How much does it cost? How decentralised cost ownership typifies the agile mindset
This was first published as part of the Velocity Agile newsletter. To subscribe for more articles like this, straight to your inbox, click here: Velocity Agile newsletter.
How much did it cost to build your product? What’s the unit economics of your product? What once may have been straight forward questions in the manufacture of physical products, has become significantly more opaque in the case of digital products. Many of the costs are shared, and therefore often the per transaction cost is small, hard to attribute and variable.
The practice of FinOps is one of a growing number of initiatives that have arisen to help shine a light on the costs of running each digital product. This program guides how to decentralise cloud costs to the relevant domain responsible. The open standard provides guidance on how best to do the attribution, and then how to use this information to inform decision making within teams.
As cloud has become thoroughly embedded into development culture, this has become more of an issue. AWS, Azure, GCP and others have increased costs over time, and the total bill can now be eye-watering, and a major operating expenditure for companies. There is even the suggestion from a16z that for some companies, it may be cheaper to repatriate and bring infrastructure back on premise. Full circle!
Cloud cost optimisation has become an industry in itself, with roles being some of the best paid in the industry. It’s popularity with executives is clear due to the obvious ROI of investing in this optimisation to the bottom line. For example, Softcat helped the NHS digital save 25% of their monthly expenditure through optimisation.
In an agile context, FinOps creates a feedback loop that can help the Product Manager and team make better decisions about the product. Bringing the costs of the product into the decision making can lead to more fiscal discipline in the development of products. For example, this may help to prioritise architecture or technical changes that help to reduce cloud costs over new features.
Which begs the question, what other costs can we provide automated data and insights to teams that can help drive better decision making?
Another area of interest may be salary costs, the majority cost of teams. Are you aware of how much your product costs to make, and what the monthly burn rate is? This is pretty rare for the team to know due to the sensitivity around salaries. This can be exasperated if the product and engineering teams are separated.
Similarly, are you aware and factor in the support cost associated with your product? Often, once development is completed, we assume a pooled cost of support operations who help to maintain the product. If this cost is high due to poor quality in the product, this is not factored into the product. This may impact the prioritisation of improving the existing product versus developing new features.
In a world where companies are being asked to focus on profitability rather than growth, helping teams to know their costs is a key component. FinOps offers an example of how agile teams can understand their costs as well as their revenues.
References
https://www.finops.org/stories/when-do-cloud-savings-kick-in/
