Business intelligence (BI) has been around for a long time in one form or another. It is, however, both becoming more important and changing as organisations see the benefits of data-driven decision making.
Looking back and forwards
In the early days of business intelligence, systems were expensive, complex and often core to a particular process, such as financial year-end accounting. They were, therefore, mostly reserved for those who needed to use them, such as the CFO and finance teams. Now, however, much more data is available, and so are intuitive and easy-to-use analytics packages. This has made business intelligence accessible to far more business units, and even individual business users.
But there are other factors at play. As with any major change – and the changes happening in business intelligence feel fairly seismic – there is an issue of timing. At the same time users started demanding more tools and better access to data, the cloud has also arrived. This has provided new opportunities to move or upgrade legacy systems and old BI tools and databases to a new environment. A 2018 report from Gartner found that many software providers have moved entirely to the cloud, and others are very much "cloud-first’ when adding new features.
It feels like business intelligence is ready to migrate, but where? There are big questions for IT departments and business users to answer about what is required when, and how to make the move.
Thinking about migration
Computing has always been about migrations. From minor software upgrades right through to whole-system changes, it is a part of life. Even with plenty of practice, however, most organisations are not all that good at it. To be fair, quite a lot of organisations are bad at change of any kind, but that’s another story.
I think there are five things that organisations can, and should, do to make migrations of any kind smoother.
1. Think about the timing
There may be particular reasons for timing your migration for a particular period. You may need to avoid particularly busy times, like year-end for a financial system. Are there, however, reasons why you will need to migrate anyway in the next few months? For example, is one of your vendors withdrawing support for an old product that will need replacing? Put these together, and you may find that the timing of your migration is actually pretty much fixed already.
2. Know what you have and what you need
It is crucial that you know whether existing hardware will be adequate for your new system. Do you have the right amount of RAM on each machine? Are there future changes that will require an upgrade in computing capacity, and if so, should you do it all in one hit? All these questions help you to prepare properly and ensure that your migration is at least partly future-proof.
3. Consider whether you want a period of ‘double-running’
Rather than migrate in a single "hit, it can pay to run both old and new systems alongside each other for a while. This gives you time to migrate different departments at different times. This means that you can avoid particularly busy periods, but also ensure that you can provide strong support to new users without your new system experts spreading themselves too thin. This type of support can make all the difference in ensuring acceptance of a new system.
4. Make sure that you inform and consult users
Over many years, researchers have found that it is almost impossible to over-communicate any change. It follows that a migration needs plenty of information. You also need to discuss timing and requirements with users. They may have particular reasons why certain timing is either essential or impossible and also very specific requirements that you might not know about. Working together with users and making them aware of additional benefits in the new environment will make migration smoother and more acceptable.
5. Take the chance to reevaluate what to migrate
Migrations are disruptive. They are also, however, an opportunity. While you are, say, upgrading a system, or migrating your existing business intelligence system to the cloud, consider whether there are other changes that you could, or should, also make. As a migration to a new system also comes with new capabilities, it is worthwhile to reassess the business requirements. You might, for example, deprecate some old content and rebuild it from scratch because the new system can fulfil requirements better or adds usability for the information consumer. This hybrid approach can add a huge benefit without very much additional disruption.
Any migration should, of course, always focus first on business need. However, it is also an opportunity to improve things, and this chance should not be wasted.