Three B2B Cases: Leading Indicators in Sales Forecasting

Three B2B Cases: Leading Indicators in Sales Forecasting

A peer-reviewed study tested external leading indicators in three industrial B2B cases. Reported error reductions are study averages—not a Morrow performance claim.

A peer-reviewed study tested external leading indicators in three industrial B2B cases. Reported error reductions are study averages—not a Morrow performance claim.

Business Cycles

Why Past Sales Don't Know the Next Turning Point

Many sales forecasts are based on an understandable principle: trends and seasonal patterns are derived from past sales and then projected into the future.

This can explain how a company has developed so far. The more difficult question is how its market will change.

Traditional statistical methods struggle to detect an approaching economic downturn when it is not yet visible in the company's own sales. Companies therefore often supplement their forecasts with manual adjustments or expert judgment.

These assessments contain valuable experience. But they can also be influenced by optimism, recent large orders, internal targets, or particularly prominent individual opinions.

A 2020 academic study therefore took a different approach. The researchers developed a method that automatically selects relevant macroeconomic leading indicators and links them with a company's internal sales development.

The approach was tested using three real-world corporate cases from different geographic markets. The companies operated in industrial B2B environments and included a supplier to the tire industry as well as manufacturers of building materials and steel.

What the Three Case Studies Show

The results were compared using the mean absolute percentage forecast error. Put simply, this metric shows how far forecasts were, on average, from the values that actually occurred later.

Across the different forecast horizons and corporate cases, the proposed method achieved:

• an average relative error reduction of 54.5% compared with a simple benchmark forecast,

• an average improvement of 25.6% compared with the best conventional forecasting method in each case.

These figures do not mean that every corporate forecast automatically becomes 54.5% more accurate. The results varied by dataset and forecast horizon.

For one of the companies studied, traditional methods were at times more successful over longer forecast horizons. This limitation is crucial: no model automatically delivers the best results in every situation and for every company.

The advantage did not come from one particular statistical method alone. What mattered was the combination of:

• suitable internal data,

• relevant external drivers,

• a limited and transparent selection of indicators,

• testing with data that were not yet known when the model was built.

Why the planning horizon matters

In the three industrial B2B cases, the researchers combined company sales histories with selected macroeconomic leading indicators. The aim was to detect relevant demand changes early enough to inform decisions beyond the next few weeks.

The paper reports an average relative forecast-error reduction of 54.5% against a naive benchmark and 25.6% against the strongest conventional comparison method. These are results of the evaluated cases and horizons, not a guaranteed improvement for another company.

What teams should take from the research

More time to see a potential demand shift can help finance, sales and operations discuss staffing, purchasing and capacity before the change appears in internal orders. But the paper also shows that performance varies by company and horizon; conventional methods sometimes perform better.

Morrow makes the principle of combining internal sales data with external indicators accessible. It does not inherit the study’s result automatically. A useful deployment must select relevant indicators for the company, back-test against a fair baseline and communicate uncertainty.

Want to know whether external indicators can improve your company’s forecast?

Want to know whether external indicators can improve your company’s forecast?

Three B2B Cases: Leading Indicators in Sales Forecasting

Three B2B Cases: Leading Indicators in Sales Forecasting

A peer-reviewed study tested external leading indicators in three industrial B2B cases. Reported error reductions are study averages—not a Morrow performance claim.

A peer-reviewed study tested external leading indicators in three industrial B2B cases. Reported error reductions are study averages—not a Morrow performance claim.

Business Cycles

Why Past Sales Don't Know the Next Turning Point

Many sales forecasts are based on an understandable principle: trends and seasonal patterns are derived from past sales and then projected into the future.

This can explain how a company has developed so far. The more difficult question is how its market will change.

Traditional statistical methods struggle to detect an approaching economic downturn when it is not yet visible in the company's own sales. Companies therefore often supplement their forecasts with manual adjustments or expert judgment.

These assessments contain valuable experience. But they can also be influenced by optimism, recent large orders, internal targets, or particularly prominent individual opinions.

A 2020 academic study therefore took a different approach. The researchers developed a method that automatically selects relevant macroeconomic leading indicators and links them with a company's internal sales development.

The approach was tested using three real-world corporate cases from different geographic markets. The companies operated in industrial B2B environments and included a supplier to the tire industry as well as manufacturers of building materials and steel.

What the Three Case Studies Show

The results were compared using the mean absolute percentage forecast error. Put simply, this metric shows how far forecasts were, on average, from the values that actually occurred later.

Across the different forecast horizons and corporate cases, the proposed method achieved:

• an average relative error reduction of 54.5% compared with a simple benchmark forecast,

• an average improvement of 25.6% compared with the best conventional forecasting method in each case.

These figures do not mean that every corporate forecast automatically becomes 54.5% more accurate. The results varied by dataset and forecast horizon.

For one of the companies studied, traditional methods were at times more successful over longer forecast horizons. This limitation is crucial: no model automatically delivers the best results in every situation and for every company.

The advantage did not come from one particular statistical method alone. What mattered was the combination of:

• suitable internal data,

• relevant external drivers,

• a limited and transparent selection of indicators,

• testing with data that were not yet known when the model was built.

Why the planning horizon matters

In the three industrial B2B cases, the researchers combined company sales histories with selected macroeconomic leading indicators. The aim was to detect relevant demand changes early enough to inform decisions beyond the next few weeks.

The paper reports an average relative forecast-error reduction of 54.5% against a naive benchmark and 25.6% against the strongest conventional comparison method. These are results of the evaluated cases and horizons, not a guaranteed improvement for another company.

What teams should take from the research

More time to see a potential demand shift can help finance, sales and operations discuss staffing, purchasing and capacity before the change appears in internal orders. But the paper also shows that performance varies by company and horizon; conventional methods sometimes perform better.

Morrow makes the principle of combining internal sales data with external indicators accessible. It does not inherit the study’s result automatically. A useful deployment must select relevant indicators for the company, back-test against a fair baseline and communicate uncertainty.

Want to know whether external indicators can improve your company’s forecast?

Want to know whether external indicators can improve your company’s forecast?

Three B2B Cases: Leading Indicators in Sales Forecasting

Three B2B Cases: Leading Indicators in Sales Forecasting

A peer-reviewed study tested external leading indicators in three industrial B2B cases. Reported error reductions are study averages—not a Morrow performance claim.

A peer-reviewed study tested external leading indicators in three industrial B2B cases. Reported error reductions are study averages—not a Morrow performance claim.

Business Cycles

Why Past Sales Don't Know the Next Turning Point

Many sales forecasts are based on an understandable principle: trends and seasonal patterns are derived from past sales and then projected into the future.

This can explain how a company has developed so far. The more difficult question is how its market will change.

Traditional statistical methods struggle to detect an approaching economic downturn when it is not yet visible in the company's own sales. Companies therefore often supplement their forecasts with manual adjustments or expert judgment.

These assessments contain valuable experience. But they can also be influenced by optimism, recent large orders, internal targets, or particularly prominent individual opinions.

A 2020 academic study therefore took a different approach. The researchers developed a method that automatically selects relevant macroeconomic leading indicators and links them with a company's internal sales development.

The approach was tested using three real-world corporate cases from different geographic markets. The companies operated in industrial B2B environments and included a supplier to the tire industry as well as manufacturers of building materials and steel.

What the Three Case Studies Show

The results were compared using the mean absolute percentage forecast error. Put simply, this metric shows how far forecasts were, on average, from the values that actually occurred later.

Across the different forecast horizons and corporate cases, the proposed method achieved:

• an average relative error reduction of 54.5% compared with a simple benchmark forecast,

• an average improvement of 25.6% compared with the best conventional forecasting method in each case.

These figures do not mean that every corporate forecast automatically becomes 54.5% more accurate. The results varied by dataset and forecast horizon.

For one of the companies studied, traditional methods were at times more successful over longer forecast horizons. This limitation is crucial: no model automatically delivers the best results in every situation and for every company.

The advantage did not come from one particular statistical method alone. What mattered was the combination of:

• suitable internal data,

• relevant external drivers,

• a limited and transparent selection of indicators,

• testing with data that were not yet known when the model was built.

Why the planning horizon matters

In the three industrial B2B cases, the researchers combined company sales histories with selected macroeconomic leading indicators. The aim was to detect relevant demand changes early enough to inform decisions beyond the next few weeks.

The paper reports an average relative forecast-error reduction of 54.5% against a naive benchmark and 25.6% against the strongest conventional comparison method. These are results of the evaluated cases and horizons, not a guaranteed improvement for another company.

What teams should take from the research

More time to see a potential demand shift can help finance, sales and operations discuss staffing, purchasing and capacity before the change appears in internal orders. But the paper also shows that performance varies by company and horizon; conventional methods sometimes perform better.

Morrow makes the principle of combining internal sales data with external indicators accessible. It does not inherit the study’s result automatically. A useful deployment must select relevant indicators for the company, back-test against a fair baseline and communicate uncertainty.

Want to know whether external indicators can improve your company’s forecast?

Want to know whether external indicators can improve your company’s forecast?