FinOps practitioners come from organizations of all sizes and from a variety of roles. This series aims to highlight our community members so we can learn a bit more from their perspectives.
We asked Tim a few questions about his FinOps and cloud finance management journey and here are a few insightful answers that might help members both new and experienced. Let’s kick it off with a quote:
“Increased cost isn’t inherently bad, and decreased cost isn’t inherently good. It’s cost relative to the value generated that counts. If you can’t tell what the value change was, you can’t make good decisions about cost.”
What is your title and how do you describe what you actually do?
I’m Tim Prentice, the Cloud Service Delivery Manager at Air New Zealand. My role reports to the Head of Infrastructure, who reports to our CTO.
I look after the cost and governance aspects of our cloud footprint. This includes our partnership with our main cloud vendor (being AWS) and all of our internal reporting and tracking of costs on the platform. There’s also a lot of work around internal client engagement and partnership between our internal customer and AWS.
How did you become a Cloud Service Delivery Manager?
I actually started out as a Microsoft Licensing specialist. I made this shift as I saw that public cloud was going to involve the same level of complexity (or “comprehensive suite of options” as Microsoft used to call it) and a need for focus on the connection between costs, incentives and business outcomes.
What does success look like in your FinOps role?
Success to me is mostly that our spend is predictable within any given month. That doesn’t mean stable or does not fluctuate though. It means that our teams have considered the implications of the things they are building and the changes they are making and have communicated to the business the impact they will have on costs.
It’s the communication and cost awareness that drives the focus on efficiency and value generation, and our ability to optimize return on investment (ROI) with constructs like savings plans and RI’s. It also has the added advantage of making cost that is probably not generating value very easy to spot, as no one has told us it was happening and it wasn’t a part of a plan.
Why did you join the FinOps Foundation?
The small community of FinOps practitioners in NZ started to talk about the FinOps community as a great place to hear what other companies were doing in the FinOps space and the founders were very approachable. As so little of this practice had the kind of industry/community support the technical aspects of cloud did, it was a great opportunity to validate our thinking and to give back to the community where we could, especially in New Zealand.
What are the most important KPIs that you track?
One important KPI is to forecast actual month-to-month cloud spending (each team can adjust their forecast at any time), coverage of RDS Reserved Instances, and performance of AWS Savings Plans (SPs). We also track a range of wastage metrics. We’re also tracking how well teams engage in advance of changes rather than post.
Where do you report in the organization? CIO? CTO? CFO?
I report into the head of infrastructure who reports to our CTO.
How has a failure, or apparent failure, set you up for later success?
Our initial focus on Tagging was probably our biggest wasted effort. It seems logical to get teams to tag everything and there was some initial enthusiasm for it. However, we quickly learned that global tagging structures were not generating much value for us or our reporting line and was becoming a real point of friction between the CCOE and the teams developing on AWS.
We have switched to an account based system for high level cost and resource buckets which is far simpler. This also lets teams create tagging systems that help them track whatever they need to locally, making the effort for them to do that work sensible and not in competition with their objectives. All w e need to do is make sure that the questions we are going to ask about cost are well articulated, then we let resource owners consider how they will answer them over time.
What is the makeup of your team? Which roles and skills?
We have a platform owner who focuses on technical alignment and product ownership, myself as the service delivery manager with the focus mentioned above, a security focused engineering resources then an outsource engineering team of 4-5 FTE depending on workload.
What is your biggest CFM challenge?
Integrating consumption costs into our larger plan, budget, execute structure of financing our IT operation. It’s been a slow process and we are making good progress but it is a very different paradigm for an airline.
What is the book (or books) you’ve given most as a gift, and why? Or what are one to three books that have greatly influenced your practice?
The most common book I handed to people was Consumption Economics by J.B. Wood, Todd Hewlin and Thomas Lah. It was a really powerful reframing of IT for my outsource partners and it holds up surprisingly well given its age.
Superforecasting by Philip Tetlock and Dan Gardner had a huge impact on how we think about the innately unpredictable nature of cloud and how to focus on just what’s valuable to make good decisions.
Accelerate: Building and Scaling High Performing Technology Organizations by Nicole Forsgren, Jez Humble and Gene Kim provided a grounding in how to think scientifically about all hype and enthusiasm from other books, blogs and podcasts and turn it into valid experiments to find what really works.
Where have you had the most success?
Empowering teams (with cloud financial management tools and best practices) to predict what they were going to do and letting that tell us as a central FinOps team if they were being successful.
While it was initially just a short term solution to a resource problem, it’s become a framework for communication that integrates very well into many disparate workflows. It also provides confidence to finance that we have the estate in hand which makes investing in commercial constructs like RI’s and SP’s much easier.
What are bad recommendations you hear in your profession or area of expertise?
One common piece of bad advice that I’ve seen is that you can rightsize based on data provided by tools that scan your environment. Usually these are presented as “you can save $XXX/month” and handed to someone senior with little regard for the cost of the activity and the unseen constraints on an environment.
Another piece of bad advice is reporting for reporting sake e.g. you should have X report going to X person every X. Just because you can produce a report doesn’t mean you should. We start from the questions being asked of us by teams, finance or leadership and dig into the why before we start generating data. Then track if the reports are being interacted with and regularly cull any reporting that isn’t producing engagement. We also tend to prefer reports that happen in response to some valuable metric, rather than at an arbitrary time.
What’s your elevator pitch for what you do to the c-suite?
We make sure teams can account for the cost they spend against the value they produce for the airline. We then optimise that spend and our relationship with AWS through negotiation and the use of commercial agreements like RI’s and SP’s.
What is one of the best or most worthwhile investments you’ve ever made? (Could be an investment of money, time, energy, etc.)
I don’t have much of a pithy answer here. It’s mostly been moving more to enabling teams by telling them the questions we need to answer to be successful, rather than enforcing reporting or tagging standards.
Do you do chargeback, showback, shameback or other?
We use showback augmented by prediction. E.g this is how you performed against your prediction.
What have you found is the most effective way to get teams to take action on improving cost efficiency?
Give them a metric that they know is being monitored and has an impact. Their predictions each month for a part of the budget for AWS, if they fall outside of this they might end up needing to do a business case for more funding etc.. This results in teams starting with a push for internal efficiency first before requesting more funding.
What is one thing you wish you could go back to the beginning of your journey and do differently?
If I could go back to the beginning of my FinOps journey, I’d use AWS accounts as our primary segregation of costs. We have some accounts that aggregate services and they are our main pain points for ownership and accountability.
How do you evangelize your work internally?
To finance and our leadership team it’s usually focused on cost savings and governance adherence through consensus rather than authority. To the teams its freedom to do what they need to without onerous reporting requirements, we keep it low touch but maintain focus on what needs to be managed effectively.
If you read the Cloud FinOps book, which chapter did you find most valuable to your practice?
As a regular part of the FinOps Community, it was mostly additional validation. It’s a great book for getting people to understand what’s involved and really helps people getting started in this space.
What tools do you use and for what?
We use CloudHealth for cost and asset reporting, Cloud Conformity for Security monitoring, and Gorilla Stack for scheduling. We are in the process of doing a market review on DLP solutions.
What advice would you give to a smart, driven college student interested in this area? What advice should they ignore?
The community has the greatest stores of knowledge. Getting involved in the FInOps Foundation, local meetups, and conferences are actually really important to getting a good understanding of the core concepts and the variation from business to business. Past that, I would recommend making sure you engage with the financial experts early. Understanding how to communicate with financial teams is critical to this role and will serve you well.
If you could have a gigantic FinOps billboard anywhere with anything on it what would it say and why?
Increased cost isn’t inherently bad, and decreased cost isn’t inherently good. It’s cost relative to the value generated that counts. If you can’t tell what the value change was, you can’t make good decisions about cost.
Inspiring more people to learn about FinOps
Huge thanks to Tim for taking the time to chat with us and provide insight into his role as Cloud Service Delivery Manager, its obligations, and how he implements FinOps best practices and tracks its success.
Keep on breaking down those silos!