[01] The data model
Every source. Every metric.
One unified model.
Shopify, GA4, Meta Ads, Google Ads, and Klaviyo all linked into one shared context. Dashboards, signals, and AI briefs read from it.
Available on Shopify App Store
Karbon Analytics
Model
Sources
One schema. Every source mapped, every customer linked.
Outputs
Inside the model
All platforms.
Every dimension. Every metric.
One place.
The model defines how every source maps into one shape, how records link across platforms, and how derived metrics get calculated. Reports query it. They don’t reinvent it.
[01] Unified Schema
Every object,
one shape.
Orders, sessions, customers, ad spend, events. Across every source, modeled the same way. Derived metrics like MER, ROAS, AOV, LTV, and CAC sit on top, calculated once and read by every output.
Karbon Analytics
Model
Schema
Karbon Analytics
Model
Identity
Meta Ads
Ad click
Klaviyo
Email open
GA4
Session
Shopify
Order
Customer #4815
unified.customer
[02] Shared Context
Customers, channels, events.
Connected.
One buyer across Meta, Klaviyo, GA4, and Shopify stays one customer in the model. Channels collapse to one canonical taxonomy. Touchpoints reconstruct as one journey. The links are pre-built; reports query them.
[03] Cross-platform reporting
Every source,
queryable together.
Compare Meta vs Shopify. Blend all ad spend into one MER. Build funnels that cross channels. Slice by either. All from one model, in one query.
Karbon Analytics
Dashboards
Cross-channel
Revenue
$124,400
+12.3% vs prev
Ad Spend
$32,100
+4.1% vs prev
Blended ROAS
3.87x
+0.42 vs prev
MER
3.87
−0.18 vs prev
Revenue by channel · last 7 days
cross-source queryManaged infrastructure
All this,
without the development overhead.
Karbon Analytics provisions, syncs, and maintains the model. You skip the months of integration work, schema design, and pipeline upkeep that a unified analytics stack normally takes to build in-house.
Data warehouse
- Snowflake
- BigQuery
- Redshift
- Postgres
ETL pipelines
- Fivetran
- Supermetrics
- Airbyte
- Stitch
SQL or dbt models
- dbt project
- SQL transforms
- Metric layer
- Schema design
In-house build-out
- Custom integrations
- Schema modeling
- Pipeline orchestration
- In-house expertise
How the model works
From every source to one model, in four stages.
01 · Ingestion
Sources stream in continuously
Sources stream in continuously. No exports, no API keys, no schema mapping on your side. Once you authorize a connection, Karbon Analytics pulls data from Shopify, GA4, Meta Ads, Google Ads, and Klaviyo on a read-only schedule and keeps the model live.
02 · Normalization
Raw fields map to a shared schema
Every source mapped to one common shape. Each platform has its own event names and field structures; Karbon Analytics maps each into a unified schema so orders, sessions, customers, spend, and events land in a consistent structure inside the model.
03 · Linking
Shared keys connect every object
Customers and events linked across platforms. Customer identity, channel taxonomy, event time, and cohort membership join Shopify orders to Meta clicks to GA4 sessions to Klaviyo touchpoints. Pre-built, so reports query the relationships instead of reconstructing them.
04 · Activation
Every output reads from the same model
Every dashboard, signal, and AI brief reads from the model. Dashboards, Daily Signals, executive briefs, attribution analysis, and AI insights all query the same connected view. A metric defined once propagates everywhere, and nothing is recomputed downstream.
Karbon AnalyticsPipelineIngestionConnected sources · streaming
Recent events
Schema map · 2 of 12 shown
many → 1revenue
Unified
unified.revenue
decimal · usd · refunds applied
customer identity
Unified
unified.customer
uuid · cross-platform
Shared keys · linked across sources
one connected viewMeta · Klaviyo · direct linked to one person
3 platform labels · 1 canonical channel
ad click joined to order via UTM + identity
cohort tag applied once · queried everywhere
Outputs · reading from the model
5 / 5 healthyUnified Data Model
One source of truth
By the numbers
Karbon Analytics already analyzes serious data. Built to handle the volume your store produces.
What reads from the model
Three outputs. One model.
Every output in Karbon Analytics queries the same unified model. Same dimensions. Same metric definitions. Same source data, no matter which surface you open.
[01]
Dashboards
Seven dashboard categories built on the unified model. Sales, marketing, customer, inventory, profitability, marketing efficiency, and operations. Every category reads from the same model.
See the dashboards[02]
Daily Signals
40+ detectors run against the model every night. The model lets a signal spot what single-platform tools cannot: cross-channel attribution shifts, identity-driven double-counting, true ROAS slips.
See Daily Signals[03]
Meta ROAS slipped to 1.7× against a 2.5× baseline. Two campaigns are pulling the blend down. Pause and reallocate before lunch. Expected recovery within 3 days.
AI briefs
AI writes the morning brief on top of the model. Plain-language summaries, ranked by impact, with a suggested next step. The model decides what is important; AI explains what happened.
See AI briefsReady when you are
See your real numbers in one place. Every morning.
Connect your store and ad accounts to Karbon Analytics. See what a unified data model unlocks in your first session.
FAQ
Frequently asked questions.
Everything about how the unified data model works, from connection through to activation.
About the model
What is the Karbon Analytics unified data model?
How is this different from a dashboard tool that just connects multiple sources?
What does the model actually do?
About cross-platform analysis
Why does this matter for a Shopify brand?
About setup
Do I need to configure the model?
Does the model apply the same rules to every account?
Still have questions?
Talk to the teamContinue exploring the platform