I agree that there's a lot of wiggle room, but there's still some ground rules, like being iterative, not having to plan for everything beforehand, deliver smaller things little by little, not being a perfectionist and turning the review process into a living hell.

But you are 100% right about the problem being people. The problems I see with agile are always because people try to find escape hatches in the process. Designers want to prepare every single detail before developers start working and hate iterating because fear of the "original vision" being a mistake, product owners don't want partial stuff in the beta channel so everything has to be in its final state, product owners don't know or don't want to reduce scope so developers end up with impossible sprints, developers want a perfect definition finished before starting work, so zero chance to iterate...

I also agree with people throwing out individual/interactions and picking a heavy-handed process instead :(

> 30% of IT projects still fail, despite “agile” now being widespread.

That's a very optimistic statistic, unless "fail" is defined to be something other than "the opposite of success".

IOW, a fail is any project that goes over time, over budget or delivers under scope.

If you're defining success as "well, we delivered eventually, and we delivered 80% of the requirements, and it only went over budget by 20%", then, sure, only 30% of IT projects "fail".

Ok, by that measure I guess over 90% of projects would fail. Which would probably be unfair in most cases, because unlike this project, where the scope as defined by the client was 100% clear from the start, most clients when starting out with a project can't really say exactly what they want, so the scope creep is not only driven by the agency, but also (sometimes mostly) by the client.

> most clients when starting out with a project can't really say exactly what they want, so the scope creep is not only driven by the agency, but also (sometimes mostly) by the client.

Most clients know exactly what their goal is ("To increase market recognition", "to increase conversion rate", etc). The scope creep can only happen in a requirements phase[1]. Once a client has signed off on exactly what will be delivered, and what the payments milestones are there's no wiggle room for the agency to continue experimenting.

If you're skipping the requirements phase because "agile"[1], then the work will be over budget, over time and under the scope.

There's very few projects that actually require the exploratory-driven process that agile uses. If you have one of those projects then farming the management of it out to an agency is exactly the wrong thing to do.

[1] There may be multiple requirements phases. For example a large "need to revamp our website" project can be split into multiple serial projects, each with their own requirements phase.

[2] I.e. We'll figure out what is acceptable to the client with constant feedback.

I think the goal is not to eliminate "failure", but reduce it. Failure will always exist but it always requires interpretation. Thus the credos of "failure is learning experience", "don't be afraid to fail" etc.

It's a very abstract term. To put in simpler terms you will always have some outcome mismatch and generally your goal should be to reduce it. But if you do not have outcome mismatch, then you should push the line so that you can look ahead and understand future areas for improvement. Or you can just stay content if "market" is not chasing you.

The problem as I see it is that projects often start without a vision, or even clear goals; even if there is a vision, it ends up getting lost and forgotten in a sea of feature requests.

Not only should we be asking if a project has achieved its original goals once it’s finished, we should be continually assessing progress against those goals throughout the life of the project - yet few projects I’ve worked on actually do this, even when I’ve explicitly requested it.

I've had someone tell me, with a straight face and serious voice, to "stick to the agile process from now on"

only 30%? I'd have guessed that most IT projects at least fail to achieve goals if they don't crater completely. the bigger the projects, the more they resemble interplanetary collisions.

Well, that’s according to the PMI. Finding the link will simply lower my level of happiness for the day, but you should be able to Google it (sorry).

The thing is that as far as I can tell, “success” is measured post-hoc, so all sorts of shenanigans might have taken place in order to determine that a project was “successful”. Having worked with my fair share of corporate project leaders, I’m quite certain that the actual number of failed projects would be much higher if measured by the original criteria used to justify the project in the first place.

Helmuth van Moltke (1880): "No plan of operations reaches with any certainty beyond the first encounter with the enemy's main force." It applies to software projects too, especially large ones. The enemy is the real world (or our lack of understanding of the real world.)

If we look at the original requirements every single project is a failure. If we look at the strategic goals and keep track of all trade offs maybe 30% is a fair guesstimate.