Aggregate Time Series Forecast

The statistical time series for all of the dealers summed in aggregate shows reasonable accuracy at 8.7% MAPE. The random errors found in each partner forecast effectively cancel each other out when added together leading to a reasonably accurate statistical forecast.

The Change Point

The times series analysis identifies a conspicuous downturn between 5 and 6 quarters in the past. This coincides with the competitive entry, but also coincides with the opening of new markets as the traditional market begins to decline.

This forecast also demonstrates extreme seasonality. This is to be expected in the customer's capital budgeting projects that include the client's products. This volatile seasonality makes accurate predictions using traditional statistical time series analysis error-prone for individual partners.

The Aggregate Forecast

This interactive chart displays the projected order input for all dealers in aggregate. The black line is the actual order input. The blue line is the forecast. The red lines indicate upper and lower bounds for the forecast and the green dots are the underlying trend.

This chart is interactive. The tools in the upper right hand corner will move the chart, magnify portions and display precise amounts when the cross hair tool is centered on a point in time.

The Seasonality

The order input over time demonstrates a strong seasonality of 4 quarters. Due to capital budgeting procedures being fiscal in nature this is to be expected but it is beneficial to have it confirmed.

Time Series Error Pattern

The time series error appears to be more pronounced in quarters with declining performance due to seasonal variations. If we were using this to forecast total order input for a declining seasonal quarter, the error would be nearly twice as large as an increasing seasonal quarter.

Client Challenge Confirmed

There is definitely a downward push on the trend line beginning with the simultaneous impact of competitive entry and market maturity. The client evidently has not successfully entered the new emerging markets.

To assess individual dealer forecasted performance will require statistical time series analysis forecasts for the individual dealers.

Plots By Dealer