October 9, 2019

New idea, pivoting from another project

Todd Heslin @toddheslin

My original idea was called AudienceHQ. It was a platform that synced Google Analytics data on a per user basis using the Google Analytics cookie. When a user arrives on my eCommerce website, their sessions, goals, and segment attributes would be available for developers like myself to build personalised experiences.

I might still run with it some time in the future, but I realised the problem was more fundamental than I was trying to originally solve.

The broader question I was trying to answer is this: "How do developers build data-driven products?"

I've was a data analyst for a while before getting serious about software programming. I'm incredibly comfortable with building ETL data pipelines, cleaning and querying data sets, and building API services wrapped around these data. But when I chatted with my SaaS friends, they barely jumped out of Rails' Active Record and knew almost no SQL.

I was introduced to GraphQL a couple years back and it took me a while to 'get it'. After building a working prototype for the original AudienceHQ idea, I realised that I kept wanting to query more and more data. It was here I realised I just wanted to make a GraphQL request in my client-side application and this to translate to the complex SQL against my data warehouse.

Graph Compose is born.