The 2024 State of Platform Engineering? Fledgling at Best
It’s a hard road ahead for platform engineers, as few organizations have achieved true platform engineering — though how could they when they aren’t measuring its results?
But if platform engineers continue to earn at least a salary grade higher than DevOps engineers, they will likely remain up to the challenge.
These are just two of the biggest reveals from this year’s “State of Platform Engineering” report.
The New Stack sat down with Sam Barlien, head of community at Platform Engineering, ahead of the release of this year’s report. He reflected on why organizations are falling short of long-term investment and strategies to unlock platform maturity, including the role of internal developer platforms (IDPs) in a DevOps organization and, perhaps most importantly, improvement, why to measure it all.
Platform Engineering: Still a Shiny New Thing
This was the third annual report, sponsored by Humanitec and Gitpod; researchers surveyed 281 platform engineering professionals.
The majority of the organizations surveyed — 56% — have had platform teams for less than two years. A mere 13% of respondents reported working in “platform engineering” for more than five years.
Why the sudden uptick? What drives organizations to adopt platform engineering?
The researchers, Barlien said, settled on the definition of platform engineering as, "the discipline of designing and building toolchains and workflows that enable self-service capabilities for software engineering organizations in the cloud-native era. Platform engineers provide an integrated product most often referred to as an internal developer platform covering the operational necessities of the entire lifecycle of an application
This is intentionally broad enough, he said, to include data platform engineering, but it hones in on platform roles serving internal customers, who are overwhelmingly developers.
First Comes DevOps, Then Comes IDPs
This year’s results once again argue that platform engineering isn’t replacing DevOps, but rather, it’s just the next stage in digital transformation evolution. A platform helps engineering organizations move closer to achieving the promise of DevOps.
Overall, respondents said that an internal developer platform is being built in their organizations to pick up the slack where DevOps has faltered. Automation is a key DevOps principle, yet nearly half of all survey respondents complained of doing too many repeat tasks due to a lack of automation.
Another DevOps principle is a drive for responsibility throughout the lifecycle — you build it, you own it. But the average tech stack has grown exponentially, often seven layers thick. This distracts the developer from actually delivering value to customers. An IDP abstracts away cloud native complexity to reduce developer burnout and cognitive load. It can also further bring products, processes and people together.
Of course, 13% of respondents admitted they also relabeled their DevOps team's platform engineering “to sound more fancy,” so Gartner may be right that this discipline has slipped into the “trough of disillusionment.”
This year’s “Accelerate State of DevOps” — more commonly known as the DORA report — found discouraging results. For instance, the DORA researchers found that, with platform engineering, throughput decreases by 8% and change stability decreases by 14%.
Since this was the first year the DORA team members looked to measure platform engineering, they didn’t claim to have answers — though they posited that speed to market could simply mean more changes are released overall, which could decrease overall stability.
From DevOps to Platform Engineer
This annual report uniquely shows platform engineering as an evolution of DevOps in terms of job roles.
The report positions infrastructure platform engineers and DevEx platform engineers above infrastructure, operations and development roles in an employment pyramid. In comparison with DevOps roles, there are fewer of these platform “value drivers” in charge of easing the base’s pain with standardization and automation.
(Source: "State of Platform Engineering Report, Vol. 3," Gitpod and Humanitec)
With that in mind, in 2024, platform engineers in Europe earned 23% more than DevOps roles and 27% in North America, according to the report.
While the differences between platform engineer and DevOps roles were murky back in 2022, this report shows that, in 2024, platform teams are clearer about what is — and isn’t —under their jurisdiction. Platform engineers told researchers they are in charge of:
Only a small fraction of responders felt that, as platform engineers, it was their responsibility to act as a developer help desk (19%) or to create dashboards for executives (9%).
The salary differential between platform and DevOps engineers has halved since last year’s report, which had platform engineers earning 42% more than DevOps engineers. This could be a sign again that platform engineering is stll important, but not as much as perhaps the industry originally thought.
“Whether that's a closing gap because of big layoffs,” Barlien posited, or “whether it's a closing gap because DevOps salaries went up and platform engineering salaries went down, [it’s] not sure.”
What the report has consistently measured over the last three years is that someone with a platform engineer job title has years more experience than someone with a DevOps job title. In fact, 28% of platform engineers surveyed this year had more than 15 years of experience in tech.
How Mature Are Platform Teams?
Just over a year ago, the Cloud Native Computing Foundation released its Platform Engineering Maturity Model.
This is a framework to measure platform engineering initiatives and to identify areas for improvement, across:
- Investment.
- Adoption.
- Interfaces.
- Operations.
- Measurement.
These aspects are quantified from least to most mature as:
- Provisional.
- Operational.
- Scalable.
- Optimizing.
Last year’s “State of Platform Engineering Vol. 2” found that platform engineering organizations were so nascent that they had yet to put in place best practices including:
- Change management processes.
- Cross-organizational buy-in.
- A clear platform team structure.
- A Platform as a Product mindset.
- A platform adoption strategy.
As platform engineering enters its third year as a trending tech topic, this year’s report looked to quantify its maturity.
In short, most teams are far from mature.
“The thing reflected here is for the people who are doing platform engineering well and it's working for them, it's spectacular,” Barlien said. “And the people who are not doing it well, it's the complete opposite of spectacular.”
Only about 9% of respondents were indeed mature by the CNCF Platform Maturity Model’s standards. In contrast to those few, Barlien and his team saw an overall lack of clarity as to what platform engineering even is — from those who reportedly have roles as platform engineers.
“A lot of the questions, people filled ‘other’ and their answer was ‘I don't understand this,’” he said, of what he thought were straightforward platform engineering questions.
While many companies are embracing platform engineering, the report read, only a few are achieving potential in a mature and scalable way.
But, if the average team interviewed is between zero and 24 months old, this shouldn’t be that surprising.
Platform Teams Don’t Know How to Measure
This year’s “State of Platform Engineering” did reveal the loudest secret in engineering or any science: you can’t improve what you don’t measure.
About 43% of platform teams surveyed self-described their feedback mechanisms as “ad-hoc" or “inconsistent.” That’s a long way from the CNCF’s standard of quantitative and qualitative feedback.
“People have no idea what they're doing,” Barlien said. “And that’s kind of the main takeaway.”
In fact, 45% of platform teams don’t measure anything at all, while 37% just measure DORA metrics. While DORA’s focus on throughput and stability are important aspects of the developer experience, they certainly don’t paint a detailed picture of DevEx.
“When you don't measure, then you don't know whether you are actually making a difference or not making a difference,” Barlien said, “and then all of a sudden, six months later, you do a check-in, everything feels worse.”
Another 26% of survey participants reported collecting data but not actually analyzing it.
When asked how metrics have improved since introducing platform engineering, only 22% reported significant improvements, while 32% saw slight gains. In contrast, 17% reported no noticeable change, while 27% were uncertain.
That doesn’t just hurt platform teams’ motivation — it makes it hard to prove that they’ve earned their budget, and deserve to keep their jobs.
The report also pointed out that 39% of teams reported that their platform is “centrally enabled with a focus on user needs.” But if they aren’t measuring for that, that description could be based on assumptions: “The gap is an indication that many organizations have a much better image of themselves than the reality.”
Add to that, there’s an almost 18% discrepancy between those organizations who admit they don't know to what extent metrics changed since they started platform engineering — 27% — compared to those who answered that they do not measure at all — 44%. This suggests that 18% of respondents may be assessing things based on anecdotal evidence, informal observations or ad hoc measuring.
“People simultaneously say, ‘Oh yeah, everything's amazing.’ And then they say, ‘Oh no, we don't measure anything,’” Barlien reflected. “It's not surprising when you look at something like DORA — instability and cognitive load are going up.
“Well, if these people are not measuring anything, they're not clear on what they're doing. Then it's understandable why some of these things aren't working the way they might want.”
What’s Missing? A Product Mindset
“Platform engineering is this step between product management and something staunchly technical and engineering-focused,” Barlien said. But, he added, many platform teams still don’t understand this.
Most platform teams still lack product management understanding, he said, including:
- User research.
- Measuring for improvement.
- Understanding what customers (your developers) want.
- Understanding if your developers are using your platform.
- How they are using it.
“Unlike with something like DevOps or agile, where it takes the same kind of headspace that an engineer is already familiar with,” Barlien reflected, “platform engineering is often asking an engineer to enter a completely new headspace.”
Without using the exact term, this report points to a Platform as a Product mindset as the main differentiator between the scarce respondents who are doing platform engineering well and the vast majority who are falling short.
What about you? Do you have a story to share on your platform engineering journey? The call for speakers for PlatformCon is open.
