Every business keeps track of things: customers, properties, equipment, jobs. In Atlas, the way you describe each kind of thing is called a classification, and it’s where the Enterprise tier begins.
The definition
A classification is a blueprint for a kind of thing.
Suppose Evergreen Landscaping, the example business in the Atlas tutorials, wants to keep track of every property it looks after. That’s a classification: Service Properties. It holds everything that describes a property.
What a classification holds
- Record schemas say what one record of that kind looks like — for a property, a Property Profile built from reusable data schemas like Service Address and Site Details.
- Sample schemas describe measurements taken over time.
- Topics are the individual things themselves — one for each property, such as 42 Juniper Lane.
- Pipelines are the process each topic moves through.
- A security policy decides who can reach all of it.
In Atlas’s documentation you’ll also see the term classification module; it’s the same idea, emphasizing that classifications are containers that model one part of your business.
Two positions: data modules and resources
A classification takes one of two positions:
- A data module sits at the top, and its topics stand on their own. Service Properties is a data module: each property is independent.
- A resource sits inside a data module, and each of its topics belongs to one topic of the module — like a site survey that belongs to one particular property.
Resources go one level deep, so decide early which of your things stand alone. Data modules and resources covers the choice in detail.
A network that mirrors your business
Classifications connect. A customer connects to their properties, and each property connects to its service history. Together, classifications form a network of connected data structures that mirrors how your business actually works — which is why Atlas describes the Enterprise tier as organizing not just your forms, but the whole business.
Good habits
The Atlas video series recommends three habits:
- Name the things, not the tasks. Properties and Customers, rather than Inspecting or Billing.
- Make one classification for each kind of thing you keep track of.
- Decide what stands alone before you build, because resources only go one level deep.
Where classifications live
Classifications live in the File System, under Data Modules. Every top-level classification in your organization is listed there. Select one and a details panel opens with what you can do to it: open it, edit it, change its security policy, or archive it. Open one and you’re inside it, where its record schemas, sample schemas, topics and resources are all items in the same File System, with a breadcrumb showing where you are. See Finding your way around the Atlas File System.
Why classifications matter
For a business, organizing data into classifications pays off quickly:
- Your data lives in one place instead of scattered across spreadsheets and inboxes.
- Every record follows the same structure, so reports and handoffs are reliable.
- Security policies control who can see and manage each part of your data.
- As the business grows, you add classifications instead of starting over.
Choosing your first classifications
Most businesses need fewer classifications than they expect. A good first set covers the things that everything else refers to:
- Customers — who you serve.
- Locations — properties, sites or premises.
- Assets — vehicles, equipment or tools, if you track them.
Everything else usually hangs off these: surveys belong to locations, service history belongs to locations, maintenance belongs to assets.
Start with two or three and add more as you learn. Adding a classification later is straightforward; untangling one that was modeled wrongly is harder.
Classification, topic, record: how they fit
These three terms do different jobs, and mixing them up causes most early confusion:
| Term | What it is | Example |
|---|---|---|
| Classification | The blueprint for a kind of thing | Service Properties |
| Topic | One individual thing | 42 Juniper Lane |
| Record schema | What you keep on file for that kind | Property Profile |
| Data record | The filled-in information for one topic | 42 Juniper Lane’s profile |
A classification is a type; a topic is an instance; a record is what you know about that instance.
Common mistakes
Naming tasks instead of things. Inspecting describes work; Inspections describes things you keep.
One classification for everything. A single “Records” module with every kind of thing inside loses the structure that makes classifications useful.
Nesting too deeply. Resources go one level deep. If you’re planning three levels, reconsider which things stand alone.
Leaving ownership undecided. Every top-level data module wants a user group with Manage access, because that’s the group its work is raised to.
Frequently asked questions
Do I need Enterprise for classifications? Yes — classifications are part of the Enterprise tier, along with security policies and API access.
Can a classification hold more than one record schema? Yes. A classification can define several record schemas; inside a topic, the New menu offers what the classification defines.
Can I change a classification’s security policy later? Yes. Selecting a classification in the File System opens a details panel that includes changing its security policy.
What happens to a classification I no longer use? It can be archived from its details panel, which takes it out of use without destroying its history.
Next steps
- Create your first data classification — every field on the form, step by step.
- Modeling your business in Atlas — how to decide which classifications you need.
