IBM Planning Analytics

In the market for a new planning application? Make sure you do the math first!

Blog Home > In the market for a new planning application? Make sure you do the math first!

In the market for a new planning application? Make sure you do the math first!


Reading Time: 3 minutes

A new planning application can help take your organization’s planning process to a whole new level by reducing your dependence on spreadsheets, automating the process and reducing errors.

One important consideration to keep in mind, however, when evaluating planning vendors, is whether the application simply has the capacity to hold all of your company’s critical business data.

This may sound like an odd question to ask. But there are cloud solutions out there that have severe limitations on data volume. You don’t want to find yourself in a position where your new planning solution can’t handle your data requirements, with the result that you can’t fully realize the benefits you’d hoped for at the start of your evaluation. These restrictions can hinder analysis, limit access to detail and require significant data movement, maintenance and reconciliation work that you didn’t originally plan for.

Understanding “sparsity”

To understand why data volume is such an important issue, you first need to understand the concept of “sparsity.” If you consider how your company’s data appears in a table or cube, you will know that there are a lot of zeros!  Not every combination of Account, Cost Center and Month, for example, will contain a non-zero value. In fact, for most organizations, there will be far more zero cells than non-zero cells.

The reason why some cloud solutions have data volume issues is that they do not handle sparsity well. They allocate space in the database for all cells (or combinations) regardless of whether they have a zero value or not. So, in the above example, they will allocate storage space for every single account for every single cost center, for every single month. That’s a lot of space!

Do the math!

Some solutions begin to have storage space issues at about 2 billion cells and reach an absolute limit (at which the system will not function) at about 5-8 billion cells. Beyond 2 billion cells, the administrator starts having to make design compromises to get all of the data to fit in. Examples include:

  1. Combining two or more dimensions into a single dimension. So, for example, you would need to concatenate the product and cost center codes together to form a new single dimension.
  2. Split the cube up into a number of smaller cubes. For example, you may need to put the West region’s data in one cube and the East region’s data into another cube. Then use a third summary cube to show the combined data in summary.

In these situations, you’ve had to sacrifice the granularity of your analysis just in order to get the tool to work properly. The good news is that it’s easy for you to determine whether your company’s data will fit into the cloud solution database. Simply add up the size of each of your critical business dimensions and multiply them together.

Example*

  • Cost Center              1,000
  • Account                    100
  • Location                    200
  • Time                           12
  • Year                            03
  • Version                      10

Total cells = 1,000 * 100 * 200 * 12 * 03 * 10 = 7.2bn

In this fairly typical example, a user would already be very close to the absolute limit of the solution mentioned above. Thus, design compromises would be necessary almost from the start. Sub-optimal, to say the least.

To summarize, if you are evaluating planning solutions currently – do yourself a favor – do the math! These few simple steps will help you make the right planning too selection!

And for additional factors to consider when choosing a planning solution, take a few minutes to review this Planning, budgeting and forecasting: Software selection guide.

 

*In this calculation, I assumed a monthly planning horizon. If your organization plans by week (52) or day (365), then adjust accordingly. For versions, I assumed;

  • Last Year’s Actuals
  • This Year’s Actuals
  • Budget
  • Forecast
  • Target
  • Quarterly updates (4)
  • Best Case

 

Leave a Comment
0 Comments

Leave a Reply

Your email address will not be published.Required fields are marked *

More IBM Planning Analytics Stories

Data Science

Timothy Walker

Recap: IBM Data Science and AI at Think 2019

Customers, business partners, and IBMers alike walked out of Think 2019 full of ideas and excitement for the vision driving IBM Data Science and AI.

IBM Planning Analytics

Timothy Walker

Melton Renewable energizes financial planning and analysis with IBM Planning Analytics

Melton Renewable Energy UK used IBM Planning Analytics to improve its financial performance management and accelerate budgeting and forecasting cycles.

Event

Paul Bard

See you at Think 2019!

Think 2019 is next week in San Francisco. And the learning and networking starts even before the conference formally begins.