Scrum defines no tester, architect or team-lead accountability — how do you keep that expertise owned?
answer
- The framework names three and stops
- A seat buys clarity and a handoff
- Specialists are Developers in this vocabulary
- Watch where finished work queues up
basics
~20 sScrum names only Product Owner, Scrum Master and Developers, so a tester or architect is a Developer in its vocabulary. The expertise stays and the accountability becomes collective — which holds only where a team makes it explicit rather than assumed.
solid answer
~50 sThe framework names three accountabilities and stops, because a named specialist seat inside the team creates a clear owner **and** a handoff — and the handoff lets everyone else stop caring at the moment they pass the work on. Everyone building the Increment is a **Developer**; specialist depth is a property of the people, not a seat. That only works when collective accountability is arranged for rather than declared. In practice a lead puts in place: a shared quality standard the whole team applies to every item; work shaped so more than one person can finish any given kind of item; specialists teaching rather than queueing, through pairing, review and written decision records; and an escalation path outside the team for decisions genuinely bigger than it. Job titles, ladders and line management are untouched — the framework is silent on them, and reading silence as abolition is the usual mistake.
go deeper
Know that the framework names only three accountabilities and that a tester or designer is a Developer in its vocabulary. Your company job title does not change; the framework simply does not describe it.
Explain the tradeoff a named specialist seat creates: it gives a clear owner and, with it, a handoff and a queue. Be ready to say what your team does instead — a shared standard, review, more than one person able to finish an item.
Show how you keep collective accountability real on a team with uneven skills: where you pair, what you refuse to let one person be the only route for, and how you spot work quietly piling up in front of a single specialist.
Own the design decision. Choose which expertise stays inside the team, which becomes a service outside it with a stated response time, and how depth is rewarded when the framework offers specialists no seat — because it says nothing at all about ladders or pay.
## Why the list of accountabilities is closed Scrum names a Product Owner, a Scrum Master and Developers, and then stops. Everyone who builds the Increment is a **Developer** in the framework's vocabulary — the person whose company title is test engineer, the one whose title is solution architect, the technical writer, the data specialist. The framework is not claiming those skills are interchangeable or unimportant. It is refusing to give any of them a **seat**, because a seat inside the team creates two things at once: a clear owner, and a handoff. The handoff is the expensive half. Name a tester accountability and testing becomes a stage that work is passed to, which means a queue, which means items that are finished-except-for-one-thing, which means everybody else is entitled to stop caring about quality the moment they hand over. Name an architect accountability and structural decisions become somebody else's, and the people writing the code stop making them well. ## What collective accountability actually requires It is not a slogan, and declaring it changes nothing. It is a set of concrete arrangements: 1. **A shared standard applied to every item.** The whole team agrees what must be true before work counts as finished, and it applies to everything, with no per-item negotiation under deadline pressure. 2. **No single-route work.** The practical target is that at least two people can carry any given kind of item to completion — a far weaker and more achievable requirement than everyone being equally good at everything. 3. **Specialists teaching rather than queueing.** The deepest expert's output should be other people's raised floor: pairing, review, written decision records, a short internal walkthrough. If their output is a personal work queue instead, the accountability has quietly become theirs alone. 4. **An escalation path outside the team.** Some decisions genuinely exceed a team — a regulated approval, a safety case, a platform-wide constraint. Those belong outside as a service with a stated response time, not inside as a fourth accountability. 5. **A way to see the failure.** Watch where finished work waits. A permanent queue in front of the same person every Sprint is a handoff by habit. ## What the seat gave you, and what has to replace it | What a named specialist seat provided | What replaces it under collective accountability | | --- | --- | | An obvious person to ask | Decisions written down so anyone can answer | | A gate before release | A shared standard applied by whoever does the work | | A career-visible identity for depth | A ladder outside the framework, plus visible teaching | | Somebody to blame | A team change after every unusable Increment | | Guaranteed specialist attention | Deliberate pairing and skill spread, planned not hoped | ## Where it fails in practice Consider a six-person team building billing for a community-energy supplier, with a 47-item Product Backlog and an on-call rotation that takes three of the six for about half their time. Exactly one person understands the tariff-rating calculation. Nothing is formally wrong: no seat has been named, the team calls itself cross-functional, and every Sprint roughly a third of the items still funnel through that one person — who is also on the rotation. The framework did not cause this and will not fix it. Collective accountability made it the team's problem to notice, and the team noticed only when somebody looked at where finished work was waiting. The repairs are unglamorous: pair on the next tariff item even though it is slower; write down the three decisions that live in one head; refuse to plan a Sprint whose critical items all need the same pair of hands. ## What the framework does not say A frequent misreading is that flat accountabilities imply a flat organisation. The framework is silent on job titles, seniority, pay, hiring, line management and career ladders, and silence is not abolition. A principal engineer is still a principal engineer; they are also a Developer in the framework's terms while on the team. Pretending otherwise produces the two familiar failures: seniority that leaks back in unofficially with no accountability attached to it, and specialists who see no path for depth and leave. ## The judgement call a lead owns Deciding what stays inside the team is the real work: - **Continuous, teachable, needed every Sprint** — keep it inside and spread it deliberately. - **Intermittent, scarce, and binding on more than this team** — put it outside as a service with a response time, and hold the team accountable for asking early rather than late. - **Needed every Sprint but held by one person** — treat that as a risk with a date attached, not a staffing preference. The honest summary in an interview: the framework removed the seat, not the work, and the work now has to be arranged for on purpose.
- When is a dedicated specialist seat outside the team the right call anyway?When the expertise is genuinely scarce, its decisions bind more than this team, and the work is intermittent rather than continuous — a regulated approval, a safety case, a rare platform skill. Put it outside as a service with a stated response time, keep the team accountable for asking early, and do not import it as a fourth accountability inside the framework.
- Does collective accountability mean everyone must be able to do everything?No. It means no piece of work can only be finished by one person, which is a much weaker requirement. The practical target is at least two people able to carry any given kind of item, with the specialist raising the floor through pairing and review rather than owning a personal queue.
- How would you know collective quality accountability is failing?Watch where finished work waits. If items pile up in front of one person, if defects are routed by default to whoever wrote the code, or if the team's own standard is quietly waived to close a Sprint, the accountability has been delegated back to an individual without anyone saying so.
Collective accountability is like a shared bill: it works only when somebody says out loud how it will be split, because otherwise everyone assumes someone else has it.
saying these in an interview costs you the question
- Claims Scrum abolishes job titles and seniority in the company
- Says testing is done by whoever happens to have spare time
- Adds a tester accountability to the framework to fix quality
- Treats collective accountability as nobody in particular
- Assumes cross-functional means everyone does everything equally