I have done quite a number of inflight reviews and post-mortems of troubled and failed large system implementation projects.
One pattern that emerges very clearly is the one of missing milestones whilst keeping the go-live date the same.
It rarely ends well.
I see it again and again. Multiple important milestones are missed. Sometimes by months. And the ones that are marked as completed have their original scope reduced.
For example system integration tests (SIT) without all interfaces being completed and no production like data.
Or user acceptance testing (UAT) with systems that are not ready or contain so many bugs that end-to-end testing is not possible.
Astonishing is that in most cases both the project sponsor and project manager seem to be convinced all is “green” and it will work out until the project folds like a house of cards.
When you look at a typical large system implementation project it is still largely implemented like a waterfall. This includes ERP systems, CRM systems, Core Banking, etc.
And this has not changed with the rise of software as a service (SaaS) offerings like Salesforce, SAP S/4HANA, Workday, etc.
Yes, the design and build phases are now iterative, but at a certain point your full solution needs to be tested end-to-end. This means one or more SIT phases and an UAT phase that includes all upstream and downstream systems and processes.
You also need time to fix all the findings of your testing, and to do re-testing. If you are lucky one cycle is enough. Usually it is not.
You also need to train all your users and your support teams on the new solution and processes. Ideally on a solution that actually works.
And when you are ready to go, you have a cutover phase from your old solution to your new solution.
So yes, you design and build iteratively, but the rest is still shaped like a waterfall.
And this means that if you miss important milestones and you don’t change the go-live date you will steal time from the very important phases that come at the end of such a project.
Starting these late phases without having completed the previous phase just does not make sense and will drive your test team and end users crazy.
Missing milestones does not mean your project team is doing a bad job, but they obviously underestimated the time it takes to do certain things.
Chances are this is a pattern that is repeated for the later phases of the project.
So you will probably need more time for these phases as planned. Not less.
In my experience there are only two probable outcomes of such projects:
1) They never go live
2) They go live too early
The latter can be even worse as the first.
See here and here for some prominent examples from multi-million projects that never went live, and here and here for projects that went live too early.
You will find many more examples among my project failure case studies.
In a nutshell: missing milestones and not changing your go-live date is a great leading indicator for trouble in the future.
