Dale Vile, originally published on The Register
No one disputes the fact that there’s a lot of data kicking around in most businesses. Neither are there arguments about how hard it is to find relevant information when you need it. Sure, there are exceptions, and some very progressive businesses manage their so called ’information assets’ extremely well, but for most of the rest of us, we know things are far from perfect, and the issues surrounding data quality, completeness, coherency and consistency are all too familiar.
As a result of all this, we have seen some great developments over the years in ideas and technology to help, and vendors are continually finding new ways of making our lives easier and our businesses more effective.
Despite some genuine advances, however, a lot of the information-related issues that users have complained about for years don’t appear to be going away. Many still struggle with finding what they want when they need it, collating data across fragmented systems, and suffering the age-old problem of inconsistency and multiple versions of the truth when they do pull it all together. Part of this is no doubt due to the relentless growth in data – you solve a problem in one area, only to have one or two more pop up in others while your back was turned. Keeping up with the growth and proliferation of data can be a very real challenge.
Meanwhile, a lot of the marketing and positioning we hear from vendors is focused on the glamorous end of the problem. The macho ‘big iron’ part of the market with a focus on high volumes and high speed is a pretty lively part of the industry, for example. Here it’s all about data warehouses and the ‘ETL’ (extract, transform and load) tools for populating and managing huge aggregated data stores. We then have a whole bunch of really sexy stuff around ‘analytics’, with intelligent rules-based engines that can find the needle that really matters in the huge information haystack, and even predict whether the CEO will be popping champagne to celebrate at the end of the quarter or hitting the hard stuff to drown his sorrows.
And you want end-user tools you say? Well, there’s no shortage for searching, collating, aggregating, visualising, slicing, dicing, rotating, filtering, trending etc – all now coming as standard in the latest desktop and portal solutions. More so than ever in the history of IT, the super-user has an armoury of weapons to tackle pretty much any information-related activity they feel the need to get into.
But is a lot of the stuff they and other users get into really necessary? We all know users can get distracted by their personal interests and the enjoyment of dabbling with technology for the sake of it, but it would be unfair to focus on this when they are typically so poorly served by the core business systems where the information they need usually resides. And from a business perspective, users having to jump through hoops to get hold of and validate even the simplest piece of information in order to support an action or decision is not great for either performance or efficiency.
With this in mind, there might be merit in implementing the latest ideas and solutions encapsulated by the jargon – ‘business intelligence‘, ‘master data management‘, ‘corporate performance management‘, ‘enterprise dashboards, ‘business scorecards‘, ‘predictive analytics‘, ‘prescriptive analytics‘ etc – but it might also be worth asking ourselves whether there is anything that can be done about the less glamorous end of the problem concerned with straightforward information integration and access.
The chances are that there is a lot of low-hanging fruit to be had by identifying specific information needs in specific places – eg the points in business processes where information is required to evaluate a situation and select the appropriate way forward. In the old days, we used to call this ‘decision support’, and while this phrase brings with it a lot of baggage from the 80s and 90s, it does actually describe the fundamental business requirement in a much more objective way than most of the above buzz phrases and jargon.
As an example, a basic query or report embedded in the appropriate form in the appropriate application could make a lot more difference to business performance than a huge ’boil the ocean’ business intelligence project in some scenarios. Similarly, a small piece of integration work to allow information to be pulled from one application into another at run time could make all the difference from a performance perspective. Process interruptions could be prevented by avoiding the need for a job to be handed off to another department simply because that’s the only place where the necessary information can be accessed for handling exceptions.
The truth is that if you define the problem as ‘decision support’, it is not just information management and business intelligence that are important. Disciplines such as service oriented architecture (SOA) and business process management (BPM), whether you implement them formally or informally, also have an important role to play, as do desktop, web and mobile access technologies. The trick is to think in terms of a decision support architecture or framework that encompasses all the necessary dimensions and layers, and to pick off problems in a manageable way within this.
In the meantime, though, how does all this play out in your organisation? Do you have huge amounts of budget and resource allocated to big central BI projects while a lot of grass roots information and decision support requirements are neglected? Or perhaps you have found a way to achieve the right balance.