How to build data platforms using the modern data stack
“Where is the data and how do we connect to it?” These are fundamental questions your team will need to answer before embarking on a data and analytics project. That is, of course, once you’ve determined which data is actually required for the endeavor.
This post will show you exactly how this is accomplished – through the lens of Untitled Firm, a company that builds and manages modern data stacks for a living. The modern data stack is a suite of tools that power data integration: a data pipeline, cloud-based data warehouse or data lake as a destination and a data visualization platform. It’s a radically new approach to data integration that saves engineering time, promotes end-user accessibility and scalability to quickly meet an organization’s growing data needs.
Internal vs. external data integration
There are essentially two types of data projects an organization can pursue – internal and external.
An internal data project is when an organization connects to its own data. In this scenario, data access is an easy problem to solve. The login information and keys needed to connect to data sources belong to the organization. They are already in its hands.
For external data projects where an organization needs to connect to its customers’ data, this is a trickier problem to solve. Why? Unlike internal projects, external ones don’t have access to data source credentials in its possession. They are in the hands of its customers.
This prompts the logical question: how do you get access to data that doesn’t belong to you? Let’s dive into how to solve the data authentication problem, so that you can build analytic products that pop.
Possible solutions to the authentication dilemma
Every data-oriented software product that connects to customer data needs to pick one of these two options.
- Simply ask your customers. Smart, sophisticated customers will send you their source system credentials in encrypted form through a secure network. Less sophisticated customers will copy the username and password for their production Salesforce account, paste into an email and click ‘Send’. Yikes. In either case, you subject your customers to a clunky exchange that is often not secure. After all, even with the first option, the moment you decrypt your customers’ source credentials, you become liable for having them in your possession. The security blanket is compromised.
- Let your customers grant you access, entirely on their own, without you ever touching their credentials. Instead of manually creating connections to data sources on your customers’ behalf, customers can authenticate connections to their data themselves. Not only is this a far superior experience for your customers, it’s a much faster way for you to begin working on your customers’ data and have the capacity to scale to more customers.
Untitled Firm uses Powered by Fivetran to help its customers connect to their data from within the Untitled Platform. Fivetran Connect Cards, the primary feature of Powered by Fivetran, are the front-end authentication experience that Untitled’s customers engage with to connect their data sources. Working hard is the Fivetran API which uses the customers’ credentials to authenticate a connection between the specified data source and Fivetran, so that it can begin syncing data through its pipelines to Untitled’s data warehouse. The entire exchange is brokered by Fivetran — not Untitled — thus offloading the hassle and security risk to Fivetran’s servers.
If you’ve ever used a personal finance application like Intuit Mint, Charles Schwab or Yahoo! Finance, then you’ve experienced this type of functionality first-hand.
The front-end authentication experience
To understand how this process works for Untitled Firm, let’s look at its platform’s experience from the perspective of a customer:
Step 1: Log in to the Untitled application
Step 2: Click on a data source supported by Untitled Firm
Step 3: Pick a destination
This is where the customers’ data will be stored.
Step 4. Enter credentials for the data source into a simple setup form
Step 5: Observe notification that says ‘Success! You have connected your data source to Untitled’
The entire experience is designed to reduce as much friction as possible for the customer, so they can spend their time analyzing their data – not connecting it to the Untitled platform.
The back-end architecture and orchestration
Behind the scenes, Untitled Firm employs a series of steps to aggregate the raw data using dbt Core and then applies models of its own to make the data ready for visualization and presentation to the customer.
As a result, Untitled enables its customers to gain actionable insights to questions, such as:
- What are the characteristics of top performing employees?
- What is the most effective use of advertising dollars?
- Which factors are most responsible for customer churn?
If you’re building a data product or analytic application of your own, consider implementing a similar approach. Your customers will thank you for it, and you’ll thank yourself. By automating the data onboarding experience, you solve a time-intensive process for both you and your customers enabling you to quickly scale your application so that you can support your growing customer base.
Learn more about why Powered by Fivetran is the easiest way for your customers to connect data to your app.
Start for free
Join the thousands of companies using Fivetran to centralize and transform their data.