In 2021, researchers from Microsoft, GitHub, and the University of Victoria published a paper that quietly changed how serious engineering organizations think about productivity. It's called the SPACE framework, and five years later, it's still the most useful mental model I've found for measuring developer work without destroying developer morale.[1]
I've implemented SPACE-inspired measurement at two companies. One was a 40-person startup. The other was a 600-engineer division of a public company. Both times, it uncovered productivity problems that purely metric-driven approaches had missed entirely.
S - Satisfaction and Well-being
How developers feel about their work, tools, and team. This isn't soft. It's predictive. A declining satisfaction score is a reason to ask what is making the work harder. Compare it with delivery and quality evidence before deciding whether output has changed.
P - Performance
The outcomes of developer work. Not outputs (lines of code, PRs merged) but outcomes (customer impact, reliability, quality). Did the code do what it was supposed to do? Did it work correctly? Did it make users' lives better?[2]
A - Activity
The countable actions: commits, PRs, deployments, code reviews, documentation updates. Activity metrics are the most dangerous dimension because they're the easiest to game. SPACE explicitly warns against using activity metrics in isolation. They only make sense in combination with the other four dimensions.
C - Communication and Collaboration
How effectively developers work together. This includes code review quality (not just speed), knowledge sharing, mentorship, and cross-team coordination. In my experience, this is the most frequently ignored dimension and the one that matters most as teams grow.
E - Efficiency and Flow
How smoothly work flows through the system. This captures interruptions, context switches, build times, CI wait times, and the ability to enter and maintain flow state. It's the dimension that resonates most with individual developers because it describes their lived experience.
Figure 1. An activity count gains meaning when read alongside the other dimensions.
# Why SPACE Works Better Than Single-Metric Approaches
The key insight of SPACE is that no single metric captures developer productivity. You need signals from multiple dimensions, and those signals should come from multiple sources: system data, surveys, and observation.[3]
Here's a concrete example. Team Alpha has great activity numbers. High commit frequency, lots of PRs merged, fast deployments. But their satisfaction scores are dropping, their code review comments are becoming superficial, and their change failure rate is creeping up.
What's happening? The team is under pressure to ship, so they're cutting corners. The activity metrics hide the problem. SPACE surfaces it because it forces you to look at all five dimensions together.
Team Beta has low activity numbers. Fewer PRs, fewer commits. But their satisfaction is high, their performance metrics are excellent (low bug count, high customer satisfaction), and their code review discussions are deep and substantive.
What's happening? The team is working on hard problems that require more thinking and less typing. They're producing fewer artifacts but higher-quality ones. A single-metric approach would flag them as underperforming. SPACE recognizes their contribution.
Measure everything for one quarter without taking action. You need baselines before you can identify trends. Resist the urge to "fix" things immediately.
The real value of SPACE emerges when dimensions disagree. Activity is high but satisfaction is dropping? You're burning people out. Efficiency looks great but performance is declining? Your fast pipeline is shipping buggy code. Communication scores are high but activity is low? Too many meetings, not enough building.
Figure 2. Conflicting signals help the team choose what to investigate.
Show the full SPACE dashboard to the team. Not just management. The team. Developers who understand what's being measured and why are more likely to engage with improvements. Developers who feel surveilled will game the numbers or leave.[4]
# The Contrarian Take: Satisfaction Is Your Leading Indicator
Most engineering leaders treat satisfaction as a lagging indicator, a consequence of everything else. I think it's the opposite. Satisfaction is the leading indicator that predicts everything else.
When developers are satisfied with their tools, team, and work, every other metric follows. They write better code. They review more carefully. They mentor juniors. They innovate. They stay.
When satisfaction drops, check whether workload, tooling or collaboration has changed. Activity can move in a different direction, so a rising commit count alone cannot settle what the team is experiencing.
If you only track one thing from SPACE, track satisfaction. It won't give you the full picture, but it'll give you the earliest warning.
Run this exercise quarterly. It takes 30 minutes of team time and gives you a SPACE health check.
Step 1 (5 min): Each developer rates each SPACE dimension 1-5 based on their experience over the past quarter. Anonymous. Paper or digital, doesn't matter.
Step 2 (10 min): Aggregate the scores. Show the team the averages. Any dimension below 3.0 is a discussion topic.
Step 3 (15 min): For each dimension below 3.0, the team identifies one specific, actionable improvement. Not "fix CI" but "reduce CI time from 18 minutes to under 10 by parallelizing the test suite."
Step 4 (next quarter): Measure the same dimensions. Did the interventions work? Celebrate wins. Escalate persistent problems.
This isn't a replacement for proper instrumentation. But it's a starting point that any team can implement in an afternoon. And I've seen it surface problems that sophisticated dashboards completely missed, because dashboards can't capture how frustrated a developer feels when the staging environment breaks for the third time this week.
SPACE isn't perfect. No framework is. But it's the best framework I've found for measuring developer productivity without making developers miserable in the process.
The alternative, picking a single metric and optimizing for it, always ends the same way: Goodhart's Law kicks in, the metric becomes the target, and it ceases to be useful. SPACE's multi-dimensional approach makes gaming much harder because improving one dimension at the expense of another is immediately visible.
Use SPACE. Implement it lightly. Trust your developers to care about the results. That's how you build teams that are both productive and happy to be there.
Frequently asked questions
What do the letters in SPACE mean?
SPACE covers satisfaction and well-being, performance, activity, communication and collaboration, and efficiency and flow. The dimensions describe different aspects of developer work, so a single activity count gives an incomplete view.
How is SPACE different from DORA?
DORA examines software delivery performance, including deployment throughput and instability. SPACE also considers developer satisfaction, collaboration and the experience of work. Teams can use delivery data within a broader SPACE discussion.
Does SPACE require a fixed set of metrics?
The post proposes a lightweight set of signals that a team can collect consistently. Its example dashboard and quarterly snapshot are implementation choices. Select measures that fit the questions your team needs to answer.
What should I do when SPACE dimensions disagree?
Examine the work behind the numbers with the team. High activity alongside falling satisfaction can prompt a discussion about pressure and work quality. Check the explanation against other evidence before choosing an intervention.
Your PR data reveals team bottlenecks invisible in standups. Here's how to measure TTFR, review distribution, and cycle time with the REVIEW framework.
Small frictions cost more than big blockers. 47 micro-frictions consume 96 minutes per engineer per day. Here's the TAX Return framework for finding and eliminating the developer productivity tax.
Set up continuous code quality monitoring with CI gates, trend metrics and alerts. Establish a baseline, define thresholds and review changes with your team.
Engineers spend 38% of their time on repetitive, automatable work. Here's a toil audit process that identifies where time disappears and a framework for systematic elimination with measurable ROI.
Developer experience is a systems design problem, not a culture problem. Platform engineering reduces cognitive load from 12 concepts to 3 for common tasks. Here's the LOAD assessment framework.
The 10x developer myth is actively harmful to engineering teams. After tracking individual metrics vs team output, I found that the highest performers often had negative correlations with team throughput.