Thread by @patkua

I had an interesting question for the talk I did today for @agilesgorg about my view on developer productivity like BlueOptima. I hadn't heard of it before but doing some reading now and 😱. Here's a quick thread with some initial thoughts

First there's a reminder about en.wikipedia.org/wiki/Goodhart%… (When a measure becomes a target, it ceases to be a good measure), or the #TOC version of this, "Tell me how you measure me and I'll tell you how I behave"

Leaders and managers need to think really, really, really hard about the metrics they use and the obtuse ways that people will game them. Also... engineers have an astounding talent for gaming the system, so beware... but back to the tool

I never take it as a good sign when you have to ask for a demo of a product before understanding it, so I read this article: reviews.financesonline.com/p/blueoptima/ It seems to have three metrics - ACE (Actual Coding Effort), ART (Analysis of Relative Thresholds) and some Performance Benchmarks

ACE seems to forget that often the best code that an engineer can right is no code (i.e. solve the problem without writing code). Code is expensive to maintain. Secondly measuring code is like measuring activity or effort. Not all code delivers the some value.

A really great engineer is also doing more than writing code. If they're working on a product with a team (who's not these days), you're missing out all on the value added activity they do (like sharing knowledge, providing feedback or advice to other team members).

How does ACE measure this? It doesn't. Which means bad managers will not focus on it it because it's not measured and reported 😡

This blog post medium.com/@BlueOptima/bl… seems to indicate that projects are "cost-efficient". If your organisation has this mindset, you've got bigger problems...

If you're worried about "cost-efficient" it probably means the software you're building isn't valuable enough. If this is the case, you've got a strategic problem where you should have probably gone with a SAAS or vendor product instead of building it yourself.

I can't really find enough about ART. Software Development Benchmarking seems to promise benchmarking productivity, quality and cost against other companies but doesn't describe its method or how. I'll challenge the why

A scrappy startup *shouldn't* be comparing themselves a FAANG company. If you're doing so, you're over-engineering. I'm sure these are all different even within and across FAANG companies because context matters.

Tools that promise managers a measure of *individual* developer productivity is the case of just because you build it doesn't mean you should

Also to close off this thread, if you haven't read the excellent #Accelerate book by @nicolefv et al, do so. These are the best measures backed by research for now and I'll refer to a 2013 article that I think still holds true today: martinfowler.com/articles/useOf…

s/right/write/g