It is a Tuesday afternoon in a quarterly planning meeting. The CFO has the headcount sheet open. Someone in the room, maybe the VP of engineering, says, "We've got eight Scrum Masters across the product org. What does that line actually buy us?" The room goes quiet for a beat too long. Somebody offers, "They run the agile ceremonies and help with delivery." The CFO nods, makes a note, and moves on. Three months later, four of those roles are gone.
That is how this decision usually happens. Not from research. Not from a delivery-performance review. From a moment in a budget meeting when nobody in the room could give a confident, specific answer to a fair question. In January 2023, Capital One eliminated 1,100 roles from what it called its Agile job family.
Their public statement was honest in a way most companies are not: the role had been critical to earlier phases of transformation, and the natural next step was to integrate agile delivery practices directly into core engineering. That captures what most leaders feel right now and cannot quite say out loud. The role helped, then something shifted, and now the question is whether the help is still earning its cost.
This piece is for the leaders making that call. Not a defense of the title; an honest accounting of what gets lost when you cut without thinking, and a checklist that separates two very different decisions: eliminating the title, and eliminating the function.
Why the cut looks safer than it is
When a Scrum Master is performing well, most of their work is invisible to leadership. That is the central problem. The function is preventive: removing impediments before they become misses, coaching teams through conflict before it erodes safety, surfacing systemic patterns to leaders who would otherwise see only the symptoms. And preventive work, by its nature, looks like nothing happened.
Cut a role whose value is preventive and you do not see the cost on day one. You see it three quarters later, when the things that used to get caught early are showing up as escaped defects, missed dependencies, and team health problems that have already calcified. By then, the connection back to the cut is too diffuse to defend.
Real talk. If your Scrum Masters were spending most of their time facilitating ceremonies and updating Jira boards, the cut probably will not hurt much, because the function was already mostly absent. That is not a reason to celebrate. It is a reason to ask why the role drifted, and whether you are about to make the same mistake with whatever replaces it.
What you actually lose when the function disappears
When an effective Scrum Master leaves and the function is not picked up by someone else, four specific costs show up.
Impediments stay unresolved. An impediment is anything outside the team's authority making their work harder than it should be: a licensing decision waiting on legal, a cross-team handoff that keeps failing because two managers have different definitions of "done," an infrastructure dependency that needs leadership escalation. Without the function, those impediments still exist. They just stop getting surfaced. Teams work around them, slow down, and stop talking about them. Lead time creeps. The connection back to the cut is hard to see.
Team conflict goes uncoached. Healthy teams surface tension early and resolve it in ways that strengthen the team. That doesn't happen by accident; someone is coaching it. Without that coaching, conflict goes underground. People stop disagreeing in meetings and start disagreeing in side conversations. Retrospectives become status reports because nobody trusts the room enough for honest feedback. Psychological safety is a delivery requirement, and it degrades quickly when nobody's job is to protect it.
Retrospectives lose their facilitator. A real retrospective inspects how the team works, decides what to change, and commits to one or two improvements. Without a skilled facilitator, you get something else: people read out what went well and what didn't, the manager nods, no decisions get made, and the improvement engine quietly seizes up. You can keep the calendar invite. The function is gone.
Leadership loses its read on systemic issues. An effective Scrum Master is a translator between the team and the organization. When two teams have overlapping accountabilities, when a policy is creating friction nobody intended, when a leader's behavior is signaling something they do not intend, the Scrum Master can name it without it sounding like complaining. Cut that channel and you do not stop having those problems. You just stop hearing about them until they show up in delivery data, attrition, or a failed audit.
The honest counter: when the cut is the right call
A Scrum Master who is mostly running ceremonies and maintaining Jira workflows is genuinely cuttable. AI is rapidly absorbing meeting facilitation and reporting. Engineering managers can pick up team-level coordination if they are properly resourced. The version of the role anchored in process custodianship was always vulnerable; the current downturn just made it visible.
The 2020 Scrum Guide saw this coming and tried to fix it. The language deliberately shifted from "servant-leader for the Scrum Team" to "Scrum Masters are true leaders who serve the Scrum Team and the larger organization." That change was not cosmetic. It was the Scrum community trying to reposition the role away from the version that was about to get commoditized. Most organizations missed it, never updated their job descriptions, and ended up with the version of the role that was being cut.
So the right question is not "Should I cut the Scrum Master?" It is: "Is the function this role was supposed to perform actually happening, and if I cut the title, who picks it up?" Sometimes the honest answer is nobody, because the function was never really there. Save the budget. Sometimes the honest answer is nobody, because we never noticed how much of it the Scrum Master was actually doing. In which case, the cut will cost you more than it saves, and you will not know it for nine months.
The "title or function?" checklist
Walk through these eight questions for the team in scope. The questions are not yes-or-no. They surface where the function actually lives in your organization right now, and where it would need to live after the cut. If you cannot give a confident, specific answer, that is the answer.
- Impediment removal. When a team-level impediment requires escalation, who currently owns getting it resolved, and who picks that up after the cut?
- Conflict coaching. When two team members are in unproductive conflict, who steps in to coach the conversation as a neutral party, not as the boss?
- Retrospective facilitation. Who currently makes sure retros produce specific, committed improvements rather than discussion that ends when the meeting ends?
- Systemic feedback channel. How do team-level systemic issues currently reach you, and what is the new channel after the cut?
- Outcome facilitation. Who currently helps the team translate "ship the feature" into "validate whether this feature changed user behavior"?
- Cross-team coaching. When two teams have a recurring handoff problem, who coaches them through resolution rather than just scheduling another meeting?
- Onboarding. Who currently coaches new joiners on team norms and how to operate effectively inside the team?
- Psychological safety. Who currently notices when a team's safety is degrading and does something about it before it shows up in attrition data?
Run the eight for one team. Honestly. If you can answer all eight with a specific person and a specific mechanism for the post-cut state, you are eliminating the title and the function is moving cleanly. If you cannot answer four or more, you are eliminating the function and calling it a budget decision. That is a different operating choice, and the consequences will show up in places that do not look connected to this decision.
Common traps
Assuming engineering managers can absorb the function without changing how they spend their time. An EM already running performance reviews, hiring, and one-on-ones cannot suddenly add team coaching, impediment removal, and retrospective facilitation without something else giving way. If you are moving the function to engineering management, decide what the EM is going to stop doing.
Assuming AI can replace the human coaching layer. AI can absorb scheduling, status compilation, and ceremony reminders. AI cannot read the room when two senior engineers are in a quiet conflict, and it cannot coach a leader through realizing their behavior is signaling something they do not intend. The same diagnostic applies here as in any role decision: are you eliminating an activity, or eliminating an outcome?
Cutting the role and quietly replacing it with three others, or skipping the conversation with your existing Scrum Masters. Delivery Lead. Iteration Manager. Agile Practice Lead. If the function was real, somebody has to do it; renaming without redefining is a marketing decision. And the practitioners in your organization right now have a clearer view of where the function is and is not happening than you do. The ones operating at the level the role requires will tell you what would actually break if their function disappears, if you ask them in a way that makes the answer safe.
Try this next week
Pick one team where the role is on the table for elimination, renaming, or absorption. Block 30 minutes. Walk through the eight checklist questions for that team. Write down a specific name and a specific mechanism for each answer, both for today and for the post-decision state. If you cannot name the person, the answer is that nobody is doing it.
Count where you have a confident answer for the post-decision state. Six or more, your plan is real; make the call with confidence. Four or fewer, you have your map. The next conversation is not whether to cut the title. It is what the function actually is, and where it should live going forward. That conversation will save you nine months of regret.
If you want a structured way to run this exercise across multiple teams, that is the kind of work I do in coaching engagements with leadership teams navigating role transitions. Sometimes the right answer is to cut, and the work is making sure the function moves cleanly. Sometimes the right answer is to redefine the role, and the work is helping the practitioners and the organization redraw the boundaries together.
Read Next
From Hero to Catalyst: Why Your Leadership Style is the Biggest Bottleneck
If the checklist surfaced gaps in how team-level visibility reaches you, this piece sits at the leadership end of the same conversation. About one in ten managers operate at the Catalyst level, where you stop being the hero and start designing systems that make heroism unnecessary. Worth a read before you make the call on the function this post is about.