Business Analytics Blog

Trend Forecasting

KPI measurement as a snap shot in time is a risk in itself. Without trend forecasting of the metrics used to assess KPI, there is no indication of whether a business or market is on the mend, on the way out, or what may be the cause of a problem. Seasonal trend analysis and the ability to "normalize" metrics are the only true way to benchmark with other outside sources or company history.

KPI assessment and measurement is purely speculative without reconciliation and verification of the core data. So when someone asks if you can really measure risk, the answer is yes. Do most practice reconciliation? Most businesses reconcile their bank balance. Few reconcile inventories more than taking annual inventory. Some reconcile floor plan accounts only because the lender sends a representative periodically for flooring checks. But for the most part, most businesses do not reconcile or verify the data enough to rely upon it. The adage, "garbage in-garbage out," is especially true with KPI measurement.

Steps to identify the direction of any metric, company or market:

• Calculating seasonal trends of the metric
• Remove trends from data (Normalize or Seasonally adjust)
• Apply linear trend to the normalized metric and chart on a graph comparing quarterly moving average of the same normalized metric
• Monitor actual metric with forecasted metric on an additional graph

Proper grouping of related metrics and monitoring relative movement is also key to understanding change. Viewing mounds of printed numbers without visualizing the trends and relative timing of key metrics becomes overwhelming for the most intelligent analyst.

Assessing KPI strategies without the history & the tools of budgeting and forecasting software is asking for failure. Frankly, without two plus years of solid monthly data, my advice is to asses risk at your own risk. It has come time that the lender, CEO, department manager, and even the employee should require monthly metrics to monitor performance and bring awareness of changing trends. This practice nurtures responsibility and "taking ownership".

KPI monitoring and based totally on mathematical analysis is ridiculous. With the aid of a few tools combined with solid judgment, book keepers and managers can become analysts and consultants.

Predictive analytics is on the forefront of analysis. It also encompasses "Relative timing" of other metrics. When metrics are compared to other metrics, there are patterns that emerge, which cannot be seen by traditional assessment. However, understanding that the change in one metric will relatively have an effect on other metrics, which provides unfair insights resulting in advanced business decisions.

For example; in the recreational vehicle industry where discretionary spending is driving sales, large ticket sales historically had a lag time behind inverted prime. Fuel prices and a few other economic metrics impact sales as well. Understanding predictive analytics requires knowledge of relative timing and reaction between key metrics. With the understanding of inverted prime, fuel prices, new home construction starts, an RV company could adjust large ticket inventories and adjust for the market change with a couple months advanced lead time.

Forecasted seasonally adjusted COGS allows calculations of target seasonal inventory levels based on desired annual turnover. When seasonal trends are removed from sales data, a rolling quarter moving average compared to the linear sales trend visualizes market change directions. When collectively monitored with inverted prime, you can add judgment to the equation, and adjust inventories in advance. The understanding of monitoring metrics holistically allows the birds-eye-view into the future.

KPI Monitoring

...is pointless unless the right combination Key Performance Indicators (KPI) are grouped and evaluated together. KPI definitions and importance are different for many companies' and/or study groups. A ratio is Key if it is actively measured, monitored and managed and deemed "Key" by management. I have listened to and read many discussions regarding the evaluation of specific KPI measurements, arguing "which KPI is best?" The answer is none. More times than not, the most adamant conversationalists often have become near sighted, and ignored other crucial indicators. Furthermore, since nearly every business is seasonal, traditional KPI formulas used by most, are useless without applying seasonally adjusted annual rates (SAAR), trend analysis, and seasonality forecasting.

For example; if May is the beginning of the selling season, October is the end, and there is the same inventory on hand in May as in October, do you have the same days inventory on hand? The answer is absolutely not. Do you think improperly calculated KPI by not allowing for seasonal trends could affect turnover, return on assets, target inventory? Unquestionably, yes. Can you accurately predict target inventory levels without knowing your desired turnover, and forecasted COGS based on seasonal trends? Only if you guess!

The Wright Brothers Had Only Three Instrument Gauges: A stop watch, tachometer for engine speed, and an anemometer to measure wind speed. Those indicators were sufficient at the time, but the Silver Dart did not fly very far compared to the space shuttle of today. By today's standards, inventory based KPI, such as gross sales, gross profit and current inventory will get you about as far as the Wright brothers' first flight. Without data visualization technology, could you imagine astronauts flying the space shuttle... with only the help of reports... reviewing historical data after the end of each flight? Try combining the analogy of "hitting the ground at full speed" with "its water under the bridge". It's called your broke and it's too late!!!

In today's business climate, understanding the complexity and relativity of KPI is crucial to long term success. In time, ignorance will land most businesses out of business. Predictive analysis and seasonality forecasting with KPI measurements of inventory/products based business, should be performed on every product and brand to optimize product mix. Product analysis requires a combination of sales analysis and Inventory analysis. Below are samples... but not all of Product Analysis metrics.

Product Analysis (By Manufacturer, Brand & Body type)

• *Gross Margin %
• Gross Profit $
• Sales seasonal trend
• Gross Sales
• *Sales (SAAR)
• COGS
• COGS seasonal trend
• COGS Forecast
• Forecasted COGS
• *Target Inventory
• Current Inventory
• *Last Inventory
• Average Inventory
• *Days Inventory (SAAR)
• *Turnover (SAAR)
• Desired Annual turnover
• *Return on Asset
• Unit Sales Annualized Linear Trend
• Unit Sales quarterly moving average (SAAR)
• *Warranty Claims % of COGS
• *Warranty Shortages % of Warranty Claims

All of the above metrics are important, but not all are necessary to monitor. Some metrics are used only to calculate other KPI which are regularly monitored*. Seeing pages of metrics in a report can be overwhelming for some. Inability to see KPI trends over time on a graph dilutes usefulness. But once the results are properly grouped and displayed visually, analysis becomes much easier to understand. Trends can be identified, and actuals benchmarked against forecasts, trends and budgets. With better understanding of history and trends, better management of inventory and sales decisions in the future can be made.

Market trends change quickly. National trends do not always represent company trends. Predictive analytics and seasonality forecasting allows a birds-eye-view to see trends and adjust business practices to changes in a specific market. Most businesses did not see the market change until August of 2008. Solid indicators warned of the change four months prior to the crash for many businesses. Analytic tools would have given most companies time to adjust inventories and overhead in the height of their selling season.

View the sample companies total unit sales market trend. The blue straight line represents the general direction of company sales. The linear sales trend is calculated using 12 months rolling history after removing seasonal trends. The Red trend line represents the Quarterly moving average of actual sales after removing seasonal trends.

This graph provides a solid understanding of when the market conditions are changing. Keep in mind that trends can change due to outside conditions which are out of control of the business such as the recession of 2008-2009. Internal factors such as inventory management and sales processes can also change trends.

A scorecard for each manufacturer or product line should also be readily available to identify product trends, aid in inventory management as well as assist purchasing negotiations. Manufacturer's representatives respect solid market information and are at most times more willing to help resolve real inventory problems rather than apply unsubstantiated pressure to increase inventory levels.