Why Squash Merging Is Quietly Breaking Your Developer Productivity Metrics
Squash merge has become the dominant default for small-to-medium product teams in 2026, and for good reason, clean, linear history, simple reverts, no noisy WIP commits reaching main. Recent academic research specifically studying this practice found a real, measurable side effect most teams haven’t considered: squash-merging systematically erases the intermediate commit history that developer productivity tools and researchers rely on, making commit-count-based metrics genuinely meaningless once a month of real work collapses into a single commit.
Key takeaways
- Squash merging is a genuinely good default for most feature-branch, PR-based workflows Clean linear history, single-commit reverts, and no pressure to enforce commit hygiene on shared branches are real, documented benefits.
- The same practice erases data that engineering metrics and tools depend on Academic research specifically documents that commit-count-based productivity metrics and developer collaboration networks become meaningless once squash-merging collapses weeks of incremental work into one commit.
- Adoption of squash-merging is uneven across teams and projects, which biases any comparison Since some teams squash and others don’t, comparing raw commit-based metrics across teams compounds the distortion rather than averaging it out.
Why Squash Merging Won as the Default
The case for squash-and-merge as a default is genuinely strong for the feature-branch, pull-request workflow most teams use on GitHub or GitLab today. It removes the need for informal, hard-to-enforce commit hygiene rules on feature branches (no more "no WIP commits," "follow Conventional Commits format," "you must rebase before merging"), since only the final squash commit ever reaches main regardless of how messy the branch's history was. Reverting is also cleaner: one git revert undoes exactly one pull request's worth of change, where a merge-commit strategy leaves you deciding whether to revert the merge commit itself or something more granular.
Teams that adopt rebasing-heavy workflows specifically (rebase-before-merge, avoiding unnecessary merge commits) have reported productivity gains attributed to fewer merge conflicts and a more navigable history, with one industry source citing up to a 20% improvement, a figure worth treating as directional rather than a rigorously controlled academic finding, since it comes from industry reporting rather than a peer-reviewed study specifically isolating that variable.
Squash merging’s benefits for main-branch cleanliness are real and worth keeping. The specific thing worth addressing is what downstream tooling (dashboards, productivity metrics, code-archaeology tools like git bisect) silently loses access to once it happens, and whether your team is compensating for that loss anywhere.
What Recent Research Specifically Found
A 2026 empirical software engineering paper examining the foundations of software measurement specifically documents squash-merging as a threat to the validity of commit-based research and metrics, illustrating the difference between a traditional merge commit (which preserves every intermediate commit in a feature branch's actual development trajectory) and a squash merge (which condenses that entire trajectory into one commit, erasing every intermediate step). The paper's core finding is direct: developer productivity metrics based on commit counts become meaningless once a month of real, incremental work is squash-merged into a single commit, and developer collaboration networks built from co-change relationships between commits become sparser and less informative as a direct result, since they can no longer see the incremental changes developers actually worked through together.
The compounding problem the research specifically names: squash-merging adoption is neither universal nor random across projects and teams. Some teams and repositories squash consistently, others don't, and some mix strategies by branch or by contributor experience level, which means any analysis or dashboard comparing raw commit-based activity across teams isn't just missing some data uniformly, it's comparing genuinely incomparable measurement bases, systematically favoring or penalizing teams based on their merge strategy rather than their actual engineering output.
What to Actually Do About This
Practical responses, not a case against squash merging
The research finding applies most sharply where commit counts feed performance reviews or team comparisons directly.
The intermediate commits aren’t gone from the platform, even if they don’t reach main.
Standard merge preserves useful branch topology for some long-running integration scenarios; squash suits short-lived feature branches.
Uneven adoption within one team, not just across teams, compounds the measurement distortion.
Who Should Weight This Most Heavily
- Squash merging’s main-branch cleanliness benefits remain real and worth keeping
- PR-level intermediate history typically remains accessible even after a squash merge
- Switching to outcome-based engineering metrics fixes the underlying measurement problem directly
- Commit-count-based productivity dashboards and performance metrics are directly undermined by squash merging
- Uneven squash-merge adoption across teams compounds measurement distortion in cross-team comparisons
- Developer collaboration network analysis specifically loses fidelity as a documented side effect
Exploring the wider developer toolkit
See our full cloud and developer tools guide for Git hosting, CI/CD and infrastructure comparisons.
Our Sources
Where this comes from
The core research finding on squash-merging and commit-based metrics is drawn from a specific, dated 2026 empirical software engineering paper examining the foundations of software measurement, cross-checked against independent industry reporting on squash-merge and rebase workflow adoption and their reported productivity effects.
-
2026 academic paper cited directly
The core finding on squash-merging eroding commit-based metrics and developer network analysis drawn from a specific, dated empirical software engineering study.
-
Industry adoption and productivity claims flagged by source type
The ~20% rebase-workflow productivity figure is explicitly attributed to industry reporting, not the peer-reviewed research, and presented with that distinction intact.
-
No claims of our own repository analysis
This article synthesizes and cites published academic and industry sources; it does not present our own original commit-history analysis.
Frequently Asked Questions
Frequently asked questions
Does squash merging actually harm developer productivity?
No, the research specifically finds that squash merging harms the accuracy of commit-count-based productivity metrics and tools, not developers’ actual productivity itself. The practice’s main-branch benefits (clean history, simple reverts) remain genuinely valuable.
Should engineering teams stop using squash merge because of this research?
Not necessarily, the more direct fix is to stop relying on raw commit counts as a productivity signal, since that measurement approach is what the research shows breaks down, rather than abandoning a merge strategy with real, documented workflow benefits.
Is the intermediate commit history actually gone after a squash merge?
Not necessarily from the platform itself, pull request review history and the original branch’s commits typically remain accessible via the PR itself even after a squash merge, though they no longer appear in the main branch’s own commit log.
Why does uneven squash-merge adoption make the problem worse across teams?
Because it means commit-based metrics aren’t just missing data uniformly, they’re built on genuinely incomparable measurement bases across teams, systematically favoring or penalizing teams based on their merge strategy choice rather than their actual engineering output.
What should engineering leaders use instead of commit counts to measure productivity?
Outcome-based metrics, shipped features, resolved issues, pull request review turnaround time, are more robust to merge-strategy differences than raw commit or line counts, which the research shows can be rendered meaningless by squash-merging.
Final take
- Squash merge's main-branch cleanliness benefits are real and well-documented
- 2026 research shows commit-count-based metrics become meaningless once squash-merging is used
- Uneven adoption across teams compounds the measurement distortion in cross-team comparisons
Squash merging earned its position as the modern default for good reasons, clean history, simple reverts, no commit-hygiene enforcement burden, and none of that is in question. What 2026 research specifically adds is a documented, previously under-discussed cost: commit-count-based productivity metrics and developer collaboration analysis genuinely break down once squash-merging collapses real incremental work into single commits, and uneven adoption across teams compounds rather than averages out that distortion. The fix isn’t abandoning squash merge, it’s abandoning commit counts as a productivity proxy, regardless of which merge strategy your team uses.
