Estimated read time: 3 Minutes
Author: Dr. Manzoor Mohammed
Engineering teams normally don't care about cost - but they do care about great engineering and learning. The trick is getting engineering teams on board with a cloud optimisation programme by solving their problems.
A great strategy is to look at cost and performance in parallel from the start to win hearts and minds. The FinOps teams will be doing fantastic work on building the foundation of looking at cost by doing tagging, RIs, quick wins, etc...
While this is occurring you can follow these 3 steps to start getting your engineers on board for changing mindsets:
- Do a deep dive on end-to-end performance
- Understand engineers' challenges without talking about costs (yet)
- Help engineers fix stability issues and other performance problems
1. Do a deep dive on end-to-end performance
Deep dive on the product and architecture from a performance and capacity angle, with the mindset, that "great performance reduces cost". Looking at it with this mindset will allow you to bring insights to the engineering teams. A great example here was identifying unnecessary activity and unrealistic performance goals across the ecosystem & organisation (including AI/ML platforms). This helped save engineering time on keeping an aggressive performance goal in place.
This is the approach that we used when working with a multi-billion revenue San Francisco based tech firm. See more.
2. Understand engineers' challenges without talking about costs (yet)
Focus on understanding performance, how do the different parts from different teams come together? Who is responsible for each part?
Once you know this, find the insights and prepare the right questions. Then share the insights with the engineers to get them interested. Once you've engaged in a conversation with them, you'll be able to understand their challenges and language.
One example of this is from our work with a large telecommunications system that supported 400M+ users. We worked with their team to identify and remove a major code inefficiency that would have derailed a major integration. You now really understand the tech and the challenges the engineers face.
3. Help engineers fix stability issues and other performance problems
The next step is to find the engineering challenges that are causing stability issues. Dealing with these, removes excess consumption and frees up capacity, as well as getting the engineering teams on board with cost savings. By now you've gained the trust and confidence of the engineering teams. You haven’t talked about cost at all until now.
We engaged with a major SaaS company that was having ongoing customer-impacting incidents. We identified this was due to an imbalance in the CSPs platform. Up till that point, the engineering teams and CSP (Cloud Service Provider) had been blaming each other.
At this stage, you'll have built loads of goodwill with the engineering teams and they'll be ready to talk about costs. On part 2, we cover how to successfully reduce 50% of cloud costs while improving service stability by 95%.