HyperModel: five tips to improve your experience
Are you considering HyperModel for your ecosystem, or have you recently deployed it? Fantastic! We've compiled a shortlist of the top five tips to help make kicking off your HyperModel experience as smooth as possible.
Check out these top five tips to improve your experience:
1. Recognize that the Hyperblock has not changed
- With HyperModels, only model/workspace sizes and the available memory are increased.
- The Hyperblock engine remains the same, meaning:
- Calculation performance is the same. HyperModel performance is dependent on modeling best practices and how the model is constructed.
- Hyperblock dictates the maximum block cell count.
- The following functions maintain the same limit set of 50 million cells, which may be easier to reach with HyperModels:
2. Embrace sparsity
- Sparse modules are not inefficient when it comes to calculations. The Anaplan Hyperblock engine is designed to work with multi-dimensional structures. At its heart, the Directed Acyclic Graph (D.A.G.) indexes data to calculate only what is needed when upstream data points have changed.
- Model size can often be reduced with increased multi-dimensionality by reducing list sizes. For example, this could apply to transactional lists that include dates. For more information, see Data Hubs: Purpose and Peak Performance.
Read The Truth About Sparsity: Part 1 for additional information about sparsity.
3. Understand the effects of text
Text can quickly consume memory and potentially slow calculations. Scaling any text formatted line items or text-based calculations in a HyperModel could cause memory-related issues.
- Avoid importing data as text where possible to maintain memory and calculation speed. Instead, import as List items.
- Create new lists to allow for more list-formatted items in data modules. We recommend this method rather than importing lists as text and conducting a FINDITEM search.
For more information, see Data Hubs: Purpose and Peak Performance - Anaplan Community.
4. Understand top-level summary
- Try to reduce using summaries and top-level summaries to maintain optimal HyperModel performance. The increased scale with HyperModels means there's more data to aggregate, which may result in slower summary calculations.
- If a summary is only needed on one or two dimensions in a multi-dimensional line item, we recommend turning off the summary and creating a new line item for that specific summary.
5. Test and optimize
- Do it yourself: optimize models with best practices by leveraging the Planual, PLANS, and D.I.S.C.O.
- Leverage our packaged services and reach out to your account team for more information:
- Model optimization
- Model concurrency testing
Are you ready to scale up on HyperModel? Or, are you already leveraging it?
Let us know your thoughts in the comments below.