The role of traditional OBI (Oracle Business Intelligence) and EPM (Enterprise Performance Management)
This is the first of a number of blogs on Oracle Analytics. Don’t expect deep technical content – this is more about strategies and options with the various software solutions and is aimed at people who are trying to understand Oracle Analytics without the marketing spin. As a blog, I can’t guarantee accuracy and would value comments or corrections at any time.
Before any conversation about big data, predictive analytics or information discovery, it helps to first understand the core information in our business. From an anecdotal analysis of surveys, it would appear that 70-80% of organisations have not yet successfully implemented a core BI solution. The percentage would be lower for larger organisations, but there would still typically be a debate about the value of their investment to date.
Many are on a journey and making steady progress, but with the limitations of siloed source systems, a lack of understanding of organisational metadata, a lack of corporate sponsorship and a fear of biting off too much, many of these applications will generate only a small proportion of the answers we crave.
Departmental solutions can be little more than dressed-up operational reporting and it is no wonder that cheap BI solutions and Microsoft Excel thrive in so many organisations.
If you want a good operational reporting solution then Oracle BI Publisher does a job. It abstracts the data from the formatting and from the output device, allowing you to format reports in MS-Word, create visually rich reports, publish them to an intranet or website and even to burst reports (create localised versions) and email them automatically. You can buy BI Publisher without buying OBI, which makes the cost comparable with some of the cheap alternatives.
If, however, you want to provide interactive features, such as drill-down, complex filtering, more advanced visualisations and automated alerts, then a simple implementation of OBI may be sufficient. Again, with the right deployment, OBI does not have to be an expensive option. For instance consider acquiring 25 licences for power users and an enterprise licence for BI publisher, to roll out reports to a much larger audience, who need to see the information, but do not need the richer functionality of OBI. BI Publisher can use the OBI metadata as its source for reporting, so even a complex OBI implementation can be rolled out broadly.
Business Data Modelling
Discoverer was a tool that Oracle touted to support adhoc analysis, particularly, though not exclusively, for its E-Business Suite customers. Although Discoverer is end of life, it did contain a number of the key ingredients of OBI, including data collection (as long as from Oracle databases), metadata modelling and presentation tools. Sadly, Discoverer fell somewhat into disrepute as organisations attempted to create adhoc reports against source databases, which often created performance bottlenecks. Also, Discoverer did not have a highly-developed visualisation engine, nor a dashboarding tool, so was often used as an extraction tool only, leaving MS-Excel as the tool of choice for visualisation and helping inadvertently to create the MS-Excel cottage industry which has done so much damage to many organisations.
OBI was purchased as part of the Siebel acquisition, and Oracle quickly realised it was richer than Discoverer and end users, executives and IT staff preferred it.
The real benefit of OBI for an enterprise is found in the deployment of metadata, partly as it forces an organisation to come to an agreement over data terminology (leading to greater trust), partly because metadata, measures and hierarchies can be used to create a rich reporting environment without having to build many complex reports and partly because it facilitates the joining up of data from multiple sources into meaningful aggregated data. For instance a charity may store all its expenditure in one application, and all its income in another, but bring the two together in a BI environment. Using metadata and OBI, they can now create true income and expenditure analytics.
Prebuilt OBI Analytics
Leaving aside the very necessary cost benefit analysis, the quickest way to a broad and deep analytics solution is to implement prebuilt analytics, most clearly when built against E-Business Suite, but also workable against any other core business applications.
These applications may not (and could not) contain everything a client might want from a prebuilt solution, but they do provide the fastest way to start, delivering out of the box data extraction, data warehouse, hierarchies and measures and exemplars of reports and dashboards. A second benefit is that they provide prebuilt metamodels, which are much harder to argue against than a roomful of colleagues. These metamodels are built on industry best practice, will not deliver vertical capability, but will deliver the core functions around Finance, HR, CRM, Project Management and Supply Chain.
Essbase and Hyperion - EPM
Quite aside from the core world of analytics reporting, organisations also need tools that allow them to analyse data in a different, multi-dimensional way, also known as what if analysis. This is not exactly the same as an operational or analytical reporting solution, which is looking at real data to determine trends. This instead allows you to throw scenarios at data to predict or model new data. So a simple use case might allow an organisation to enter last year’s actuals, apply a 3% uplift to costs for inflation and determine budgets for the following year.
Essbase is a tool which can do this easily, assuming you provide the source data and the rules/predictions. Essbase can also be used for any predictive modelling, not just financial modelling, so for instance it can be used to model student intake or product sales.
Hyperion applications, such as Planning and Project Management are akin to OBI Analytic Applications, in that they include prebuilt templates, workflow for approvals automation and other features to make the application simpler to embed into an organisation’s processes.
Hyperion also delivers a front-end visualisation tool called Smartview, which allows users to create visualisations of the data in Essbase. This is very popular with Hyperion users, but lacks the breadth of functionality of OBI.
Oracle has invested recently in bringing their BI and EPM environments together, such that OBI can now surface Essbase data and Smartview can be used as a front-end for OBI. Oracle has also hinted that it will focus attention on OBI visualisation development, but will continue to support its Smartview users. BI Publisher can also be used as a front-end for OBI, as mentioned above, so each tool can have its place in the Oracle estate, without overlapping significantly.
In the Oracle world, as in any vendor-reliant community, organisations have to understand their needs and focus before selecting tools to support them. Getting their core Analytics platform right is probably the most important part of any broad information program. Without a platform, key business knowledge and value gained from other approaches such as knowledge management, big data and predictive analytics will be reduced. For instance, a social media sentiment analysis report showing how people are buying a particular product cannot help an organisation unless it can then quickly respond through supply chain intelligence. Or a predicted swell in student numbers cannot be dealt with unless the organisation understands its capacity to grow.