One of the biggest ongoing discussion points, and in fact a key decision factor to be bedded down prior to implementing a new Enterprise Performance Management, or Business Intelligence application is whether it should be implemented as an on premise system, or as a hosted or a cloud solution.
We have come fairly quickly from a place where we had hardware on raised pallets in dark corners of basement floors. These were shrouded in the mystique of IT empires, and the access to information on them was definitely not available to Joe Blogs at his desk, except that it was delivered to him (sometimes in hard copy print) after applications in triplicate had duly been authorized!
We are now in the place where as far as possible, the business attempts to give the individual user access to self served information on their personal device, at any time.
The analogy I often use for this is one where my grandfather one day clung to one end of a firmly planted tulip and said to me; “Look how strong I am, I have the world on the other end of this!”
The internet used to be the domain of the few, and now it is almost impossible for an individual to function without it.
So where does this leave us in respect of the best approach to delivering what our Business needs to it’s various user communities? Putting our “simply said” hats on, let’s examine some of the key facts. We can loosely break these into four categories:
The complexity and scale of the application, including the size and regularity of data integration requirements
The skills and effort required to implement and maintain a suitable environment for the application
Company policy (effort to change, often linked to four!)
The cost (taking into consideration life of hardware).
Complexity and Scale
Where we have large intricate applications, with voracious appetites for frequent large scale updates of metadata and data it is sometimes best to leave the application on premise. I would recommend that in this case, a thorough review of the application is done anyway. Sometimes models that are big and ungainly are the product of years of “enhancements” and add ons to cater for a changing business. It is also common that these “monster” applications are being used for everything from planning to consolidation, to management summary and operational detailed reporting! Taking a step back and revisiting the “raison d'etre” for the application can result in beneficial streamlining and sometimes rebuilding of a more suitable “fit for purpose” application. The new application may well then fit within the bounds of an application suitable for cloud.
This is not always the case, and it is important to state that it is not a sin to have an on premise solution! The millions of dollars spent by oracle on developing it’s engineered hardware and software is testament to this.
Applications which are more suitable for cloud are typically those that:
- Have a fixed purpose (e.g. Planning Application, Reporting Application, Close Management Application)
- Have less need for large scale movement / constant change of metadata and data (This can be catered for on cloud, but can be hard to achieve without additional tools such as Qubix Cloudbridge)
- Applications in organizations where IT departments are under pressure, both from a funded cost perspective as well as simply having bandwidth issues.
As Cloud technology develops, even some of these factors will become less important to consider. Who would ever have thought that we would have “Big Data” on the Cloud!
Skills and Effort
Hardware has come down in price over the years. However, the prospect of setting up a robust secure environment with correctly installed EPM or BI software requires skilled resource. Qubix has fairly often come across situations where an Organisation has attempted to use internal resource to perform Oracle EPM and BI installations. Often the internal resources doing this work are very capable people. However the unique complexities of these hardware ( if using Oracle “Exa..” machines) and software installations dictate that the more cost effective approach is to pay a company that performs these installations every day. They understand the methods to be employed, along with the unique idiosyncrasies of each product, to ensure a robust platform for the solution is provided.
Some larger companies have their own dedicated Hyperion dba’s, who are extremely well qualified and up to date, a number of Oracle Aces amongst them. Where this is the case, it can still make sense to have and maintain one’s own “on premise” environment.
Cloud has changed the topography of where the installation is, but more importantly who is responsible for it.
In the case of Oracle cloud, It brings to mind that infamous phrase: “Look Ma, no hands!”
- Oracle handles the complete installation
- Oracle deals with all the patching
- Oracle deals with all the backups and redundancy.
For a company to be up and running, they simply need to make the purchase and get switched on. It will still be necessary to get some assistance from consultants to develop the application, but a major part of the time and cost is removed from the equation. The advent of software subscription brings with it the opportunity to pay for consultancy via subscription as well.
We have policies in our Organisations designed to protect us. These relate to both internal and external compliance requirements. Where these are constantly reviewed in line with a growing business, they instill good practice, and facilitate robust business processes. Where they are not, they become like a noose gradually tightening around the neck of progress. This may well be the case when considering on premise versus cloud.
Those that are sometimes hardest to impact are those to do with security, and the idea of having sensitive data residing on a cloud somewhere off premises. In my experience the care taken around securing environments and their data is generally more robust in the Oracle environments than some Company’s own or third party data centres.
The fact that it is a subscription means that it could live in the P&L and move off the Balance Sheet. The hoops to jump through for Capex can be obviated.
It brings high end hardware and applications in reach of even small organisations. With the opportunity to buy as few as 10 licences at a time, even a single department within a larger organisation can afford to buy what they need. This can cater for the need to create “non-finance” applications to underpin a Marketing planning process, or a Sales Department wanting to impact logistics with their forecast demand, and so on..
As mentioned previously, there is still a need for outside consultancy to help develop the application. This is commonly being sold as part of the subscription. Depending on the scale and complexity of your application, you can generally choose an amount of consultancy to include with the subscription.
Please see: http://www2.qubix.com/cloud-planning-tco for some examples which include software and various “packages” of consultancy.