In our previous blogs, we’ve seen how a forecasting and demand planning assessment can give actionable insights that enable the planning team to focus on moving towards smart-touch forecasting and increase the quality of their decisions. Furthermore, we have some first results on the effect of decisional guidance: It generally improves forecast accuracy, but a planner’s willingness to accept the guidance can vary.

Planner types

Just like there are many different kinds of people, there are many different kinds of planners. We will highlight three examples:

• An optimistic planner adjusts too heavily, typically in an upward direction, often decreasing accuracy.
• A planner can furthermore show anchoring behavior. The adjusted value is in the right direction, but the adjusted value is too close to the statistical forecast (increasing accuracy, but not attaining the full potential).
• Finally, an overreacting planner adjusts in the right direction but overshoots the actual demand.

A theoretical data scientist might now simply say: “Easy! We identify which kind of planner we’re dealing with, and accept or reject their enrichments based on a fancy machine learning model, thereby maximizing the accuracy of the final forecast”. This is, however, not a productive approach: It is pitting the human and the machine against each other, instead of empowering planners to make the best possible forecast. Instead of ‘human versus machine’, we could be maximizing the potential of ‘human with machine’.

So, why do we care about planner types?

Personalized feedback

Imagine for a moment that you’re talking to two planners, let’s call them Anna and James. Anna is a typical optimistic planner: She has a positive outlook on life in general and that filters through to the demand planning enrichments he makes. James on the other hand is an anchoring planner, and a bit more cautious.

In the real world, talking with Anna and James would be a very different kind of conversation. Similarly, if we want to improve their impact, a planning system should give very different kinds of feedback. To Anna, when she adjusts a forecast upwards, the machine may suggest: “Are you sure? Dramatic upward adjustments like this have shown to be overly optimistic in the past. Please consider both the size and the timing of the uplift.” To James, when he adjusts a forecast upwards, the suggestion may be: “Are you sure? If you have good reasons to increase the forecast, the uplift will probably be more than you think.”

‘Smart-touch forecasting’ through automated enrichments

The examples above are a natural first step. Once a planner has made their adjustment and the system sees a potential for significant improvement, we give targeted feedback. This approach puts the planner in charge and leaves it up to them to accept or reject the machine’s suggestion. This is a small first step to take: When the recommendations make sense and prove their value, over time demand planners will come to trust the machine.

That’s when it’s time for the next step: automated enrichments. Taking as input the time series data, historical enrichments, and other internal and external drivers, we can use machine learning techniques to automatically recommend enrichments. A planner can then focus on validating those suggestions.

With modern advances in the explainability of machine learning models, techniques such as Shapley values (e.g., SHAP), and local surrogate models (e.g., LIME), we can create an understanding of the behavior and outcome of the machine learning models. For instance, the planner can see what inputs, trends, and drivers have caused the recommendation. This further enhances trust in the algorithms and enables the user to make an educated judgment call on the validity of the recommendation.

Conclusion

In the current day and age, with a ‘war on talent’ in the job market, our aim should be to empower and engage our planners. With a step-by-step approach from cognitive insights in a forecasting and demand planning assessment, through personalized feedback, and finally automated enrichments, we can truly enable smart-touch planning. By giving our people the right tools and the right information, they can have a tremendous impact.

If you have questions or would like to discuss your enrichment process, please contact Dan Roozemond, or get in touch with us.

The post Putting the ‘smart’ in smart-touch forecasting appeared first on EyeOn.

As described in our earlier blogs, forecast data enrichment processes are complex. It is not always clear who in the demand planning team adjusted and for what reasons. While in many companies the number of forecast items to take into account has grown exponentially, the experienced ‘old school’ planners who knew the ins and outs of each planning item are gradually phasing out. Today’s demand planning teams would therefore benefit greatly from being supported in a smart way to make effective and focused forecast adjustments. Providing these effective and focused data enrichments is one of the key steps of our cognitive insights approach to improve your forecast by nudging the planner to make better enrichment decisions:

1. Create awareness
2. Assess previously performed enrichments
3. Guide planners in providing effective enrichments
4. Automate predictable enrichments where possible

This blog focuses on guiding planners to provide effective enrichments. But how to achieve this? The concept of decisional guidance provides the answer to that question.

Decisional data enrichment guidance

Decisional guidance was introduced in literature in the ’90s already: Mark Silver defines it as the way a planning system supports the decision-maker with structuring and executing the decision-making process. Within intentional guidance, literature distinguishes two types: informative guidance and suggestive guidance (Montazemi et al., 1996; Fildes et al., 2006).

Informative guidance: Giving a planner unbiased, relevant information without any suggestions on actions to take. Example: “Based on historical sales, average demand was 598 products”.

Suggestive guidance: Proposing a specific action to the decision maker. Example: “Based on historical sales, the system advises to adjust the forecast for March from 123 to 598 products. Do you accept this change?”

As research on decisional guidance in the context of planning and forecasting is lacking, we decided to perform a study on how decisional guidance can be implemented in planning systems to improve the performance of judgmentally adjusted forecasts. Jochem Geurts, master student Operations, Management and Logistics of Eindhoven University of Technology, performed this research.

The research of Jochem consists of a data analysis on a dataset of one of our customers, and an experiment in which we determine which form of guidance works best in what situation.

Data analysis

To validate the added value of forecast enrichment, we used a dataset of one of our customers. The dataset showed the current process of judgmentally adjusting statistical forecasts. Jochem found that in the current situation, on average planners do not improve the forecast performance. Specifically, adjustments made for products categorized as low to medium volatile decreased the forecast performance. This decline in accuracy is mainly due to the already high quality of the statistical forecast (average forecasting error of 29% for statistical forecasts versus 42% for the enriched forecasts). Adjustments made on volatile products did improve the forecast performance. Next to the distinction between the volatility of the products, Jochem also looked at the direction of the adjustment. The analysis showed that planners are good at choosing the direction of the adjustment, but they have difficulties predicting how much higher or lower the new forecast should be. These two findings are combined in the experiment in which decisional guidance is applied.

Experiment

The experiment showed that decisional guidance has a positive effect on forecast accuracy. When making a distinction between informative and suggestive guidance, we showed that both forms of guidance have a positive effect on forecast accuracy and there was no big differences between the effect of these two different types. For products with medium to high volatility, decisional guidance on the size (how much) significantly improved the forecast performance. For products with low volatility, decisional guidance on the adjustment (direction) showed nearly the same effects in terms of performance as decisional guidance on the size . It can be useful to use decisional guidance on the adjustment as extra check for planners if they are certain to adjust the forecasts of products with low volatility.

Next, the experiment shows that the participants were more intended to fully accept advice which involved small changes compared to large adjustments. Since large adjustments typically add most value to the forecast performance, we recommend to be careful with proposing many small adjustments as decisional guidance. It might lead to a decreased willingness to accept the large adjustments as advice. These results will be used to further improve the forecast data enrichment process.

Decisional guidance: towards smart-touch forecasting

The promising results of both the analysis and experiment confirm that providing the right decisional guidance to planning teams on making- and accepting changes can significantly improve your forecasting performance. This is a great step towards true ‘smart-touch forecasting’.

If you would like to know how decisional guidance can improve the data enrichment process in your company, please reach out to Bregje van der Staak.

The post How decisional guidance helps planners make effective data enrichments appeared first on EyeOn.

Boosting your supply chain demand planning capability will allow you to improve the quality of your demand plan, quickly sense plan deviations, and effectively take action to get a grip on demand. But where to start? Many companies find it difficult to pinpoint the root causes limiting current performance, evaluate whether performance issues are related to the enrichment process, and how to prioritize improvement initiatives.

What you need is an objective assessment. This is one of the key steps of our cognitive insights approach to improve your forecast by nudging the planner to make better enrichment decisions:

1. Create awareness
2. Assess previously performed enrichments
3. Activate planners in providing effective enrichments
4. Automate predictable enrichments where possible

A clear overview and concrete action steps to improve your supply chain demand planning capability

We help customers by objectively assessing their current demand planning capability including the impact on supply planning and inventory management. The outcome of the assessment is a clear view of current strengths and weaknesses, identification of quick wins and a signed-off roadmap that describes concrete steps to improve the demand planning capability.

Depending on the needs of your company, the assessment can entail a number of components: a quantitative demand and forecasting benchmark based on a statistical data analysis, or a qualitative demand management review based on the actual state and interviews. You can find more details on the forecasting and demand planning assessment on our website.

The cognitive component in the forecasting and demand planning assessment

In our forecasting and demand planning assessment we focus on the cognitive component. We have a structured approach to assess where planners add value to the supply chain demand planning process. With our assessments, we leverage our dashboards focusing on cognitive insights using forecast performance metrics such as forecast value add. With these dashboards, we provide the planning community with insights into which enrichments were effective in the past and which did not lead to increased forecasting performance.

The following insights can be derived from the dashboards:

• What is the overall impact of the enrichments on the quality of the forecast;
• For which products the forecast deteriorates with enrichments;
• Where does Sales and/or Marketing improve the performance;
• Is there a structural bias under- or overshoot?

Our visualisations in PowerBi will turn the demand data into actionable insights that enable the full planning team to focus and take high quality decisions. By analysing past performance, the dashboards provide advice that can be used in the current demand planning cycle.

We focus on the following insights:

• Impact: We show the planning team the impact of the enrichments performed in previous cycles. We indicate how many products have been enriched, the impact on forecasting performance and the type of adjustments made (negative, positive or no enrichment). By providing these insights, the planning team is able to get insight into their past actions and the effect it has on forecasting performance.
• Ineffectiveness: By focusing on the top products, customers or regions, a planner can find out which groups’ forecast accuracy decreased, how many products where changed and what kind of changes were made. By showing the ineffective enrichments, we are able to bring focus in the enrichment process. These products shouldn’t have been enriched as the enrichments had a negative effect on forecasting performance.
• Classification: By using ABC / XYZ analysis, we determine where to focus enrichments on and what products to leave to statistical forecasting. For each of the ABC/XYZ categories, we show the impact on forecast accuracy. By highlighting this, we help the planning team to focus on the products that are difficult to forecast and need human attention.

Whether it’s during the assessment or as part of our planning service, we make sure to discuss together with your planning team how to use the insights on cognitive enrichment. This way we take your team along in how to make effective enrichments, enabling you to overcome human biases such as optimism in the planning process.

Are you interested in learning more about forecast and demand planning assessment, the cognitive component or our Planning Services dashboards, feel free to reach out to us! 

Read part four of this blog series where we talk about how decisional guidance can help planners make effective enrichments.

The post Supply chain demand planning & forecasting assessment with a cognitive component appeared first on EyeOn.

In our first blog, we introduced our vision on smart-touch forecasting. A key element of this approach is what we call ‘cognitive forecast enrichments’, in which we nudge the planner to make better enrichment decisions.

To take the planner along in providing effective and efficient enrichments, we recommend the following four steps:

1. Create awareness
2. Assess previously performed enrichments
3. Activate planners in providing effective enrichments
4. Automate predictable enrichments where possible

This blog focuses on creating awareness.

Realizing the pitfalls of human forecasting behavior

The forecast enrichment process of many companies is still in the early stages. It is often unclear which person made the enrichments, or the forecast history including the different enrichments is not properly tracked. By analyzing the enrichment process and its different steps, we aim to uncover the pitfalls of human forecasting behavior. However, how do we analyze the enrichment process?

Performing a quantitative analysis is one of the first steps to evaluate your enrichment process. In such a quantitative analysis, we touch on the following topics:

By taking a detailed look at the (type of) enrichment process in place, the stakeholders involved, the tools that are being used, and the continuous improvement cycle, we can identify the pain points in the process that can be improved by means of for example coaching or process re-design.

Building enrichment capabilities

Next to the qualitative analysis, awareness can be created by means of a training session. We provide a one-day training ‘Smart-touch forecasting’ which is designed to introduce the key concepts and core requirements needed to design and implement robust forecasting enrichment processes. This will drive your business performance by balancing the use of advanced analytics with focused value-adding enrichments. The EyeOn master class ‘Smart touch forecasting’ can both be facilitated in-house and as a standard master class in which multiple companies participate simultaneously.

After creating awareness, it is important to also act and assess your enrichment performance. We will discuss in our next blog post how we propose to evaluate the added value of your enrichment process.

If you have questions or would like to get more information on how to create awareness in your organization, feel free to contact Bregje van der Staak!

Read part three of this blog series where we talk about forecasting and demand planning assessment with a cognitive component.

The post Create awareness on human forecasting behavior appeared first on EyeOn.

Due to increasing data availability, advanced data science techniques such as machine learning have become a powerful tool to reduce forecasting bias and create more accurate statistical forecasts. However, in practice, human planners still include additional information, not known to the statistical forecasting algorithm, to come to a final forecast.

Integrating planner enrichments to reduce forecasting bias

Despite efforts over the last decade to improve statistical forecasting capabilities and include more input data, this forecast enrichment process is still of eminent importance in most companies. But to really complement the statistical forecast the enrichments should be of high quality. However, research shows that planners do not always add value to the statistical forecast when enriching, and can even make the forecast worse. This inaccurate enriched forecast causes production re-planning that creates purchasing, financing, and scheduling difficulties, next to service level issues and imbalanced inventories. Only by enriching in a structured and focused way planners can truly add value.

At EyeOn we believe that we can integrate the planner enrichments in a smart way. By creating smart-touch planning solutions, we can automate the forecasting process for products that are easy to forecast, we can provide recommendations where it is more difficult, and flag where human intervention is needed. We call this process smart-touch forecasting.

As shown in the figure below, planners obtain a statistical forecast from the planning system and receive feedback on their enrichment behavior while interacting with the planning tool. This combined effort of planning system and planner results in the best forecasting performance.

The challenge of human bias

When evaluating the enrichments of human planners, we need to be aware of their cognitive biases. As humans are not capable of dealing with too large amounts of data, they can provide suggestions that are irrational.

Decreasing human forecasting bias with data-driven nudging

Therefore a key element of our smart-touch forecasting approach is what we call ‘cognitive forecast enrichments’, in which we nudge the planner to make better enrichment decisions.

In order to take the planner along in providing effective and efficient enrichments, we recommend the following four steps:

1. Create awareness
2. Assess previously performed enrichments
3. Activate planners in providing effective enrichments
4. Automate predictable enrichments where possible

The road towards smart-touch forecasting

The blog series ‘Improved Forecasting Using Cognitive Insights’ will dive into each of these four building blocks in the coming months. Each month, we will present a part of our approach in building an effective forecasting enrichment process. Read here the next blog on creating ‘Awareness’!

Are you ready to discover how smart-touch forecasting can build a solid foundation for your demand planning as well as bring more value to the role of your demand planners? The Fast Forecast Scan provides you with rapid insights into the main demand characteristics and forecastability of your business. It reveals the highest possible statistical forecast accuracy that can be reached and identifies the main opportunities for improvement. All in just a few days. Watch the on-demand demo of the Fast Scan.

If you have questions or would like to discuss your enrichment process, please contact Edward Versteijnen!

The post Reducing forecasting bias through smart-touch forecasting appeared first on EyeOn.

How to improve forecast accuracy: decide which products need human intervention

At EyeOn, we use the ABC/XYZ classification as guidance for planners to decide which products to focus on. Products are classified based on their volatility, from ‘stable demand’ to ‘high volatile demand’, and on their importance for the total turnover, for example ‘A products’ are important, whereas ‘C products’ are less important.

This ABC/XYZ classification results in nine quadrants, which are classified in terms of the human involvement needed.

• For those products that are important in terms of turnover and relatively easy to forecast (light green), a human planner can review the statistical forecast provided.
• For products that are not very important in terms of turnover, or are stable and medium important (dark green), we recommend using statistical forecasting only. Human involvement is not recommended for these products.
• Lastly, for products that are important for the turnover and hard to forecast (dark purple), we recommend performing manual forecasting. Human planners can focus on these products to increase forecasting accuracy. The same holds for the products that are new to the market (NPI) or in their end-of-life (EOL).

Enable planners to focus and make future adjustments smarter

By automating the forecast for most of your products, you enable the planners to focus on key products where their expertise is most needed and has the highest impact. These focused adjustments performed by the planners can be used to see where planners add value and can be analyzed and used to make future adjustments even smarter! Which, in the end, will give a significant boost to your forecast accuracy.

The road towards smart-touch forecasting

Do you want to learn more about this topic? Don’t miss our blog series Towards smart-touch forecasting, where we dive into the four building blocks to pave your road toward advanced smart-touch forecasting using cognitive insights.

Are you ready to discover how smart-touch forecasting can build a solid foundation for your demand planning as well as bring more value to the role of your demand planners? The Fast Forecast Scan provides you with rapid insights into the main demand characteristics and forecastability of your business. It reveals the highest possible statistical forecast accuracy that can be reached and identifies the main opportunities for improvement. All in just a few days. Watch the on-demand demo of the Fast Scan.

Need personalized advice? Reach out to our expert Bregje van der Staak!

The post How to improve forecast accuracy with smart-touch forecasting appeared first on EyeOn.