Blog posts like these are a staple on HN and I’m always mildly offended by them.

It seems to me that there’s a class of programmer that will take an overpaid job at a terrible BigCo, spend their evenings writing ranty blog posts about how terrible it all is, culminating in the inevitable “I quit” article, only subsequently to accept a job at a different terrible BigCo.

At no point does this article, or most articles like it, do any effort to realize that things are not always like this. Even if there is some unavoidable law that huge companies inevitably fill their ranks with idiots, like this article suggests, you do not need to work at a huge company.

Most people do not work at huge companies. There’s lots of amazing tech companies with fulfilling jobs and they want you, now. But there’s this super prevalent idea that keeps getting pumped around the blogosphere that it’s absolutely impossible to not work at a terrible huge company and therefore you cannot possibly escape, and I quote, “going home and despairing”.

There’s some weird kind of suffer porn going on here that’s just unnecessary. If you choose to trade your sanity for a high salary then go right ahead but don’t act like you had no choice.

> This is because Power BI tracks usage metrics and almost all dashboards are completely unused.

This is what I term the “dirty secret of BI” - for most organisations, most dashboards are unused, except by the teams creating them.

This is because reports need to be useful to be used, and most are not. In my opinion useful means one of:

1) Aiding people with actual decision making power to make decisions

2) Improving the ability of “front line” staff to complete their day to day tasks better

PowerBI and similar tools are poorly suited to the above because 1) requires mgmt to understand the data they are looking at, which is difficult without context in a grid of charts, and 2) requires a thoughtfully designed, fast UX that is basically impossible to deliver in a dashboard.

I’m now working on an open source BI tool, Evidence.dev (https://github.com/evidence-dev/evidence) which is targeted at these two use cases. It enables you to add context to data inline, design the UX for the use case, and is just fast.

Previous discussions on HN:

https://news.ycombinator.com/item?id=35645464 (97 comments)

https://news.ycombinator.com/item?id=28304781 (91 comments)