The Narrator Platform
The Narrator Platform - Under the Hood
Narrator implements a consistent and company agnostic Universal Data Model (UDM) to make answering data questions, running analyses, and deploying algorithms quick and easy.
The Universal Data Model (UDM)
One 11 column table in your data warehouse, where each row represents a single occurrence of one of your customers doing an activity (like an event stream using all your existing data).
Tools to easily write queries on top of the UDM and audit individual customers.
A Better Data Science Community
As more companies model their data using the UDM, we can start sharing analyses and algorithms in order to build a thriving Data Science community.
The Current Approach:
Simple questions are not so "simple"
Asking data questions requires only knowledge of your business while providing the answer requires knowledge of your data system.
Answering even "simple" data questions can often take weeks to return a response.
Simple Data Question...
"How many people that came to the site called us?"
...Complicated Data Answer
... I need website data from Segment and call data from Salesforce
... Segment has an `anonymous_id` and Salesforce has a `contact_id`
... `anonymous_id` does not easily map to an email
... contacts have duplicate emails
... how do I handle multiple sessions and multiple calls?
After a week of work, the analyst responds with a 200 line query to generate the table that the stakeholder needs.
If there's a follow-up question (there's always a follow-up), this whole process repeats itself.
A platform to ask and answer questions the same way
The Universal Data Model
Your core business concepts and customers’ activities are defined up front (takes less than a week) and written in simple SQL on top of the source of truth tables in your warehouse.
Simple queries written on top of the source of truth
SELECT u.email as customer, ‘subscription_upgraded’ as activity, s.id as activity_id, s.upgraded_at as timestamp FROM subscriptions AS s JOIN users AS u ON s.user_id = u.id WHERE s.status = ‘upgraded’
via Internal db
Opened Marketing Email
Signed a Contract
Benefits to the Universal Data Model
Add and update activities at any time and take advantage of all historical data
Define concepts as activities once, allowing everyone to use consistent definitions
Follow the simple structure, and identity resolution happens automatically
Assemble Datasets For Each Question
Because of the UDM, we can now assemble datasets similar to the way stakeholders ask questions.
Take advantage of UDM concepts like customer, time, and relationships between activities:
Benefits to Assembling Queries on top of the UDM
Generate queries off of one consistent table so your numbers ALWAYS match
Join based on human concepts like occurrence (first, last, within_minutes), time, and customer, instead of foreign keys
View data and metric summaries to quickly become confident in the logic
With the Narrator Platform, iterate through data questions much more easily, so follow-up questions are encouraged!
Easy to write SQL (average of 14 lines vs. 1000).
No redundancy—each concept is clearly defined only once, and new activities don’t need to be created often.
Everyone can answer their own questions and assemble datasets in 5 minutes.
Coming Soon: Shared Analyses and Algorithms
Narrator is currently focusing on developing analyses and algorithms that we can run on top on your datasets. Imagine reading about an algorithm and then going into Narrator to run it immediately.
For example, all you'll need to do is assemble a dataset with a column for features and a column for conversion, and you'll be able to generate the following report: our sample report.
We love giving data advice
Sign up for a free 30-minute chat with a senior data analyst.