Predict with Oracle BI – Look into the glass bowl
Oracle BI – Thoughts (9) – Cadran publishes a series of articles about the ideas surrounding Business Intelligence in combination with Oracle JD Edwards ERP. In these articles various considerations and reflections are discussed, which can be helpful in making the right decisions in the implementation and application of both systems. The previous article discussed the development cycle and maturity steps and the reporting options in JD Edwards that are part of this. This article goes one step further and we look at the possibilities for predicting the future.
Oracle BIEE applications
Oracle Business Intelligence enables you to achieve strong improvements in the organization in both the quality of data (master data and transaction data) and in the processes. At an operational, tactical and strategic level, analyzes, graphs and KPIs provide the right management information to improve the organizational, administrative and operational processes.
Previously, we have seen that a KPI (Key Performance Indicator) is a strong tool for helping Oracle BI to identify situations that require adjustment and intervention. To this end, we provide facts or measured values of valuations or targets so that Oracle BI can assess them. An example of this is sick-leave. If this is, for example, 4%, then this is only a fact. If we add the judgment “too high”, Oracle BI can signal this and help you steer.
In Oracle BI itself, it is possible to look ahead in time in all kinds of ways on the basis of the source information. When smart trends are set up in dashboards and analyzes, an estimate of future developments can be made. A practical example of this is an analysis, in which the top 50 customers of last year are used to estimate whether the customer will show more or less sales this year. This can be done by extrapolating the expected sales of this year with a simple linear progression. The formula on which this is based can, however, be made as complex as is conceivable.
With information from Sales Forecasting from JD Edwards, an even greater range of possibilities for future predictions is unlocked. Depending on how smart data is fed into that module, Oracle BI is able to quickly make this transparent with visualizations.
Oracle BI is able to provide you with proactive signaling through so-called Agents. These are jobs that are periodically executed automatically. Based on KPIs and the criteria that are linked to this, they can be identified by means of e-mail or SMS and an Alert Section on your main dashboard. Users and / or user groups can also subscribe to the alerts in which they are interested, as a result of which these alerts will behave like newsfeeds.
KPI Watch Lists & Scorecards
Users are also offered the possibility to compile so-called KPI Watch Lists (not to be confused with Watch Lists in Oracle JD Edwards), so that a bundling of separate KPIs arises that give an overall assessment of the performance of a particular (sub) process in the organization. An example of this is the monitoring of the warehouse process, which is done based on KPIs such as turnover rate, number of backorders, number of backlogs or short-comings in stock, order picking speed, and so on. Only when all individual meters are green, the above overall KPI will turn green. These Watch Lists are then connected to departments, creating an even higher level. For example, a process such as Sales Order Management can be provided with several of these watch lists and linked to the responsible departments, such as Customer Service, Logistics and Accounts Receivable, which are involved in this total process. This functionality is known in Oracle BI under the name Scorecards. This functionality can then be lifted to an organization-wide level. This allows targets to be set, such as a target margin for the current year. All components, departments and processes that influence this can be monitored and controlled at the highest level in this way. Only when all the underlying elements meet, is it clear that this goal has been realized. Using tools such as budgets, goals, forecasts and trends, the management becomes increasingly intelligent.
Conditional Dashboarding & Actionable Intelligence
A strong concept of Oracle BI is so-called Conditional Dashboarding. A section on a dashboard, such as the top worst paying customers, may only be interesting when certain limits are exceeded. Oracle BI is able to apply such conditions in analyzes and dashboards. This is also known as Actionable Intelligence. A link to a detailed analysis or drill-down option is only offered if there is actually something that needs attention. This is Management by Exception. In a world where there is a lot of automation, we are only interested in things that require attention. Performance Management Above Oracle BI tools are Oracle EPM (Enterprise Performance Management). They use the tools in Oracle BI and take it even further to make it even more proactive and dynamic to monitor and manage the organization on performance. These tools originate from Hyperion and are integrated into the total software package that Oracle BI can offer. It is nice to see which customers have delivered the most sales and how the trend for the coming year might look like. Predective analyzes may be able to answer questions such as: Which customer will make the most profit? Which product will customers no longer buy this year? What if demand falls behind on the forecast? This shifts towards what-if analyzes and the probability associated with the different scenarios. Historical reports can support this, but they are not omniscient. However, they are one of the sources that will feed what-if analyzes. On the basis of these figures certain scenarios can be created, with which to investigate what would happen to those figures if … …
- more people would work in the warehouse? …
- would a new product line be added to the assortment? …
- a new distribution center would be opened in the south of the country? …
- the new catalog is launched a month earlier? ….
With these scenarios a medium-term prediction for the next five years can be made. A variety of parameters can act as volume adjusters and thus model different outcomes. These different results can be stored as photos, so that this data can be used in Oracle BI to compare them with the actual results and data after those five years. For example, it is possible to determine how certain choices actually worked out over the years.
A term that we come across several times a day. But is it just a buzz-word or does it mean more? In the coming article I want to go deeper into the concepts and techniques that have to do with this and the practical details that an organization can give to it.
Author: Rick Brobbel
BI Consultant at Cadran Consultancy