.png)
Five architecture thinking hacks for your next IT project
Written by Mark Smith - Senior Consulting Architect
If you’re a NZ SME about to replace a system your business depends on, whether that’s job management, accounting or a CRM, you can’t afford for it to fail. Budgets are tight, and a stalled project ties up the same few people who keep the business running.
The Standish Group has tracked IT projects for about thirty years, and the reasons they go wrong have barely changed. Projects are too big, goals are fuzzy, the people who’ll use the system are kept at arm’s length, and the person in charge gets pulled onto other things. The good news for smaller businesses is that small projects succeed far more often than large ones, but there is still some upfront work that improves your chances.
Part of the reason that IT projects struggle is unexpected complexity. Software on its own is complicated, but it isn’t complex. Add people, habits, the things only one person knows, and the undocumented workarounds, and it can get complex.
In complex systems, small changes can have big effects, and big changes can have almost none. Early decisions are hard to undo, and no plan can predict exactly what will happen. More planning won’t remove that uncertainty, but the right kind of thinking upfront goes a long way.
Here are five things we think give you the most bang for the least effort, and why they work. Most take days, not months.
Agree what this project will and won’t fix. Write it down and make sure everyone has seen it. Early decisions are hard to undo, and “fuzzy goals” usually means no one agreed them up front or where people tacitly assume different ones.
Put your capabilities on one page. List what your business does and flag the few you think set you apart. Spend customisation effort there and run the rest close to out-of-the-box. Get this wrong and you either customise low-value processes or force what sets you apart into a platform that won't support it.
Find out how work really gets done. Look for the spreadsheet someone keeps open, and the person everyone rings when things go wrong. These exist to cover gaps in the current system. If the new system doesn’t cover them too, they’ll reappear.
Try small things first. Run one team, one site or one process on it for a few weeks. You can’t predict how people and systems will react to each other. A cheap test teaches you more than a detailed plan.
Agree what you’ll keep an eye on. Pick a handful of signs, including soft ones like the side spreadsheets creeping back. Check them at a regular meeting with whoever owns the decision. People adapt to whatever you measure, so treat the numbers as a prompt to talk, not a scorecard.
None of this needs a specialist team or a formal method. A whiteboard, honest conversations and the right people in the room will get you most of the way. If you’d like a hand, it’s work we do with organisations of every size.
Planning your next IT project? If you’d like a hand working through your options or making sure you’re asking the right questions, we’d be happy to chat.
Comments