elmarMar 18
Ruth — of systems & design

“1. You don't know what to build

This is the one nobody wants to talk about because it's embarrassing. Your PM hasn't talked to a real user in two months. Your requirements arrive as a Jira ticket with three sentences and a Figma link to a design that was approved by someone who's never used the product. Your engineers are making fifty micro-decisions a day about behaviour, edge cases, and error handling that nobody specified, because nobody thought about them.”

https://andrewmurphy.io/blog/if-you-thought-the-speed-of-writing-code-was-your-problem-you-have-bigger-problems

If you thought the speed of writing code was your problem - you have bigger problems | Debugging Leadership

AI coding tools are optimising the wrong thing and nobody wants to hear it. Writing code was already fast. The bottleneck is everything else: unclear requirements, review queues, terrified deploy cultures, and an org chart that needs six meetings to decide what colour the button should be.

Debugging Leadership
Mar 17 at 8:24pmWeb
Show threadDawn AhukannaMar 17
@RuthMalan “You don’t know what to build …” because it has to be figured out by the team and co-created with the target audience. This is why experience (interaction and interface) design is not about delivering visual static frames snapshots (I call them Polaroids). It’s about providing a solution to enable entire workflows and tasks.
Show threadMay Likes TorontoMar 17
@RuthMalan 2 months? That's child's play.
Show threadMay Likes TorontoMar 17
@RuthMalan Some PMs will have you know they've never talked to a customer.
Show threadGerardMar 17

@mayintoronto @RuthMalan Yeah, a place I worked at had feature requests that sat on the backlog for years.

The backlog tool was an internally developed system and I spent one hackathon seeing what lessons I could learn from all our historic data. Especially interesting was a graph I made: I drew one line for each major feature request. The x axis was calendar dates, the y axis was our ETA for that feature as calculated on each date - that is, how long it would take to complete, including the time to complete all the higher priority features before it.

The interesting result this graph demonstrated was that once the ETA for a feature was more than about 2 years in the future, it usually became longer rather than shorter as time went on. There was a kind of event horizon visible in the graph, beyond which planning was meaningless.
The honest and courageous thing to do would have been to remove everything beyond 2 years in the future from the road map. I leave it as an exercise for the reader to guess whether that was what we actually did...

(Incidentally, I believe the 2 year threshold was specific to that product and sector, and may be at the longer end of the curve. I would anticipate that this planning horizon would change according to things like the regulatory system, the type of customers, the rate of change in the market, and so on.)

Show threadMay Likes TorontoMar 17

@GerardThornley @RuthMalan Currently doing a tour of the go to market exec team to give them the news that I'm killing feature requests, to figure out where the objections will be.

It's been easier than I thought. I promised to not just delete the whole backlog this time around.

Show threadSébastien BarbieriMar 18
@RuthMalan Goldratt "The Goal" of course!
Show threadSébastien BarbieriMar 18
@RuthMalan just finished the whole article. I'm just living that right know in my new job.
So I think I will read Goldratt again. And maybe Gene Kim...
I told my team... Is this real? Are you OK with this?
Thanks for the share!
Show threadTim SwanMar 18

@RuthMalan
Stop spying on me!

(yesterday had so many micro decisions it turned into a macro decision).

What's the fix for the badly scoped requirements and misaligned priorities? Because optimizing the right things doesn't fix those.

Show threadKay Ohtie 🔜 FWA6d ago

@RuthMalan Bookmarking this to share with my boss. It's not something he's unaware of, but it puts into concrete form a frustration he and I have shared with recent projects being sent to me to "pls push to prod".

Tickets coming in from an office that rarely deals with students that only mention what to do by naming the page code name in the application platform (some awful 7 letter indecipherable thing). Pages built that can't be tested by developers because, and styled to look exactly like the mid 2000s pages they were asked to remake.

And no thought given to "how can we streamline this for students".