Anyone who works with IBM Planning Analytics (TM1) will quickly come across two terms: rules and feeders. They are at the heart of TM1 and determine whether a model functions correctly or whether users are left frustrated by empty cells.
In this article, we explain step by step what these concepts are, how they relate to one another, and why they are essential for any TM1 application.
Note: For historical reasons, the term “TM1” is still often used interchangeably with “IBM Planning Analytics,” but today it officially refers only to the database engine.
Why do we need rules?
TM1 cubes store not only data but also calculated metrics. A classic example: Revenue is calculated as quantity multiplied by price.
In Excel, we would have to enter this formula into every single cell—a process prone to errors and time-consuming. TM1 works smarter: A rule defines this logic once, and it automatically applies to all relevant cells in the cube. This ensures consistent calculations, saves time, and simplifies maintenance.
How Rules Work
Rules are calculation instructions that TM1 executes when a cell is accessed. Here is a real-world example:
- Revenue = Quantity × Price
- Profit = Revenue – Costs
- Profit Margin = Profit ÷ Revenue
The key point is that rules do not store values. Instead of storing values, a rule defines the calculation logic centrally within the model. TM1 applies this logic whenever a value is needed.
Consolidations vs. Rules
TM1 comes with a powerful built-in feature: totals and consolidations are generated automatically as soon as dimensions are structured hierarchically. For example, a sales hierarchy automatically totals the revenue for all products without requiring additional rules.
Rules are only needed when dealing with non-standard calculations. Profit margin is a typical example—it isn’t the result of a simple addition but must be recalculated for each level.
Feeders – the unsung heroes
This highlights both the strength and the complexity of TM1. The system stores only calculated values to save memory. This means that if a rule references data that does not exist in many cells, TM1 sometimes does not realize that it needs to perform a calculation. As a result, the cell remains empty.
Example:
- Rule: Revenue = Quantity × Price
- Input: The price is available, but the quantity is not (yet).
- Without a feeder: TM1 does not display any revenue, even though a value should be calculated.
Solution: Feeders. They act like little signposts that tell TM1: “If there’s a value here, you need to perform a calculation there.” In the revenue example, this means that as soon as quantity or price is entered, TM1 knows that revenue should also be calculated.
Common pitfalls
Beginners, in particular, often run into common problems:
- Rule overrides consolidation: If a rule is formulated too broadly, it can unnecessarily override automatic totals.
- Too many feeders: Setting too many feeders makes your cube large and slow.
- Empty cells despite the rule: In most cases, this is simply because the correct feeders are missing.
- Unclear documentation: Without comments, no one will understand later why a rule exists.
Best Practices for Rules & Feeders
- Define as few rules as possible, but as many as necessary.
- Use feeders strategically and test their effectiveness—it’s better to be precise than to apply them across the board.
- Document rules in a separate file or with clear comments.
Case Study: Sales Planning
Let’s take a cube with the dimensions Product, Region, and Time. We want to calculate revenue there:
- Input: Users enter quantity and price.
- Rule: Revenue is automatically calculated as quantity multiplied by price.
- Feeder: As soon as either quantity or price is available, TM1 “knows” that revenue must also be calculated.
- Consolidation: Regions or product groups are automatically summed up.
Why a rule makes sense here:
Revenue is regularly analyzed as part of the planning and analysis process. Since rules are calculated every time they are accessed, this approach is ideal for metrics that are evaluated daily or on an ongoing basis. If, on the other hand, revenue were needed only rarely – such as once a year for financial closing – a process (e.g., using TurboIntegrator) would be the more efficient alternative.
Conclusion
Rules and feeders aren’t rocket science; they’re the core control mechanisms of TM1. Rules define how metrics are calculated, while feeders ensure that these calculations take place exactly where they’re needed. It’s the interaction between these two mechanisms that ensures models function correctly, efficiently, and transparently.
Those who understand when it makes sense to implement a calculation using a rule and when a process is the better choice can avoid common mistakes, reduce performance issues, and establish a solid foundation for planning and analysis.
Contact us to learn more about how IBM Planning Analytics can help your business!
