Modeling

Made to fit how your business runs.

Your teams define customers, revenue, and joins differently than the next company. The modeling layer captures yours.

Book 15-min demo See how we deploy
Models › daily_marketing_performance
Built 6h ago
Aggregates daily marketing performance by county, business line, and channel. Default model for Marketing questions about performance breakdowns and channel attribution.
Properties 15
Relationships
Aggregations 16
Aa Business Line string
Aa County string
Day date
Aa Marketing Channel string
Aa Market Tier string
# Spend number
# N Qualified Leads number
# N Contracts Signed number
# CAC by Geo Market number new
Operator 9:42 AM
Add CAC by geo market.
Ran Update Model
AI context

Operator runs on your model.

Operator answers questions from the model, not the raw tables. Ask why CAC spiked last week, and the math runs through your customer definition, your revenue rules, your joins. No guessing at what a customer is.

Operator 9:42 AM
Grounded in
paid_channel_attribution model
daily_marketing_performance model
customer = active within 90 days definition
Why did Meta CAC spike last week?
Ran Analyze Spend
Ran Build Cohort
Meta prospecting CAC hit $72 last week, up 34% w/w. Driven by a campaign launched May 8 targeting cold audiences. Other channels flat.
AI enrichment

Enrich your data.

Switchboard uses AI to generate data your sources don't have. Configure a property, give it a prompt, and it lands as a real column in your model: filter on it, group by it, build dashboards with it.

Property Name
weather
Type
LLM
Inputs
Prompt
For this zipcode, return the weather conditions on the order date.
Type @ to reference properties or tools.
Properties
@orders.shipping_zip
Tools
Web Search
Output
Value Type
String
orders refreshed 4h ago
id customer items weather LLM
8847 Marina B. 3 Snow
8848 David L. 1 Light rain
8849 Priya K. 5 Clear
8850 Jay T. 2 Heavy rain
8851 Sam R. 4 Clear
↓ Filter, group by, and chart this column like the rest of your data.
Lineage

Trace numbers to their source.

Each model knows what it's built from: sources, upstream models, and the Documents and Automations downstream. Run history travels with it, so the data behind any number is auditable end to end.

When Operator answers a question, it traces the math. When you change a definition, it shows what depends on it before the change ships.

Model lineage daily_marketing_performance
Sources
Shopify
HubSpot
Meta Ads
Google Ads
Models
customers_base
ad_spend_base
Output
daily_marketing_performance
All models built · 4h ago View runs →
How we deploy

We deploy alongside the people who run the business.

Senior data engineers embed in your business and run weekly sprints. They talk to ops, finance, growth. They learn how a customer is defined at your company, how revenue rolls up, which joins your team relies on.

The model that results isn't a template. It's a working description of how your business runs.

Implementation
Modeling sprint, week 2
Active
Day
12
Sources
12
Models
47
Joins
18
Recent modeling work
+ Mapped "customer" with growth team Day 1
+ Unioned DTC + wholesale storefronts Day 3
+ Defined CAC + LTV with growth lead Day 5
+ Built revenue attribution with finance Day 8
Modeling cohort retention by SKU Now
Get started

Modeled to fit your business in days, not months.

Book a 15-minute demo. We'll walk through a sample Switchboard instance and show how your business gets modeled.

Book a 15-minute demo