In the second part of our series on the future of forecasting, we identify what you could be missing out on if Excel is your main go-to when identifying your leading indicators. Read on, and you’ll find out the advantages of using a forecasting model over Excel to identify leading indicators.
Here are the limitations
The number of leading indicators
It’s a numbers game. Why not get the highest-performing forecast by getting access to a large number of leading indicators.
The regression function built into Excel can only handle 16 variables. Using four lags for each variable limits the total number of leading indicators to only 3, as the main variable needs to be lagged as well.
Indicio can analyze up to 50 leading indicators, giving you the advantage of identifying the valid indicators among all the noise.
Are your leading indicators relevant to your forecasts, and how do you determine that?
One of the ways to do so is to conduct a quality check to minimize spurious correlations and wrong interpretations. Unfortunately, Excel does not offer the possibility to double-check this. With only rigid indicators in place, you could run the risk of generating an inaccurate forecast, leading to potential losses.
By employing automated forecasting based on the latest academic algorithms, you can quickly identify the specific indicators that enhance your forecast and rule out those that hinder better performances.
The relationship between variables and their evolution is no longer clamped to their past values; rather, it evolves in real-time with your sales.
Simple linear model
When making critical business decisions, you want to have a pool of models to capture and assess as many perspectives as possible.
With Excel, you only have a simple regression tool, and relying on one model’s insights presents a risk. Having access to a rich library of econometric and machine learning models allows you to quickly and accurately identify the model that offers the best performance at any horizon.
Indicio automatically applies a rich library of econometric and machine-learning models to identify the model offering the best performances at any horizon.
Advantages of using a forecasting model to identify leading indicators
The forecast horizon is not limited to lags
When using correlation to identify leading indicators in Excel, the user relies on a coefficient associated with previous observations of the data. Therefore, the underlying forecast is directly dependent on the chosen lag. However, building a model to identify indicators does not rely on a set amount of previous observations. Indicio identifies relevant indicators by including or excluding variables or both to a model and assesses whether they improve forecast accuracy. Hence, the user can choose the forecast horizon without dependency on lags.
Captures indicator group effect
As aforementioned, Indicio builds an econometric model (VAR) to identify leading indicators. Vector autoregression is a workhorse model in macroeconomics that defines each indicator as a function of other indicators. This way, instead of treating each indicator’s impact separately, the model captures interactions between them and their influence on your sales.
Penalizing funny correlations
95.86% is the correlation coefficient between the number of civil engineering doctorates awarded and mozzarella cheese per capita consumption between 2000 and 2009 in the United States. Does this mean that one should look at Ph.D. candidates to forecast cheese consumption? No, and Indicio does otherwise. Coupling LASSO to the VAR model and tuning it to minimize forecast error, Indicio can select valuable features in the data and discard the unnecessary information.
Forecast leading indicators
Another perk offered by modeling the variable of interest and the indicators is the possibility to forecast all of them simultaneously. Each indicator is given its development path as part of the VAR model, interconnected with all other indicators used as input. The relationship between variables and their evolution is no longer clamped to their past values; rather, it evolves in real-time with your sales.
Easy to verify results
The backbone of Indicio is to make statistics accessible to everyone. Indicio creates previous forecasts that you can compare against actual data to backtest and evaluate model performance. These forecasts are visually available on the platform, making it easy for the user to see how a model would have performed in the past. The forecaster can follow what the model suggested at any given time to better understand the dynamics of its sales or capacity.
Find out how to use the same methodology to forecast your data by booking a free demo.