4 ms·
One startup, our MVP launch date (very hard date; miss it, and the company is over) was moved to earlier, by customer's manufacturing schedule change, just as h
by neilv 4y ago
One startup, our MVP launch date (very hard date; miss it, and the company is over) was moved to earlier, by customer's manufacturing schedule change, just as half our small, high-powered engineering team got sick.
So I took the existing plan, from before the latest reality happened, and:
* Turned plan into refreshed sorted spreadsheet rows of prioritized tasks and estimated durations and assignments.
* Culled tasks very heavily and sometimes painfully.
* Kept in mind that the system had to work immediately at launch, and had to keep working, with production line uptime and correctness. (For example, there were some non-obvious system robustness/resilience tasks that I added, even as I was removing most other tasks.)
* Looked for opportunities to do things smarter, from the remaining essential tasks.
* Made sure the sick people weren't assigned to anything that absolutely had to happen.
* Kept updating the plan, sometimes multiple times a day, as tasks were completed, problems found, etc.
* Made sure that the estimates always said we'd hit deadline.
All this planning was rapid and intense, not "let's take a day to plan, and schedule some meetings, and have regular check-ins, etc.", and captured in a single view (the spreadsheet).
It was successful, ended up with 100% software uptime and correctness, for the entire contract period, of over a year. Some of that was because the technical cofounder who'd done all the initial work had experience with critical systems, and I followed through in the same spirit. Some of it was luck. Some of is was being vicious and smart about what work was done. Some of it was misery, knowing that the company, and a lot of people's livelihoods and career goals, relied on this coming together in time and working.
As for how to shield people from ulcers in a crunch: if anyone has to get ulcers, it should be leadership, quietly, from figuring out how to make this come together successfully, without the entire team exploding with ulcers.
And the team should see that leadership is all-in on making this be a success, and every individual implicitly buys in on staying committed to the team and project, and the team rises to the smart things that they are asked to do.
(Note: In this case, I used a spreadsheet, but in other cases, the single source of truth view for the entire project would more likely be a good Gantt diagram, or maybe a Kanban board. What's important are that it always reflects the best information and current decisions&activity, everyone is working from it and understands their parts in context rather than as checkoff tasks out of context, and it shows the project coming together.)
- BurningFrog 4y ago> * Turned plan into refreshed sorted spreadsheet rows of prioritized tasks and estimated durations and assignments. > > * Culled tasks very heavily and sometimes painfully. This is how any software project should run, continuously.
- orblivion 4y agoAs a side note, why is it that whatever Very Smart organizational system using Trello or Asana that we have set up in our own lives, or in our company, (and now I have your company as a second example), when shit gets real it comes down to the "real" list which is a spreadsheet or in my personal case a text file? What does that say about our Very Smart organizational system?
- CharlesW 4y agoI think it says "don't get religious/superstitious about tools". Many can work, but they also don't replace project management people.
- neilv 4y agoIMHO: 1. Some things done in everyday practice have big inefficiencies, or give the illusion of productivity and progress, and desperation can force a team to realize that. 2. A lot of tools are driven by enterprise sales, which means they target needs of large companies with very different needs than a startup-like team (in which it's easier to have everyone on the same page, aligned, and barriers removed). 3. A lot of tools are driven by enterprise sales, which can mean that the people who select and approve them aren't the people with experience using them. (And sometimes a team is forced to use poor tools, with adverse impact on effectiveness and morale, and then you get into a budget meeting, and see the SaaS fees the company is paying for those tears (TaaS).) 4. There are definitely projects I can't handle in a spreadsheet (most of them). But, when SHTF, and priorities change, the number of things you're focused on gets smaller and more in everyone's heads, the spreadsheet or text file might be most rapid to work with. (For SHTF deadline-hitting in another, much larger, startup, I would've forced everything into a canonical Gantt chart, just because the interdependencies were too complex, and declared the tons of Jira sprint Issues and manual team reports to be confusing and time-wasting noise, and everyone works&reports first from the Gantt.)