# The Statistical Value Chain

When decisions are made in both private and public organisations, the decision-makers are supported by analysts who provide key decision support. As such, the quality of decision support provided by the analyst directly affects the quality of critical decisions. At Q2M2, our approach is built on a simple method that decision-makers can employ to evaluate the process of decision support from a statistical perspective. We call this approach the “Statistical Value Chain” (SVC): a consecutive benchmarking checklist with eight steps that can be used to evaluate decision support with statistical precision.

Click here to download a pdf of the Statistical Value Chain.

The Statistical Value Chain (SVC) handles data in a cradle-to-grave perspective, from the extraction of raw data to its use for decision support. The better the data is handled at each step of the Statistical Value Chain, the better the resulting decision support—and therefore the better the final decisions. We are proud to present the Statistical Value Chain to you.

### Step 1: Define

The Statistical Value Chain will always start by **defining in time and space** the boundaries of the investigation. In defining these boundaries we also define the population that we want to investigate, the group of things or items that we want to know more about—for example, a specific cornfield in the present year, all soybean fields in a given country in the year 2006, or a batch of printed circuit boards in the year 2018. In most decision support contexts there are many populations from which to collect data, and we refer to this as a *system*. To make valid investigations, we need to define the time and space in a clear manner. If this is not done, nothing else can fall into place.

To make valid investigations we need to define the time and space in a clear manner. If this is not done, you should ask – what then?

### Step 2: Sample

Step two of the Statistical Value Chain is to make a **representative** sample of the system in order to** avoid any bias** that could lead to flawed decision support and incorrect decisions. For example, when investigating a population’s transport habits it is not representative to ask/investigate only what people in the capital use for transportation.

If step three of the Statistical Value Chain is not done in a representative way, you should ask – what then?

### Step 3: Describe

Step three of the Statistical Value Chain is to **calculate and quantify** the factors defined and measured in steps one and two—for example the minimum, average, and maximum income for a specific group of people in a given year. This step of the Statistical Value Chain might not be the most difficult step, but mistakes here are critical. Assuming that the data you receive from third parties is always correct and without error can lead to flawed decision support and incorrect decisions.

You should ask – how and by whom was this step of the Statistical Value Chain undertaken?

### Step 4: Investigate

Step four of the Statistical Value Chain is to **investigate** by summarising the different populations and find relationships in the system. For example, when investigating the total amount of greenhouse gases from diesel engine cars in Europe in one year, we must answer a number of questions. How many diesel cars are there in Europe? How far do those cars drive in a year? How much greenhouse gas does each diesel car emit per km? At the end, we can summarise these numbers and say that in the year 2010 the total greenhouse gas emission in Europe from diesel engine cars was, for example, 100 million tonnes.

You should ask – how and by whom was this step of the Statistical Value Chain undertaken?

### Step 5: Forecast

Step five of the Statistical Value Chain is to apply the data that has been defined, collected, described, and summarised, and to use this data to make a **data-model for forecasting**. It could, for example, be desired to forecast, predict, or estimate the weather for next week, or the production of milk in a country in the year 2025. The better step five of the Statistical Value Chain and the previous steps have been performed, the better the forecast will be.

You should ask who did the forecast, how was the forecast done, and what data went into the data-model used for the forecast?

### Step 6: Baseline

Step six of the Statistical Value Chain is to **estimate a baseline.** What will likely happen if no changes are introduced? For example, a company could consider continuing—without any changes—a production of diesel engine cars. What does the company expect their future revenue baseline to look like? What about the future baseline for greenhouse gases emitted from the diesel cars produced?

You should ask who developed the baseline? How was the baseline in the Statistical Value Chain developed? Were all forces that can affect the baseline included in the baseline study? What data was used to develop the baseline?

### Step 7: Alternatives

Step seven of the Statistical Value Chain is to **define and create alternatives **to the estimated baseline. For example, if the government is considering changes to a tax, then this will be an alternative to the existing tax structure. How will this change affect taxpayers and public funding compared with the future baseline?

You should ask who developed the alternatives? How were the alternatives in the Statistical Value Chain developed? Were all forces that can affect the alternatives included? What data was used to develop the alternatives?

### Step 8: Valuate

Step eight of the Statistical Value Chain is to **place a value** on each of the previous seven steps. For example, putting a monetary value on the total emission of greenhouse gases from diesel engine cars in Europe.

You should ask how the valuation was performed? Who did the valuation? What data went into this valuation and how was this data obtained?

It is our experience and conviction that our methods can create value in the vast majority of industries and contexts.

The Statistical Value Chain is published in 2013:

Herrmann, I. T., G. Henningsen, C. D. Wood, J. I. R. Blake, J. B. Mortensen, and H. Spliid. 2013. The Statistical Value Chain (SVC)—A Benchmarking Checklist for Decision Makers to Evaluate Decision Support seen from a Statistical Point-of-View. International Journal of Decision Sciences 4(2). July–December 2013.