Data Model

The Data Model is the catalog of analytics fields available for a client (or your whole account). It lists the metrics and dimensions Jepto can query from your connected data sources, plus any custom fields you create with formulas.

When you build a report card, use Data Chat, or ask Jepto AI to create a chart, the fields you pick come from this catalog. Keeping the Data Model clear, and adding the custom metrics your team actually uses.

At a glance

• The Data Model shows what you can measure and break down for the selected client.
• Fields come from connected data sources (Google Ads, Meta, GA4, and others), plus Unified and Custom fields when available.
• You can browse, search, and filter the catalog, and create, edit, or delete custom fields.
• Custom fields use the Formula Builder (manual or AI) and then appear in report card pickers like any other field.
• The Data Model is not the same as the Data Warehouse (pipelines / BigQuery export). See How it fits together.

‍

What is the Data Model?

Think of the Data Model as Jepto’s governed field list for analytics:

Each row in the Data Model has:

• A human-readable Name (label), often matching language from the platform UI.
• A stable field key (for example googleAds.clicks or custom.…) used behind the scenes when cards and AI build queries.
• Kind (Metric or Dimension), Type, and an optional Description.

You do not write SQL to use these fields. You select them when configuring a data query on a chart or table card — or describe what you want in AI Edit / Data Chat, and Jepto resolves against this same catalog.

‍

How it fits together

Account Connections (OAuth / logins)
       ↓
Data Sources (attach platform accounts to a client)
       ↓
Data Model (fields + custom formulas available to query)
       ↓
Reports · Data Chat · AI (cards and answers use those fields)


Important: If a platform is not connected as a data source for the client, its fields will not appear (or will be empty) in the Data Model. Connect sources first, then return here to build custom fields and report cards.

‍

‍

Open the Data Model

For a specific client

1. Select the client in the client switcher.
2. Go to Data Sources → Data Model.

Typical path: /clients/{clientId}/data-sources/data-model

For the whole account (All Clients)

1. Set the client switcher to All Clients.
2. Go to Data Sources → Data Model.

Typical path: /data-sources/data-model

Client vs All Clients

• Viewing a client shows fields for that client’s connections, plus custom fields available to that client (including account-wide custom fields).
• Viewing All Clients is useful when creating custom fields that should be available to every client (see Enable for All Clients).

[Screenshot needed: Client switcher set to a named client, Data Sources → Data Model open, fields table visible.]

‍

Reading the fields table

The Data Model page is a searchable table of every available field for the current scope.

Toolbar

• Search — Filter by name, kind, or type. Placeholder: “Search by name, kind or type”.
• Data Sources filter — All Data Sources, or a single source (Google Ads, Facebook Ads, Unified, Custom, and so on).
• Field count — e.g. N fields.
• + — Create custom field.

Columns

Empty states:

• “No fields match your search.”
• “No fields available for the selected data source.”

‍

Labels vs field keys

• Label — What marketers usually recognize (“Clicks”, “Cost”, “Campaign”).
• Key — Stable id such as googleAds.clicks, unified.spend, or custom.…. Report cards and AI use these exact keys when building a data query.

You normally pick fields by label in the UI. Knowing the key helps when reading the Data Query preview on a card, debugging AI output, or writing formulas (source.field).

‍

Metrics and dimensions

Metrics

Metrics are quantitative values: spend, impressions, clicks, sessions, conversions, revenue, and similar.

Each metric has a default aggregation used when it rolls up in charts and tables — for example Sum, Average, Minimum, Maximum, Count, or Count distinct. Native fields come with sensible defaults from Jepto’s field standards. Custom metrics let you choose aggregation when you create the field.

Dimensions

Dimensions are attributes you group or filter by: campaign name, ad group, device, country, channel, and so on.

Date fields appear as dimensions with a time-related type. They power trends, report date ranges, and card granularity (day, week, month, …). In everyday language: “date fields” are how Jepto knows which period a row belongs to.

Why the distinction matters

• Chart metrics answer “how much?”
• Chart dimensions answer “broken down by what?”
• Formulas must match kind: do not SUM a text dimension; do not treat a pure text transform as a metric. See Formula.

‍

Source fields, Unified fields, and Custom fields

Source fields

Source fields come from a connected platform. Common source labels include:

Field keys are usually source.field in camelCase, for example:

• googleAds.spend
• facebookAds.campaignName
• googleAnalytics4.sessions

Labels and descriptions are written to match platform UI language where possible (what you see in Ads Manager or Google Ads), not raw API jargon.

Unified fields

When more than one distinct source type is connected, Jepto may expose Unified fields — shared / mapped concepts such as spend, clicks, campaign name, or date that can work across platforms.

Examples of unified-style concepts (exact list depends on your connections): Spend, Impressions, Clicks, CTR, Conversions, Cost per conversion, ROAS, Date, Campaign, Device, Country, Data source.

When to use Unified

• Cross-platform totals and trends in one card.
• Consistent naming when comparing Google Ads and Meta side by side.
• AI prompts like “total ad spend by platform last 30 days.”

Prefer Unified when it appears in your Data Model and matches the question. Use source-specific fields when you need a metric that only exists on one platform.

Custom fields

Custom fields are formula-based metrics or dimensions you create. They show a Custom Field badge and typically sit under the Custom source group.

Common uses:

• Total or blended spend across Google Ads + Meta (+ others).
• Ratios and efficiency metrics your team defines once and reuses.
• Cleaner campaign labels (CONCATENATE, case transforms).
• Extracting a value from a JSON/metadata field (JSON_STRUCT).

Full formula reference: Formula.

‍

Create a custom field

1. Open Data Sources → Data Model.
2. Click + (Create custom field).
3. Enter Name and optional Description.
4. Build the formula in the Formula Builder — type it manually or describe it for AI.
5. Set Kind: Metric or Dimension.
6. For metrics: set format (preset, type, currency / scale, precision, prefix/suffix, separators, compact) and Aggregation.
7. Optionally turn on Enable for All Clients.
8. Click CREATE.

Edit or delete

• Open the row menu on a custom field → Edit or Delete.
• Delete asks for confirmation: Are you sure you want to delete the "{label}" custom field?
• Native source fields are not editable or deletable from this table (you browse them; you extend the model with custom fields).

‍

Enable for All Clients

Use account-wide fields for definitions everyone should share (for example “Blended paid spend”). Use client-only fields for one-off or client-specific logic.

‍

Using the Data Model in reports and AI

Once a field is in the Data Model:

1. Open a chart or table card → Settings → DATA.
2. Pick metrics and dimensions from the picker (fed by this catalog).
3. Or use AI Edit / report AI and describe the chart in plain language — Jepto should resolve to exact Data Model field names.

See Data Query for metrics, dimensions, date range, filters, comparison, and sort.

Tip for AI: Name the outcome and the platforms (“blended Google Ads + Meta spend by campaign”). Vague prompts produce weaker field choices.

‍

Tips for getting the most from the Data Model

1. Connect the right sources first — Empty catalogs almost always mean missing data sources, not a broken model.
2. Use the source filter — When hunting for a GA4 metric, filter to Google Analytics 4 instead of scrolling everything.
3. Prefer Unified for cross-platform questions — Then fall back to source fields for platform-only metrics.
4. Create shared custom metrics once — Enable for All Clients so every report uses the same definition.
5. Name custom fields for humans — “Blended paid spend” beats “cf_sum_1”.
6. Document in the Description — Future you (and teammates) will thank you when AI and reports reuse the field.
7. Keep formulas aligned with kind — Metrics for numbers; dimensions for text/JSON extracts.
8. Don’t confuse Data Model with Data Warehouse — Warehouse is for pipelines and exported tables; Data Model is the analytics field catalog for Jepto reports and AI.

‍

Troubleshooting