← All posts
SOC 2 data flow diagram templateSOC 2 data flow diagramSaaS data flow diagram

SOC 2 Data Flow Diagram Template for SaaS Teams

Map customer data through your SaaS service for SOC 2. Copy a small data flow example, check each route against deployed systems, and keep the diagram current.

A data flow map connects customer inputs, service components, storage, backups, logs, and provider boundaries.
A data flow map connects customer inputs, service components, storage, backups, logs, and provider boundaries.

A SOC 2 data flow diagram should show how information enters your scoped service, where it is processed and stored, which providers receive it, and how it leaves or is deleted. Start with one real customer journey and draw only the routes you can verify. The example below is a template to adapt, not a diagram required by the AICPA.

What does SOC 2 ask you to explain about data flow?

The AICPA SOC 2 Description Criteria say that management’s system description should help report users understand the processing and flow of data through and from the system. The criteria do not prescribe a drawing format. They allow flowcharts, tables, graphics, and other aids alongside the narrative. A diagram is useful when it makes the SOC 2 scope and provider boundaries easier to check.

The NIST Cybersecurity Framework 2.0 also recommends maintaining representations of authorized network communication and internal and external network data flows under ID.AM-03. That is separate security guidance, not an extra SOC 2 diagram rule.

Copy a small SaaS data flow example

This fictional service accepts an account profile and a customer file. The application API stores the profile in a database and the file in object storage. The database sends snapshots to backup storage. The API sends event metadata to a log service. A customer can download the file through the API. The example does not say these paths are safe or complete for your product.

Customer browser (outside candidate service boundary)
      | F1: profile + file
      v
+-- Candidate service boundary -----------------------------------+
| [Application API, team operated]                                |
|      | F2: profile --> [Database, cloud provider]               |
|      |                        | F4: snapshot                     |
|      |                        v                                  |
|      |                   [Backup store, cloud provider]          |
|      | F3: file ----> [Object store, cloud provider]            |
|      | F5: event metadata --> [Log service, log provider]       |
| [Object store] -- F6a: file read --> [Application API]          |
| [Application API] -- F6b: file download --+                     |
+------------------------------------------|---------------------+
                                           v
                                     Customer browser

The boxes name components and operators; the arrows name information. The boundary is a candidate map for management to check. A provider label shows who operates a component, but does not decide how the CPA firm treats that provider. Add support access, data deletion, and any other route your service actually uses. Do not assume the log service receives only metadata until someone checks its current configuration and a safe sample of output.

Use this flow register beside the drawing. Copy the headings and replace each example value with the deployed route. The source reference can point to a dated configuration export, log sample, or other safe record. The owner and checked date stay blank until someone performs the check.

route,from,to,information,operator,what_to_check,source_reference,owner,checked_on,status
F1,Customer browser,Application API,Account profile and file,Team,Ingress endpoint and request fields,,,,pending
F2,Application API,Database,Account profile,Cloud provider,Connection and region,,,,pending
F3,Application API,Object store,Customer file,Cloud provider,Bucket and access path,,,,pending
F4,Database,Backup store,Database snapshot,Cloud provider,Backup target and retention,,,,pending
F5,Application API,Log service,Event metadata,Log provider,Collector settings and actual fields,,,,pending
F6a,Object store,Application API,Customer file,Cloud provider,Read permission and source object,,,,pending
F6b,Application API,Customer browser,Customer file,Team,Download endpoint and authorization,,,,pending

These rows describe proposed routes. Mark a check complete only after its owner compares the row with current source records and adds a reference and date. If a route is unknown, keep it pending and assign someone to investigate it. Do not fill the gap with a statement copied from a policy.

Check the boundary and every copy of customer data

Start from the bounded system rather than an org chart. For each information type, ask where it comes from, where it goes, who can reach it, what copies persist, and how it leaves. Include backups, logs, exports, support tools, and integrations when they receive or can access the information. Check relevant provider relationships separately through the vendor review process. A provider box does not automatically make that provider a subservice organization in the examination.

Keep the picture at a level an engineer can verify and a report reader can understand. A service component belongs on the main diagram when it carries a material flow. Put short-lived instances and low-level network detail in supporting architecture or configuration records when they matter. The asset inventory guide shows how to reconcile those records without turning the diagram into a device list.

Verify the arrows against deployed systems

Ask the owner of each route to compare the map with configuration and source output:

  1. Check deployed application configuration and infrastructure definitions for the actual endpoints, accounts, regions, and integrations.
  2. Check database, storage, backup, and log settings for destinations, retention, replication, and access. Preserve a dated export or safe reference to the source.
  3. Compare data fields with the route labels. Logs and support tools often expose more than the team intended, so review a safe sample without copying secrets or customer records into the diagram.
  4. Walk through a real creation, read, export, and deletion path. Record missing routes, uncertain provider access, and exceptions for follow-up.
  5. Have a second owner review the diagram and its source references. Record the review date and the version approved for the current system.

The diagram helps explain the system. It does not prove that encryption, access control, backup, or deletion controls operated. Keep the source configuration and dated operating evidence that supports those claims. The CPA firm decides what it will test and whether management’s evidence is sufficient.

Keep the map tied to changes

Review the data flow after a new integration, information type, storage location, region, support path, or provider changes the service. For a Type 2 period, keep the prior version and the change date so the final description does not erase a route that existed earlier in the period. Match the map to the SOC 2 system description before management approves that text.

FileGRC is a Git-native GRC workspace for SOC 2 work. JSON holds structured records, Markdown holds long-form work, and Git supplies the change history. Its System, Component, Information Type, and Vendor records can give an agent the context to draft a map and find missing links. Starter records are proposals, not proof of live data movement. Your team verifies the routes against source systems, reviews the draft, and makes the scope and control decisions. The independent CPA firm performs the examination.

Open source · MIT

Run your SOC 2 program as files in Git.

Keep policies, controls, work, and evidence indexes in a repository your team and agents can inspect. Add optional FileGRC Autopilot email and Slack reminders to keep work moving.

Frequently asked questions

Does SOC 2 require a data flow diagram?

The AICPA Description Criteria ask management to describe how data and information move through and from the system, but do not prescribe a diagram format. A diagram can support that description. Confirm the presentation and supporting evidence with the CPA firm for your engagement.

What should a SaaS SOC 2 data flow diagram show?

Show the scoped service boundary, the main information types, where each enters, is processed, stored, copied, sent to a provider, returned, or deleted, and the people or systems with access. Label each route and check it against current deployed configuration and source records.

Should a SOC 2 data flow diagram include vendors?

Include a provider when it receives, stores, processes, backs up, or can access information relevant to the scoped service or a control. Label the data and responsibility boundary. A vendor's presence on the diagram does not by itself decide its treatment in the CPA engagement.

Is a data flow diagram enough to prove a SOC 2 control operated?

No. A diagram describes an intended and reviewed route. Configuration exports, logs, access records, change history, and other source evidence show what actually happened over the relevant date or period. The CPA firm decides what evidence is sufficient.

Can FileGRC generate or verify a data flow diagram automatically?

FileGRC records Systems, Components, Information Types, Vendors, and their links in a Git-native GRC workspace. An agent can use those records to draft or check a map, but your team must compare it with deployed systems and provider facts. FileGRC does not discover every live data route or perform the SOC 2 examination.