected to the same business process as represented by the KPI, all of them are jointly responsible and need to utilize some form of analysis tool that helps them collaborate with each other to seek a solution. It is also critical that the analysis and improvement steps be thoroughly documented so that interested parties can refer back to their collaborative “notes”, learn from them and contribute more effectively to business activity analysis and improvement.

KPIs are not just graphs. Rather they are important contributors to business process analysis and improvement. If you want to make your business activities more efficient and effective, you have to create a system that identifies where you have been (past), where you are today (present) and where you appear to be heading (future). If the displayed KPI indicates that improvement is required, then a fully functioning KPI system helps you and others in your firm collaborate with each other and make changes where changes need to be made.

Share and Enjoy
  • Print
  • Digg
  • del.icio.us
  • Facebook
  • Google Bookmarks
  • LinkedIn
  • Reddit
  • StumbleUpon
  • Twitter

Leave a Reply

Navigation Hover