This is a Twitter thread I posted on November 13 as a reaction to a prompt from Dan Luu, captured here for posterity.
Some advice for line managers to help avoid similar situations (project getting cancelled). I'm not at all familiar with the referenced project, nor any contributing organisational dysfunction, just using it as a launching point. 🧵 https://t.co/BIc7x614NS— Xavier Shay (@xshay) November 13, 2022
Some advice for line managers to help avoid similar situations (project getting cancelled). I’m not at all familiar with the referenced project, nor any contributing organisational dysfunction, just using it as a launching point.
Projects get cancelled when the confidence execs have in their expected net benefits are lower than other projects they could fund for similar risk, and/or they need to lower their aggregated risk. (You can only have so many “bold bets”.)
If your project is long running with back-loaded benefits (“pay off a year down the line”): you are a high risk project. This is relatively independent of the projected benefits.
Software projects (in particular) are notorious for having both high variance costs and benefits. Your job as a line manager is to reduce variance and increase confidence over time that the project will be successful.
Your two main tools to do this are sequencing and storytelling – and they’re related.
Sequencing: front load work that reduces risk and increases confidence the project will succeed. More useful framing than typical agile “ship to production”, which is a subset of useful activities and often hard to apply to infra projects.
Example goals that reduce risk but aren’t shipping software: share a design doc, build a prototype, get clear instrumentation around the thing that needs to change.
(Sidebar: some amount of planning is good, but for most projects exhaustive planning only gives the illusion of variance reduction … for a lot of cost.)
Your team is likely doing a decent job of sequencing implicitly. The more common problem is that this is invisible, and so perceived risk and variance are unchanged. And with no feedback loop, there’s no way to go from “decent” to “great”.
Hence: storytelling. You need to explain why your sequencing reduces risk, and ensure you have artefacts to show it (the document, the graph, etc…)
Effective storytelling requires repetition. Every single time you present: reinforce what the project is doing, why it’s important for what the audience cares about, how far along it is, what has changed since last time.
Storytelling isn’t just for managing up. It’s a useful feedback loop - if you can’t tell the story, maybe your sequencing could be improved? It also improves morale on your team: a series of intermediate wins feels better than a long grind.
Those are the main points I wanted to make, here are some other quickfire observations:
You probably underestimate the scale and potential impact of short term pressures on the business. (Senior leaders probably overestimate it, but that’s not a useful insight for you.)
“Execs forget” isn’t true. They have a higher bar for clarity than you are providing them.
Storytelling is a manifestation of your reality distortion field. Developing as a manager means getting better at perceiving and tuning this field. (This deserves its own thread I think.)
Despite all the above, your project might still get cancelled. Them’s the breaks sometimes. Not much you can do to avoid getting hit by a meteor. (Thanks to Maya Hope for this analogy and for feedback on this whole thread.)
All easy to say, hard to do. I’ve done it wrong a lot and will continue to do so. If you’re on my current team, please help me identify gaps between my words and my behaviours 🙏
I liked this reply and example from Py.
Great thread! (and also ♥️ the quoted thread from @danluu)— @firstname.lastname@example.org (@Piwai) November 13, 2022
Sequencing to deliver immediate value even if the path looks longer overall is also how you build the right thing.
This approach was crucial to how we shipped the Square Reader SDK, let me tell you how 🧵 https://t.co/gOngMKswqq