Hi All, I need to do a small exercise to quantify the ROI on implementing IDP for 2000 engineers. I am using Backstage already and I was wondering whether somebody has anecdocts, success stories or tried to to this exercise in the past to demonstrate the value of it. Thanks in advance!

Might need to be a bit more specific. Lots of variables to an Internal Developer Platform depending on what the organization needs / goals are. So just a blanket ROI calculation won’t be as clear cut

fair enough. My organization's need is to improve developer productivity and efficiency and I am building my northstar metrics and the story on how having an IDP can help. Meanwhile, I was wondering whether in this channel there are people who have already drafted and implemented a strategy to roll-out IDP (possibly by using Backstage) and what features you started to implement first and how you measured success. Looking forward to hearing back from you!

I'm also working on a similar initiative, but since we had to work very quickly for internal company goals, I opted for smaller effort initially: high impact items instead of inventing a whole new platform to develop. It actually helped me a lot! And it's much easier to come up with metrics for individual items like that. And then later you can roll these individual ones up into more global metrics. (Numbers below are arbitrary) Ex. Automate manual processes in the development process 1. Key Result: Automate/optimise the PR process a. KPI 1: PR cycle time is reduced by 15% b. KPI 2: PR weekly throughput is increased by 20% c. KPI 3: Average PR kickback rate is <1.5 2. Key Result: Faster build time via gradle conventions plugin a. KPI: Plugin adoption is 50% or more by EOY b. KPI: Build time reduced by XYZ c. KPI: Reduced effort in maintaining resources (ex. Spring Boot upgrades) d. KPI: Faster migration rate to new versions of libraries and resources across all microservices by XYZ

I assume you already did your internal assessment to see where most of the energy is spent in your development process regarding time, effort, re-work, and tooling costs (free is not free). At this point, you have an idea of where the current bottlenecks are happening and can tackle the first set of objectives. I see that once the initial bottlenecks are removed, the next set emerges, and it is not always the one you expected to be (since we’re talking about productivity and dev experience). Most teams I work with tackle the environments bottlenecks first. They build a system for isolated dev, test, and staging that mirrors production and is available on-demand to all developers without risking breaking prod. This becomes the backbone of their self-service IDP. Reliable and easily available environments are usually the major frustration for devs who want to actually code and not wait or configure resources. Obviously, I’m speaking from my own soapbox, so your ROI calculations must come from your developer’s needs and wants. Would love to hear how both of your projects went! @Silvia @Olga

i'll take a step back. is calculating ROI here the right approach for this? what's the actual outcome desired internally? input into a decision? something else?

@Amir Tayabali might have some thoughts on this!

is time spent on on-call incidents a part of any kpi?