The article that sourced this statistic is marketing clickbait meant to sell an Impact Engineering book. At best it’s incredibly ignorant and at worst it’s intentionally deceptive. I lean towards the latter.

What I’m referencing is a stat from an impact engineering book by Junade Alimeant, which resulted in articles being spread around with summarizes of those findings. The ones most often cited are “96% of agile transformations fail” and agile projects have a “268% higher failure rate”. These are manipulated stats.

According to the study citing the 268% higher failure rate, it was calculated by making the connection that agile requirements engineering equated to “Development starts before clear requirements, no complete specification, significant changes late in development.”

I’m not sure where that definition was taken from, but it’s not anywhere I’ve heard of connected to agile development and never experienced development starting before we knew the requirements. Nor was there a specification that was never complete. If you started a project before you knew what you were doing and called that “agile”, yes, that’s likely going to result in higher failure. At the same time, this would also apply equally to waterfall or any other methodology. This is not a principle of agile, it’s connected to project chaos.

The last part of that could be associated with agile, since it does welcome changing requirements. However, it specifically goes out of its way to state “significant changes late in development” which seems like a pretty specific slice to take out of there unless you were trying to get the stats for projects that failed the most. Again, this is a principle of project chaos – not agile. If a project significantly changes late in development, there’s not anything that’s going to save that.

One other curious part of the study is that “lean software development” was defined as “Only working on one project at once,” which I’ve never heard as a definition of that.

The 96% agile transformation failure is a blatant twisting of words and statistics. That stat comes from the 12th State of Agile report which states on one page: “4% report agile practices are enabling greater adaptability to market conditions.” What the impact engineering article ignores is that it was an agile maturity breakdown, not a percentage of teams who had adopted agile but had failed; it’s intentional misinformation.

This article and book are getting passed around and cited as a concern about agile thinking. I can easily see this being forwarded in an email to management and a seriously damaging decision being made based on this deceptive, clickbait book. Agile thinking is not suddenly being outed as this massive pit of failure.

Don’t fall for it.

Side note: the image Copilot generated for this article was incredibly funny.