- Home
- Compliance
- SOC 2 Compliance
- SOC 2 on a small team
Running the engagement
SOC 2 when the whole company fits in one room.
The Trust Services Criteria don't scale with headcount. The same 33 common criteria apply to a company of six and a company of six thousand. What changes is how you meet them. A small team can run fewer tools, combine roles and keep policies short, as long as every control still operates and leaves evidence.
Back to SOC 2 services01 · What scales down
What a small team can honestly simplify
The criteria are fixed. The machinery around them is not.
The criteria describe outcomes, not tooling. Nothing in TSP section 100 asks for a ticketing suite, a security operations centre or a forty-page policy. An auditor judges whether your controls are suitably designed for the risks you have and, in a Type 2, whether they operated across the period. A lean design that fits a lean company passes that test.
- Fewer tools. One identity provider, one code host, one tracker. Evidence lives where the work happens, so fewer systems means fewer populations to keep complete.
- Combined roles. One person may hold several jobs, if a compensating control covers the conflict each combination creates.
- Peer review for separation. A second engineer approving a pull request is a real segregation of duties control, even with no separate change board.
- Short policies. A two-page policy that states what you actually do beats a borrowed template describing a team you don't have.
- Lighter governance. A quarterly leadership meeting with minutes can carry the oversight that a large company gives to a committee.
Short is fine. Fictional is not. Every sentence in a policy becomes something the auditor may ask you to prove.
02 · What never shrinks
The controls no team size excuses
These run on a calendar or a trigger, and a skipped one shows up as a gap.
- Access reviews
- Someone checks who can reach production and customer data, on the schedule your control states, and records the result.
- Change approval
- Changes to the in-scope system are reviewed by someone other than the author before they reach production.
- Risk assessment
- At least once a year, a written look at what could go wrong and what you're doing about it (CC3).
- Vendor management
- You know which vendors touch the system and you review the important ones (CC9.2).
- Incident response
- A written plan, a way to log incidents, and evidence that you followed it when something happened (CC7.3 to CC7.5).
- Onboarding and offboarding
- Access granted on approval and removed promptly when someone leaves (CC6.2). Small teams lose this one most often, because leavers are friends.
Why size doesn't help here
These controls protect against the risks that hit small companies hardest: a departed contractor with live credentials, a rushed change with no second pair of eyes, a vendor nobody vetted. A small population also means a single miss is a large share of it.
03 · The founder trap
When one person does everything
The fastest way to fail a design test is a single person on both sides of every control.
In many early companies the founder is the admin of every tool, approves every change, reviews every access list and also holds the keys to production. Each of those is fine alone. Together they mean nobody checks the person with the most power. An auditor reading that design will ask who reviews the founder's own access and who approves the founder's own changes.
- List every control and write the performer and the reviewer side by side.
- Circle every row where the same name appears twice.
- For each circled row, move the review to someone else, even a part-time advisor or a board member, or add a detective control that would expose misuse.
- Put the founder's own accounts into the access review, reviewed by someone else.
- Write the result into the control description, so the design the auditor reads is the design you run.
04 · Role combinations
Common role combinations and what each needs
Illustrative pairings for a team of roughly five to fifteen people.
| Combined roles (illustrative) | The conflict | A compensating control |
|---|---|---|
| Engineer who also deploys to production | Can ship unreviewed code | Branch protection that blocks merge without one approval from another engineer; deploy only from the protected branch |
| CTO who is also the identity admin | Can grant themselves access with no trace | Admin activity log reviewed by the CEO during each access review |
| Founder who approves access and holds admin rights | Reviews their own entitlements | A second leader signs off the founder's own line in the access review |
| Ops lead who handles HR offboarding and IT removal | A missed step goes unnoticed | A leaver checklist with each tool ticked, compared against the identity provider at the next review |
| Security lead who writes and tests the incident plan | Marks their own homework | A tabletop exercise with leadership present and minutes kept |
| Finance lead who pays vendors and approves new ones | Onboards a vendor without review | New vendors with system access need a second approver before the first invoice |
Compensating controls must leave evidence too. A review that happens in a hallway conversation is, for the auditor, a review that didn't happen.
05 · A workable plan
A workable shape for a small team
Fewer, sturdier controls that a busy team will actually run.
- Keep the control set just large enough to meet the criteria. Every extra control is another population to keep complete.
- Put recurring controls on a shared calendar with a named owner, so they survive a busy month.
- Prefer controls the tools enforce (required reviews, single sign-on, automatic deprovisioning) over controls that rely on memory.
- Write down where each control's evidence lives the day you design it, not the week before fieldwork.
Sources: AICPA TSP section 100, 2017 Trust Services Criteria with revised points of focus (CC1 to CC9, including CC5.1 on segregation of duties); AICPA DC section 200; AICPA guide, SOC 2 Reporting on an Examination of Controls at a Service Organization.
Questions small teams ask
Will an auditor mark us down for having only one engineer?
Not for headcount alone; for an unreviewed design, possibly.
The auditor tests whether controls are designed to address the risk. With one engineer, peer review isn't possible, so you need another control: a reviewer outside engineering, a post-deployment review of changes, or alerting on production changes. Explain the design plainly in the system description.
Can our part-time advisor act as a reviewer?
Often yes, if they really review and leave a record.
What matters is that the reviewer is independent of the person performing the control and that the review is evidenced. An advisor or board member who signs a dated access review can fill that role. Make sure they can see what they're approving.
Do we need every policy a large company has?
You need the ones your controls rely on.
Policies exist to tell people what to do and to show the auditor the intended design. Combine where it reads naturally, for example access and onboarding in one document, and drop any policy that describes nothing you do.
Is a small population easier to pass?
Smaller, but less forgiving.
With few occurrences, the auditor may test every one. There is no room for a miss to hide in a large sample, so one forgotten leaver can stand out sharply in the results.
Related guides
Design controls your team can actually run.
A readiness assessment sizes your control set to your team, and names a compensating control for every combined role.