When working with clients to design a TM1 system, I sometimes have clients ask about an approach of building one cube for a model vs. building multiple cubes for a model. What do I mean by that? A cube consists of a single table/database of data whereas a model consists of a series of cubes that are linked together.
TM1 is designed to quickly and efficiently link a series of small cubes together, but this style of thinking contradicts the approach that you may be used to within Excel. An Excel spreadsheet is a single table of data that offers the ability to link together more data via multiple tabs and multiple workbooks. This approach easily gives you the ability to start with a single set of data and then expand it.
What happens though, when you take this approach, is that you end up with your TM1 cube looking like your attic might if you just toss all your stuff in it. You might have old skis, camping gear, baby keepsakes, and more all together in one room. It makes it hard to find what you need and, in some cases, hard to know what you even have up there.
When you build multiple cubes, it’s like having an organized system to store all your stuff. You might have a closet dedicated to your sporting equipment. Camping gear may live elsewhere in a storage unit. Your family heirlooms and keepsakes are kept in water-proof containers in the garage. It’s much easier for you to find what you are looking for.
Revelwood’s approach to building TM1 models is to build small, efficient cubes that each have a separate role and purpose. These cubes link together to make a model. For example, say we are going to build a TM1 model to calculate compensation costs for a mid-sized national technology business. That model has three main components:
- Data that gets entered by the planning team
- Lookup tables for various rates and dates
- Calculated results
Each cube stores only the data that is needed for its specific function and task. For example, there may be one rate cube for Federal tax rates and another cube for state tax rates. Everyone in the company has the same Federal tax rate for a single year whereas each state has a different rate. If you were to have one large cube for all tax rates – Federal and individual states – it would be hard to navigate and may require workarounds to make the data fit (for example, creating a “federal” state).
Building smaller cubes versus one large cube has benefits beyond those that are just about organizing your data. In fact, the organization part is really just a side benefit. When you have a model that consists of smaller cubes, you have a system that is not wasting space and provides you with a flexible foundation for your TM1 applications. Because your cubes are smaller, the entire system also performs faster.
Chances are you bought TM1 because you weren’t happy with Excel. Many Excel “systems” grow into multiple workbooks with many linked spreadsheets, which turns into a messy behemoth. Why would you want your TM1 environment to look like that?