Mathias Eifert

Helping people identify, connect and interpret the dots, mostly in IT.

[Mathias Eifert @mathiase]

There is no product I have ever more hated than my @HP printer. Unable to print from Windows 11 even after installing their software. "Printer requires attention" even though it prints just fine from a Windows 10 machine on the same network. YOU PEOPLE HAD ONE JOB AND YOU FAILED!

[Attention Alchemist @alexbunardzic]

Horizontal slicing is how most software teams work. The most difficult/challenging software skill is vertical slicing. Those who master it attain software superpowers.

[Mathias Eifert @mathiase]

@ejames_c If a lot of your basic equipment is broken and nobody bothers, being good at data is maybe not your highest priority? Most orgs seem to struggle with the 101 levels, mostly due to short-term "numbers" pressure.

[Mathias Eifert @mathiase]

@ejames_c I think in either industry the gap between the best and the rest is enormous. Very few manufacturing firms do Lean/SPC well, very few software orgs do Lean/Agile well. Mostly a leadership/culture problem, only then a knowledge problem.

[Mathias Eifert @mathiase]

@ejames_c Just curious, are you familiar with Cynefin? Nigel Thurlow's (et al) FLOW system? Dan Vacanti's work on Kanban flow metrics?

[Mathias Eifert @mathiase]

@ejames_c One of the biggest shortcomings of the Agile Manifesto is the implicit assumption that all items in the backlog have value and thus all of the resulting software does. But that's because they wanted to address broken software development first. To get to some level of control.

[Mathias Eifert @mathiase]

@ejames_c Software development is/happens in a complex adaptive system, so they went with principles and heuristics rather than a prescriptive playbook.

[Mathias Eifert @mathiase]

@TheCynefinCo @snowded Would a cross-functional team be considered a “silo” in the context of this post? The specialization and shared mental model would be focused on problem / value domain rather than sameness of skill set?

[Mathias Eifert @mathiase]

@swardley Probably safe to assume that many of the more diverse, less popular folks don’t pay Elon any money so their posts get less visibility from the algorithms?

[Mathias Eifert @mathiase]

@ejames_c All that said, my background is in TQM/SPC, so maybe I am just mentally (too much?) contrasting software dev with the idea of large, repeatable runs of same things. Your application to the processes behind selling items or gaining subscribers seem close enough to that, so yeah 👍

[Mathias Eifert @mathiase]

@ejames_c Appreciate this discussion, btw.

[Mathias Eifert @mathiase]

@ejames_c @danielwithmusic Well, technically it's what's in the Manifesto and I don't claim more wisdom than that. But saying that Agile doesn't work because someone has been subjected to shallow Scrum (my sympathies!) is kinda like saying tennis sucks after playing doubles at the local senior center.

[Mathias Eifert @mathiase]

@ejames_c @danielwithmusic Afaik the statistics don't work as well without a "stable process". If there's low volume of data points and a lot of VUCA, you can have relatively high special vs common cause variation and thus the outliers would seem to be less clear?

[Mathias Eifert @mathiase]

@ejames_c @danielwithmusic I meant software Kanban. Not TPS Kanban.

[Mathias Eifert @mathiase]

@ejames_c @danielwithmusic In high complexity situations, special cause variation dominates because a) few data points (low volume) and b) novelty -> limited repeatability. Good Agile implements short feedback loops that allow for rapid adjustment. Process control tends to work better for ordered systems.

[Mathias Eifert @mathiase]

@ejames_c @danielwithmusic Sorry, but no. This thread confuses Agile with (bad) Scrum. The Agile Manifesto mentions none of these "low-value processes" - there is no "Agile process". Agile at its core is about narrowing the uncertainty funnel and adapting to emergent reality.