Hacker News Reader

Comment by datascienced | Hacker News Reader

There are not many options. Headcount might help (secondment, overtime, contractors, move from another team, new hire) or pushing other work back (or cancelling it, reducing scope) or saying no (find a workaround, go without it). I might of missed it but there is only so much you can do. A win win might be showing how doing X now is not worth it because event Y makes X obsolete or less valuable.

View on HN

You can just have a queue. I’ve seen departments that basically function as an internal vendor of services to other departments. First in first out. Fixed schedule to expect turn around on most requests. No one complains. Seems quite pragmatic and fair.

Until they become the sole group that does X (Officially), and other groups realize that having them do a simple task related to X is way more hassles and months of wait time so each team just build their own unofficial X or try to circumvent it from day 0.

I'm not say this or that method is better just pointing out what I've seen so far through different companies. Not all work is the same priority but if you try to reprioritize then you can enter the endless quagmire of inter-department/group politics than involve a lot of useless meetings. And a simple FIFO queue can be equally as problematic.

I have seen it where they are literally the sole provider of x. Cheaper than third party x vendor even because of a lack of profit margin charging internally. Still their terms are straightforward and there’s no bullshitting.

FIFO is not an efficient way to prioritise work, though. Unless all work has about the same value and urgency, in which case, I guess, that sounds nice.

If you are an specialized department, insulated from the organizational goals, and circled by people that only speak in riddles, FIFO is the best you can do.

The context was in R and D at a large organization. Are some projects better funded/further along/seemingly more important to spearhead? Sure. But that was not how this department handled the workload. Dozens of people are coming to them a day for work done on hundreds of separate projects potentially. They can’t afford to triage this workload nor even can you really. Like I said, no one complained because despite FIFO turnaround was still faster (that afternoon if you got your job in during the morning potentially) and cheaper than any third party vendor.

Queue is fine but my experience is small company’s anyway hate them. A queue self manages, so how do you micromanage!

Sprints are an editable queue I guess but even waiting 2 weeks is too much for the nanomanagers.

Only works if you have a lot of funds or your funding model isn’t broken. Where I work it’s a big dance around funding IT to build or connect stuff for Ops/Engineering. Because it’s such a chore there are lots of alternatives (ticketing, middleware, event management/collection system etc.) or even shadow IT.

Reply on news.ycombinator.com