Mastering Anaplan PlanIQ: Guide, Tips, and Tricks

Just as Tom Brady meticulously analyzes game footage to anticipate the opposing defense or LeBron James adjusts his strategy mid-game based on his opponents' movements – shouldn’t business have the flexibility to change their plans based on the needs? Shouldn’t they have the visibility into their financial, supply chain, sales, and operations? That’s what Anaplan gives.

The power to collaborate, integrate, budget, and plan for a variety of functions including: Demand planning, Supply chain planning, HR and workforce optimization, and Finance. To this hat, adding a intelligent forecasting feather make it more seamless and relevant. That’s what PlanIQ in Anaplan is for. To combine Anaplan’s connected planning with accurate forecasts.

What is Anaplan PlanIQ?

PlanIQTM is an AI/ML driven statistical forecasting technology with Amazon’s best of breed intelligent forecast engine, as well as Anaplan- native algorithms such as multi-variate linear regression (MVLR) and anaplan prophet to create accurate forecasts driven by internal and external drivers integrated within Anaplan.

That’s the definition anyway. In simple terms, it trains forecast AI and ML models with the Anaplan module data or data directly from warehouse (AWS S3, Azure, and GCP) to equip users with quality metrics for true results. The added benefit is that – though it sounds complicated, all you have to do is setup your interface (that is easy-to-use) and integrate it directly into your anaplan planning modules.

PlanIQ Benefits & Features

With our expertise as both Anaplan Implementation experts and Data Science Experts, we understand the need for having accurate (demand/financial) forecasts & to start a new process. Given the innumerable tried and tested launches of PlanIQ we’ve done, the benefits far outweigh the risks.

1. Improved forecast accuracy by up to 50% with powerful forecast engines (Anaplan’s own algorithms such as multi-variate linear regression (MVLR) and Anaplan Prophet, Amazon Forest, etc.)

2. Manually or automatically learn and train forecast models to increase accuracy

3. Eliminate the need to train data every time by leveraging the recent data

4. Schedule operations and manage forecast quotas to match your forecasting cycle

5. Create what-if scenarios for a variety of use cases without the need for data science expertise (directly by the model builders)

6. Support ad-hoc predictions

7. Attain highly accurate metrics in an accessible & easy to learn format

8. No black-box models & truly understand what drives forecast results

9. Reduce time needed for data processing

The true secret and fun with this forecasting solution is that – it doesn’t require any special expertise to configure, deploy, and operate – making it easy to setup and visualize the results.

Anaplan PlanIQ Use Cases

From demand forecasting to workforce planning, the scope of PlanIQ expands across the functions of HR, Finance, Supply chain, and Sales. We’ll not go into the definitions of each of the use cases but talk a little about what’s usually covered in the top 3 use cases: Financial planning, Demand Planning, and Workforce planning.

Financial planning

You can pull in data from Anaplan models and automatically chooses the best ML prediction model for the data (we’ll discuss the 8 Algorithms of PlanIQ below), saving hours of manual effort. Meanwhile, ad-hoc analysis allows analysts to determine the drivers that increase prediction accuracy.

Demand Planning

With PlanIQ’s demand planning you can assess, and forecast based on the baseline sales, identify total demand, cleanse history, drill down on demand drivers’, include trend and seasonality analysis, and detect and remove outliers

Workforce planning

Predict staffing needs and optimize resource allocation. Integrate headcount forecasts directly to revenue plans, operational expenditures (OpEx) forecasts, and other strategic plans. Plan for staff even during the peak holiday season (which usually yields inaccurate results), integrate multiple variables, and automate processes for staffing requirements.

From portfolio cannibalization analysis, off the shelf forecast, promotional effectiveness, to inventory level projections there are a lot many other use cases which we’re not talking about now.

Though the words AI and ML are used to describe PlanIQ, it is surprisingly simple to use it as it comes it with the interface to select the models. It is possible to complete it within 4 simple steps.

4 Steps to getting started with Anaplan PlanIQ

To get started with PlanIQ, you’ll need the Historical data (for which you want the forecast), related data that might influence the historical data like pack sizes, color, or any data that can have a casual inference, information of the attributes which is static, non-time dependent categorical text features that describe the items in the historical time series data like style, category, geographic location, etc., holiday calendars (which is also related data – think about the Thanksgiving bump in data), forecast result cubes, and the import-export actions. Now let’s jump right into how to setup PlanIQ.

1. Setup your data collection

PlanIQ can ingest up to three types of data to deliver a predictor. While only one type of data is required to forecast, adding two other sets can help the forecast learn and output more tailored forecasts to a specific scenario – which is why we add the related data and attributes.

2. Run forecast model

Once you’ve reviewed the data, you can create a forecast model. Please note that for accurate forecasts clean data and relevant inputs are needed. There are three things to be considered here: The Algorithms, Forecast horizon, and the Calendar (to cater to geographic specific forecasts).

• Provide a name for the model
• Choose the required algorithm from the drop down list (we’ll get into the details of the algorithm below)
• Choose the required frequency of forecast (depends on the granularity of the data in the data collection)
• Then click on the Create forecast mode – which might take 30-90 mins to render based on the model selected

While choosing the models it is important to pay attention to the optimization metric setting. PlanIQ supports three optimization metrics: MASE (default), MAPE and RMSE.

Once your review of the forecast is complete i.e. checking the quality metrics, you can import the backtest data (the data that is held back from the historical data which is equivalent to the forecast horizon) to check the accuracy. After this you can import the explainability (which is a part of some of the algorithms like Amazon Ensemble) – which comes in a graphical format with a comparison between the variables.

3. Create a forecast action plan

By now you should have a pre-trained forecast model in ready state with the forecast results and import models. Now you can get PlanIQ to take action.

• Select Forecast actions > New Forecast actions > Create a forecast action
• Enter the forecast action name that was created while you were setting up the training plan
• For the Forecast import- select your target workspace and model and then the explainability import

4. View results and embed them in planning

• Create cards with PlanIQ results to view explainability side by side with historical data and forecasts – by configuring the right cards.
• You can now review the forecast results and add them to the planning process to improve decision making.

When should you get PlanIQ?

Though we’ve described how to get started with it, there are few things you’ll have to take into consideration while thinking about PlanIQ. For starters, if you don’t have the adequate quality data – you should think about improving that first before getting started with Anaplan PlanIQ.

Top three parameters you should consider are:

1. Need for a forecasting tool/platform – this should be based how you current methods are – whether they’re inadequate or manual or if they’re too expensive etc.

2. Scaling ease- When dealing with multiple product lines, diverse geographical markets, or intricate supply chains, PlanIQ's advanced algorithms can help manage this complexity – in Anaplan environment.

3. Resource constraints - PlanIQ can provide advanced forecasting capabilities without the overhead of maintaining an additional data science team.

Unless you already have a very high-level forecasting tool, we’d suggest you to atleast take up a POC for PlanIQ to look at how the environment would play for you.

Getting the most of your PlanIQ investment: Mistakes, Best Practices, and Skills

One of the common mistakes that some of the customers made while implementing PlanIQ is – to not consider the skills needed. Given that it is based on Anaplan, the Model builders need to be able to create saved views, structure models, create and save imports and exports. But they would also need to have a some understanding about forecasting.

Some of the PlanIQ forecasts would look something like these:

So, the skill level to be able to interpret these and understand the parameters that we described above like the forecast predictions and the attributes, etc. would be needed.

The second consideration is about Change Management – especially when you’re moving from manual forecasts or Y-o-Y growth forecasts. This would need a little heavier change management vs when you’re already using some sort of statistical forecasting or custom ML solutions. For any of them the requisite change management is needed.

Best practices while using PlanIQ

At the end of the day, PlanIQ is an investment. Even Harry potter needed the sorting hat to pull out the sword of Gryffindor – every right use cases needs the right tools and processes for the intended output.

To get the best out of your investment, you need to keep in mind these best practices:

1. Ensuring data quality and preparation i.e. that the data is clean, consistent, and structured appropriately. Also identifying the external data for incorporating for better forecasts.

2. Set up a regular schedule for forecasts (e.g., daily, weekly, monthly) based on your needs to ensure timely adjustments.

3. Monitor regularly & compare forecast results with actual outcomes to evaluate model performance.

4. Provide training for effective use data input, model selection, and interpretation of results.

5. Integrate with existing processes, use tools like Polestar’s Anaplan Integration utility or connectors to be connected to data sources and results.

6. Use PlanIQ as a tool to augment human decision-making, not replace it. Take help from business users alongside PlanIQ's forecasts to verify results.

Additionally, we’re talking about a few more topics a lot of our clients have asked us before.

Commonly asked questions:

1. How can we bring data into Anaplan PlanIQ?

Forecast teams can drive forecast accuracy without having to first move the data into Anaplan to bringing internal and external data directly into PlanIQ, from AWS S3 or Azure Blob or Google Cloud. They can also connect to Anaplan Models for brining in the data. Alternatively, they can use the help of connectors like Polestar’s Integration Utility or Anaplan Cloud Works to connect to data sources.

2. What are the advanced metrics that are used in PlanIQ?

The quality of the forecasts (like any other statistical method) is understood by a few set of advanced metrics like:

• MAPE or Mean Absolute Percentage Error: The sum of the individual absolute forecast errors, divided by the actual values for each period. The closer this is to 0 the better.


• RMSE or Root Mean Square Error: This is the root of the cumulate sums of square of the difference in actual vs estimated value divided by number of non-missing data points. Lower RMSE value indicates more accurate prediction.


• MAE or Mean Absolute Error: This is the average absolute difference between X and Y – only compares data for similar units.


• MAAPE or Mean Arctangent Absolute Percentage Error: This is a measure of forecast accuracy of zero or close-to-zero actual values. This is particularly useful when extremely large errors can occur because of mistaken or incorrect observations.


• MASE or Mean Absolute Scaled Error: This is the mean absolute error of the forecast values, divided by the mean absolute error of the naive forecast. his metric is well suited for intermittent-demand series.


• SMAPE or Symmetric mean absolute percentage error: This scale dependent relative error is the absolute error divided by the magnitude of the exact value.


• MASE: This is the overall model quality metric which penalizes over positive and negative forecasts equally. A model can be considered a good metric if at least 60% of the items received a high quality MASE result.

3. What are the Algorithms that are supported by PlanIQ?

This topic in itself can have a whole article for itself. For the sake of simplicity, the algorithms supported are:

• Baseline time series algorithms including ARIMA and ETS
• Advanced statistical time series algorithms like Multi-variate linear regression
• Flexible local algorithms like Anaplan Prophet
• Neural network algorithms like DeepAR+ and CNN-QR

Please note that no algorithm is better than the other, the performance depends on the specific use case, datasets, context and historical patterns.

4. What are the types of outliers and how to deal with them in PlanIQ?

In general, there are three types of outliers namely: Point outliers, Contextual outliers, or collective outliers. Point outliers and contextual outliers are more common forms of outliers. Other than visualizing them, they are usually identified by calculating the standard deviation of the values.

Some of the ways you can deal with Outliers are:

• Leverage the Exclude Value functionality of PlanIQ, which will replace the outlier a value based on an automated logic
• Create a forecast model and predict the value for the period in which the outlier is present, then replace the outlier with the forecasted value
• Replace the outlier manually

In conclusion: What’s your PlanIQ need?

PlanIQ is a forecasting tool available for Anaplan users and is beneficial for those who want to add the power of AI/ML to their planning needs. Before you get started with PlanIQ think about the why, the how, and the when – in case you need help – talk to our Anaplan experts and we can guide you through the journey!