Predictive forecasting is all around us — Google’s autocomplete, weather forecasts, your GPS knowing when you’ll arrive at your destination — they’re all forms of predictive analytics.
What is predictive analytics?
Predictive analytics uses multiple techniques such as data mining, statistics, machine learning, and artificial intelligence to predict the possibility of something happening based on historical data.
Predictive forecasting may seem mysterious, but the logic is actually quite easy to understand.
The way many machines work is by following a strict set of commands outlined in computer code. An alarm clock, for instance, is programmed to perform a few relatively simple tasks — display the time and turn on the radio at a pre-programmed time.
But what if instead of programming the alarm clock to turn on the radio at a certain time, you wanted it to wake you up at the optimal time based on your sleep cycle? For that, you would need a more sophisticated alarm clock, more data, and more coding.
Assume for a moment that your alarm clock is a more sophisticated machine that can track all of your sleep data such as:
- When you go to bed
- Your sleep cycles
- All the times you toss and turn
- When you get up to use the washroom
- What position you sleep
- Your body temperature
After a while, you would have a large volume of your data of your sleep cycles and nighttime habits. Your sophisticated alarm clock could then use mathematical algorithms to look for patterns within the dataset. For instance, it notices that you hit snooze twice on Mondays, you wake up at 5:30 a.m. on Wednesdays, tend to sleep in on Saturdays, and so on.
But what if we fed the machine more data such as your productivity levels at work, your mood, your caloric consumption, your metabolic rate, your fitness routines, your social habits, and so on. It’s likely to find even more patterns and relationships within the data.
It can tell that you’re more productive on the days you get more rest. You’re in a better mood on Fridays and Saturdays. You tend to go out for drinks on paydays, and you tend to order particular kinds of foods following days where you were most active, etc.
Using all this data, it seems as though the machine would know your habits quite well, right? It also seems likely to be able to make intelligent guesses, or predictions, with a high degree of accuracy as to your future behavior based on what you did in the past.
This is precisely how predictive forecasting works. SAP Analytics Cloud predictive forecasting looks at historical data to find patterns, and then uses those patterns to make predictions about future outcomes.
Predictive forecasting for business
As we saw in the previous example, predictive forecasting uses historical data to predict future outcomes. For example, a sales manager may use predictive forecasting to project sales revenue for the upcoming season.
Predictive forecasting takes into account different values, trends, cycles and / or fluctuations in your data to make predictions. This is a powerful data-driven approach that can be leveraged to aid your planning process.
SAP Analytics Cloud predictive forecasting gives you an unbiased understanding of your key business influencers and allows you to dive deeper into your data with interactive visuals. You can then use these findings to influence business decisions.
How to use predictive forecasting in SAP Analytics Cloud
To use predictive forecasting in SAP Analytics Cloud, upload a dataset and make sure there’s historical data so the system can use this to look for trends. Once you create a chart, you can apply predictive forecasting.
First, select a chart you want to apply predictive analysis to and change it from a trend chart to a Time Series chart. Note, it must be enriched with dynamic hierarchies to show different levels of granularity.
Next, select one of the forecast options:
- Automatic Forecast — forecasts future values based on historical data within the chart
- Advanced Options — offers the ability to include additional data as a potential forecast influence
The forecast is added and is indicated by the highlighted area with the dotted line. When you select a data point, you can see the upper and lower confidence bounds (i.e. probability of accuracy in the prediction). You can also edit the forecast if you like such as change the color or add another variable.
SAP Analytics Cloud predictive forecasting helps you understand past data trends to predict any metric in the future. The predictive algorithm classifies existing information, identifies outliers, and surfaces relationships within your data to help you see and understand your business’ key influencers.
You can use SAP Analytics Cloud predictive analytics to:
- Discover key influencers of your KPIs such as revenue, churn, and productivity
- Explore interactive charts and graphs, automatically generated based on your query
- Create predictive forecasts to predict future results based on historical data
To see how SAP Analytics Cloud predictive forecasting can benefit your business, try it today for free.
Please note, you will need planning rights and a planning license to run a predictive time series forecast.