Four Out of Every 5 IT Data Analytics Projects Will Fail in 2019
4 out of every 5 IT Data analytics projects will fail in 2019 according to Gartner and it seems as though it’s getting worse. If you don’t believe that just walk up to a marketing person and ask them to show projected sales and revenue for next week on their cell phones, then watch the blank stare. The technology to do so has existed for decades, so why don’t people have it? It’s because their analytics projects keep failing, and we will examine 5 ways to succeed.
1.) Start with 'why' not 'what'
If you notice, many projects start with the what, actually, some don’t even start with that, and if you can’t clearly define what you are building, you will be doomed to fail from the start. But I digress. There is a step before the 'what', and that is the 'why'. I’ve seen project plans and business cases that spend 20 pages spelling out the what and not even 1 sentence explaining the why. Whenever you are going to put together a project, start with why you are doing it. It may seem obvious why you want to report revenue by month by location, however when you really break down why and use the 5 why technique, you may be surprised. Prior to leading BI efforts, I spent a decade as a developer, and I can’t think of one time someone told me why we were doing what we were doing. As a leader and practitioner, this question is never left out of the conversation.
2.) Dedicate the team
Another success factor is having a team dedicated to the mission. You want full time dedicated people, not people whose regular job is something else and they dabble in BI. Have dedicated FTEs (full-time employees) on the task. Even if you must bring in a consultant team, that’s ok as well, however still dedicate at least one FTE to managing the consultant efforts.
3.) Get some help
One trait that is consistent across the board is the ability to get help when you need it. As IT professionals, we tend to think we can do it all, you could have been in IT for 30 years, however, unless you have specifically built enterprise data solutions like data warehouses, then you’re going to need some help. Swallow your pride and get some assistance.
4.) Don't be afraid to quit
I’m not talking about giving up on the entire reason you are doing something, and the need to quit will be lessened if you understand the 'why'. But sometimes you will find yourself going down a dead-end road. Don’t be afraid to call it what it is and abandon that route and try something new. Many CIOs will be reluctant to admit a current path isn’t going to work. I used to think they just didn’t see it; however, I’ve witnessed CIOs being informed by consultants and team members alike and just wouldn’t abandon the path simply because it’s the one they promised. Once again, swallow your pride, regroup and redirect. Don’t go down a stubborn failure road. You will be less liked by the rest of the C suite when you end up 2 years down the wrong path and spent an extra 2 million when you could have stopped the bleeding long ago.
5.) Think big, Start small
We’ve all heard the term don’t try to boil the ocean. That’s true on a data project as well. It’s fine to have a roadmap to integrate everything under the sun, but don’t try to implement it all in one two-week sprint. As obvious as this sounds, I still see plenty of organizations trying to go this route. I’m exaggerating with the one sprint analogy, but the spirit is there. Have this in mind, one source at a time, and my project will be just fine. Build, integrate and then move to the next one. Focus your efforts.
Building quality data solutions is simple, but not easy. Identify the issue, define the 'why', document the what, build slow, deploy and train, keep the business involved every step of the way and get some help when needed. Sounds simple right? It is, just not easy. Follow these simple steps on your next project and hopefully, you will join that coveted 20% who enjoy data success. And hey, don’t worry, B I