Most features in software products are never used

There are multiple studies and statistics on line. This resembles the division between touch time and wait time and looks like it might be just as fertile a space to reduce development effort. If you could build software with only the features that will be used… you would save 60 to 80 % of the development effort.

@kentyer: the Full Kitting activity, if done properly, will prevent “featuritis” from developing.

Yes, Full Kitting has a big impact, but I think the basic problem is more deep-rooted. The feedback loop between the “producer” and the “client” is just too long…unless you expand the full-kitting to have some equivalent of Xp’s “client in the room”, which I think is very possible in our current environment.

Well, the client is in the room during the FK activity.

Of course, if you are acting at scale serving markets, then you will have proxies, like product owners or product marketing managers. If you are acting at scale and serving a multitude of bespoke project clients, they you will have key account managers or other representatives. But the clients voice is fully present or represented in the FK activity.

FK is where demand meets supply. It does not make sense to think of FK without the demand side. If you thought that was not the case, you need to dig deeper into what FK truly means.

(Also note that when you are acting at scale you have even other, more compelling economic reasons not to take on unnecessary or non-valuable work rather than just avoid feature creep…)

This helps me “dig deeper”, thanks.

It might be worth doing a blog post or a video on this… i have searched through the 3 books and find no mention of the client during full kitting, only of the “proxies”, like product owners or marketing managers. I think those people will not have enough “deep” knowledge to take action to keep down feature bloat.

You need perspective here.

In principle we would like to be functioning like XP. And that is entirely possible in the small.

But when you are in a PEST environment, it is just not materially possible. The “proxies” are unavoidable. They have to do a lot of the conceptual aggregation work. There is simply no space / time to have the direct relationship there. Even so, a properly functioning Full Kitting protocol will avoid most of the work that does not provide value.