FinOps Engineering — DX

FinOps Engineering— Developer experience, developer efficiency, value stream efficiency.

As we try to seriously consider and monitor the cost implications of our actions in product teams it is important to always remember that the purpose of FinOps is to get the most value out of cloud computing. It isn’t to simply reduce or minimise costs.

Consider that the best thing for any team to do from a pure cost perspective would be to all quit and go home. But then there is no potential for transformation, no excess value, no profit.

Often times by far the most expensive cost in an organisation is people. Labour. So it’s critically important that they can make the best use of their time and be as efficient as possible.

Due to how consumption is charged, current data available and transparent (albeit often confusing) pricing information we can work out approximately what any optimisation will save in advance. We can also fairly accurately work out the cost of labour.

Quick search says UK average software engineer salary is around £60,000. They work around 225 days per year (accounting for weekends, expected leave, bank holiday, sickness) for around 7 hours per day (accounting for breaks). That’s around £38 per hour. Blocking a team of 8 for just 1 hour would cost around £304.

This is obviously imperfect because people simply aren’t productive for all hours of a work day and any given cost-optimisation is unlikely to wipe out all productivity for a given hour — people can often pivot to something else but the context switch has an additional cost — and so on. But this is the way to properly think about these things.

This also goes for any systemic inefficiency in terms of value streams i.e all of the activities for you to release value to a user. For example if you find a bottleneck then first do no evil. Don’t slow it down for any reason. On the other hand. Bias towards reducing toil and labour in value streams even at additional expense.

It’s rare that a local cost-optimisation that lowers productivity is worth it. Of course there will always be edge cases but heuristically we should lean on the side of productivity.