Create a Key Figure Matrix and Architecture Overview

During the requirement analysis, consult the following sources of information:

• Project focus

  Check the scope and aim of the project with the sponsor.

• Interviews

  Identify the participants for the interviews (management) and determine the conditions (business strategy, products, markets, customers, targets, project scope) as a basis for the operative workshops (with the employees). It's important to define the correct coverage for personnel interviews.

• Workshops

  Hold a workshop with the employees of the departments and determine proposals for key figures (KPIs) and characteristics.

• Documentation

  Document the collected results (characteristics, key figures, calculations, descriptions, and so on) and determine which areas already have KPIs.

• Gap analysis

  Determine the differences between the previous reporting systems.

• Reporting requirements

  With the departments (through the workshops), determine the reporting requirements using the collected characteristics and key figures (documentation).

First, you must structure the information from the requirement analysis. You may need to structure the analyzed characteristics and key figures in a table display, and define the initial relationships between the characteristics and key figures. We provide two examples of table displays. Let's start with a simple key figure matrix.

The preceding figure displays an example of the key figure matrix that contains only information needed for the next steps. It demonstrates a formal way of structuring the information requirements that you've gathered. The first step is to determine which key figures must be measured. For example, quantities, volumes, profit, and costs.

The steps involved in creating a table display (key figure matrix) from a requirement analysis are as follows:

1. Gather the key figures as rows.
   1. Obtain relevant KPIs from interviews.
   2. Determine basic key figures.
2. For each key figure, incorporate the most detailed time granularity.
3. Gather characteristics as columns.
   1. Obtain the characteristics that describe the KPIs from interviews (list the basic business subjects and their attributes without distinction yet).
   2. For each characteristic, generate a new column.

4. Assign business subjects to KPIs.

   For each key figure, add a mark (X) for all characteristics that describe this key figure.

5. Group key figures by structure. The idea is that key figures with the same set of characteristics can be in the same InfoProvider.

Different interview partners may provide different answers. Keep in mind that people might use different terms for the same things, such as "product ID", and "product number".

Ask the interview partners to agree on one definition. If they can't, agree on an SAP BW/4HANA terminology that distinguishes different entities as clearly as possible.

If there are in fact, different requirements, adapt the entries in the key figure matrix in the following ways:

1. If the new interview partners mention further KPIs with other required key figures, add them to the list as new rows.
2. If the new interview partners desire a coarser time reference, leave the finer one from the earlier partners. If they require more detailed time characteristics, change the time characteristic to the finer one.
3. If the new interview partners mention further characteristics, add them to the list as new columns.
4. If the new interview partners desire further assignments, add new marks (X), but don't delete old marks.
5. Regroup the key figures.

The preceding figure displays examples of enhancements of the key figure matrix.

You can also add any kind of information to the key figure matrix that makes the subsequent steps easier.

Add the unit of measure (or currency), whether it's a fixed one ("KG" for kilogram, or "PC" for pieces) or different ones ("original unit"), or none ("1" as unit).

If you expect or know SAP BI content that covers the need, you can indicate the content object, or at least the application component.

If you know which source system, or even which table is involved, indicate it in another column. This indication helps you to recognize where harmonization issues are relevant.

If it’s clear that a key figure can’t be evaluated for a specific level of detail (characteristic), indicate this with a hyphen (-). If it turns out that this detail was requested, report the problems to the requester immediately.

If a more detailed inquiry reveals that a characteristic is used with different reference objects, such as the country/region of the customer and country of the sales organization, add more columns.

A more detailed inquiry may reveal that a key figure is used in different versions. For example, sales volume in original currency and sales volume in Euro, or net and gross values. If so, to generate two rows with the same assignments, copy the row.

Add formulas for the calculation, such as the following:

• "Currency conversion"

• "Unit conversion"

• "1:1" for no transformation

• "RKF" for restricted key figure, with a specific condition such as "cost type = ‘storage’"

• "CKF" for calculated key figure with a formula such as "items * price", or "revenue–costs"

The KPIs from the requirement analysis are broken down to the lowest level of the basic key figures. When you do this, you must also consider whether these key figures are basic key figures or calculated key figures. Add new lines for the basic key figures.

If a key figure is calculated from others, then as much detail as for the calculated key figure must be assigned to the basic key figures. For example, if profit is calculated from revenue and costs, with detailed values per product, then costs and revenue must also be provided per product. This is indicated by the red text and arrows in the figure.

The process of deriving an architecture overview model from a key figure matrix, involves the following steps:

1. Definition of data marts and their key figures.
2. Identification of sources.
3. Identification of transformations.
4. Addition of more layers according to LSA++.
5. Decision about physical storage requirements (to store snapshot data, or to manage duplicates).
6. Identification and reduction of redundancies.

When defining data marts and their key figures, start the key figure matrix with key figures grouped by the set of characteristics relevant for them. For each group, define one data mart object. For each data mart object, list the key figures.

For each pair of data mart and source, indicate how the key figures must be derived. If extra values are needed in a transformation, add more sources and targets.

On top of each data mart, add a CompositeProvider as the interface for reporting. Add a propagator layer, and maybe extra layers for data acquisition.

If values can be calculated with a simple formula, calculate it in a Query (or SAP HANA calculation view) instead of storing results in an additional data mart. Remove the data mart from the architecture model. For the propagator layer, if changes for existing keys are allowed, use a Standard DataStore Object with change log. If there are no changes in the original values, use one with compression capabilities, but without change log. If every key is provided once, use an InfoSource instead of a propagator DataStore Object (advanced). For the data mart, decide if you have a specified key, or if all characteristics are key.