I reckon most of us have heard the phrase ‘death by PowerPoint’, referring to those moments when you’d rather just give up than see yet another slide stuffed with burping, animated clip-art. There is another phrase some of us are also familiar with, that’s probably more pernicious – ‘death by spreadsheet’.
Imagine the scenario – a team of analysts and business experts are putting together the requirements for a complex new application. Will they model their business processes using a process modelling tool? They might, but let’s say they don’t in this case, because they don’t have a process modelling tool, and they’re too expensive. Do they use a UML modelling tool, producing activity diagrams and use cases? Probably not – isn’t UML for system designers and developers? How about producing high-level data models in a data modelling tool? I doubt that very much – I haven’t met many business analysts comfortable with producing data models. Okay, what does that leave? By the way, did I mention the need to manage and compare versions of those requirements, and allow several people to update them at the same time? Oh, don’t forget that we need to link those requirements to design artefacts and other stuff, like data models and processes.
Of course, why didn’t I think of it? The perfect solution – let’s use a whopping great Excel spreadsheet, and use SharePoint to manage versioning and sharing. Even better, we’ll develop a set of linked spreadsheets, each one designed for a special task. That’ll do the job, for a while at least, but for how long? What about when person B wants to edit a spreadsheet that person A is working on, so person B takes a copy to work on. Then they store it in their own SharePoint folder. Eventually, there are ten copies spread around various SharePoint folders. Never mind, the cost is still low, isn’t it? We already have licences for Excel and SharePoint. Anyway, I’m sure we can use hyperlinks to connect requirements to other documentation, models, etc. It’s easy to link to another file, and maybe to a bookmark of some kind within a file. But (there’s always a ‘but’) what happens when the name of the file or bookmark changes, or you want to link to more than one of them? Can we see the link from the ‘other’ end? Of course not, that’s not how hyperlinks work. No problem, what we need is another spreadsheet where we duplicate this rats’ nest of connections in a searchable manner.
It’s pretty low maintenance, just a few hours needed each week, making sure everything connects, and updating that other spreadsheet. After all, it must be cheaper than spending all that money on tools mustn’t it? In one way, it is cheaper – we haven’t spent any money on software. However, we do have a large overhead maintaining all this stuff. Still, we won’t need to maintain it for ever, will we? Once we’ve gone live with the second or third release, we can afford to chuck away all that work, can’t we? It doesn’t matter if we can’t answer future questions like “why did we do it this way, when that isn’t the way the package does it out of the box?”.
It’s ironic, isn’t it? The organisation is in a pickle because we’ve built all these data and process silos over the years, and the solution is to simplify things using a complex package, service-oriented architecture, a combination of both, or something else. So, we’re replacing the organisation’s process and data silos with something that integrates much better, reduces effort and cost, and is less liable to lead to confusion, error, or even jail sentences for those held responsible. How do we manage the design and implementation of this new nirvana? We build our own metadata silos (multiple spreadsheets, data models, etc) which don’t link together, each group develops their own deliverables, and little effort is spent on making sure all these silos connect with each other. I forgot, that’s what spreadsheets are good at, isn’t it? Nope, that’s what integrated tools are good at. After all, isn’t that what we’re giving the business people, a set of integrated tools for managing their business, often to replace their own version of ‘death by spreadsheet’?