Say Hello to Confident Planning with SAP Analytics Cloud
This is a blog post following a webinar presented by Antoine Chabert – “Introduction to Predictive Planning in SAP Analytics Cloud”
This blog will provide a brief overview of Predictive Planning and show you how SAP Analytics Cloud allows you to leverage this feature to enhance your planning and decision-making capabilities. For a more detailed presentation, please watch this webinar.
In advanced analytics we see features that support Planning, Reporting, Dashboards, Data Discovery, Visualisations, and Smart Wrangling capabilities. In the future of analytics, we see the implementation of Augmented Intelligence which includes Natural Language Processing, Artificial Intelligence, and Predictive to name a few. Augmented Intelligence compliments our human efforts in deriving key analytical figures by enhancing our decision-making abilities. This blog post focuses on planning in the traditional sense and predictive in terms of Augmented Intelligence.
What is predictive planning?
Planning means to apply reporting derived from human intelligence in different departments across your organization, such as finance and sales, and use this to develop a vision for your business in the coming months or years. As business users, we want to understand what could happen in the future based on actual historical data – a process known as forecasting. In addition to this, business users want to understand how they can achieve their plan, specifically, what is the right course of action – this is where Predictive Planning comes in. Predictive Planning crunches the actuals (historical data) and projects (or predicts) the likely outcomes of our KPI’s. Predictive Planning does not replace human expertise. In fact, it compliments and accelerates the planning process, and helps planners make confident decisions.
The success formula of predictive planning:
The simple formula for success in predictive planning is combining your historical data (the evolution of your data over time) with machine learning and business acumen. Machine learning creates time-series models which provides a smart baseline to explain the evolution of your data. Planners build on top of these insights with their business acumen by supplementing their understanding of economic situations and business operations that predictive technology isn’t aware of.
Demonstration of a Forecast Prediction:
In this revenue forecasting scenario, a company is selling Bikes and E-Bikes across different regions. We have been given the actual data for Q1 (2020) and the Budget 2020. The Predictive Forecast has been created by the planner themself based on the historical data from Q1 and the budget. We also have the Smart Prediction, where we see the prediction up till the end of Q4. Planners can use these numbers and compare the product across the different regions based on the budget, Predictive Forecast, and Smart Prediction. Given this information, business users and planners can now decide whether current operations are going to meet the budget across the different products or not. Therefore, a predictive forecast table created in SAP Analytics Cloud like the one above provides past visibility and future predictive ability, resulting in confident planning.
To see a detailed demonstration, please go to 10:40 of the webinar.
Demonstration of a Predictive Scenario:
Let’s take a step back. Before creating your forecast predictions, you should create a predictive scenario, where you can build a planning time-series model. To do this, pick your data source, data version, and the KPI you want to predict. For a more thorough demonstration, please go to 14:00 of the webinar.
Something unique and exciting to Predictive Planning is entities. You can use entities to create multiple time-series charts. For example, if you want to predict sales revenue over multiple products and regions and want them to be crossed together, you can ask Predictive Planning to create as many time-series charts as there are crossings of types of bikes and regions. Each of these entities is very specific to the type of bike and region. The image above shows you the error you can expect when applying that entity to predict a future outcome. For example, E-bike and France has an error rate of around 2% – meaning this is a highly accurate model.
The image above shows forecast values vs. actual values for the model of E-Bikes and France. The actuals values are represented by green and the forecasted values are represented by blue. As a planner, you can now compare from 2015 to 2020, how close the forecast was to the actual values, with a prediction of the outcome till December 2020.
Transparency in Predictive Planning:
One key notion built into Predictive Planning is to make the models as transparent as possible, so that planners and business users can clearly understand what their model is telling them.
Signal Analysis looks at trends, seasonal effects or cycles, and fluctuations within your data. In this case, we’re able to understand how the sales are evolving over time. The image above tells us there’s an acceleration in the trend up till 2019 which is also reflected by the planning model to continue in 2020. To view a detailed description of Signal Analysis, please refer to 21:00 of the webinar.
Major use cases:
The 3 major use cases of Predictive Planning noted by SAP customers is outlined in the image below. For a detailed overview of these use cases refer to 23:30.
To learn about some of the industries who used Predictive Planning as part of SAP’s successful BETA Program, please go to 28:00 of the webinar.
Refer to 30:20 of the webinar to learn more about the key differentiators of Predictive Planning in SAP Analytics Cloud.
Roadmap of Investments:
With Predictive Planning, we continue our focus on bring machine learning to the customer so they can self-serve their needs better and bring transparency to the next level. With our focus on transparency, you can elaborate and back your predicted outcomes with strong planning models. Please go to 35:20 to learn more about future investments.
Beyond Predictive Planning, SAP Analytics Cloud has many smart features to ease your planning process including Search to Insight, Smart Discovery, Smart Insights, Time Series Forecasting, and Smart Grouping. We encourage you to look into these features, and other methods to make your planning process easier and more seamless.
Join the discussion below: What you like to learn about Predictive Planning next?
For a comprehensive presentation of all the topics discussed here, please watch this webinar.
To learn about an upcoming Smart Predict Virtual Hackathon focusing on real use cases, visit this blog written by Sarah Detzler. Both hackathons, English and German, will take place at European friendly time.
Thank you Zarmina Khan, this was very clear and easy to understand.
That was a very nice explanation. I really like the examples and screenshots. Thank you.