Why Data Activation is a Real Deal

‍

10 years ago, a data warehouse was for BI. 5 years ago, for both BI and ML. Today, one of the fastest-growing use-cases is data operationalization, aka reverse ETL: a simple example is augmenting Salesforce data about customers with product usage metrics from the warehouse. Hightough's rebranding from a reverse ETL to a "data activation" company may seem like a clever marketing stunt in the cut-throat fight with Census. But it actually highlights a massive trend: increasing business automation with data. Interestingly, we don't seem to need cutting-edge ML models to automate most workflows in a typical business. All we need is a convenient bridge layer between the data stack and the SaaS tooling.

So is "data activation" just another buzzword?

Reverse ETL (which everyone tends to agree is an ok term to use) conceptually is just 1:1 data copy from a data warehouse into another tool's backend database. Data activation goes beyond that: for example, enabling data-driven workflows, where the data in the warehouse can be both a trigger and an input to an automated workflow. For example, when the product usage by an account falls below a certain threshold, create a Zendesk ticket and send a Slack message to the account owner.

Much like Zapier and its cooler and richer successors enabling API-based workflow automation, Hightouch in a way fulfills Martin Casado's prophecy about the imminent remaking of all SaaS apps into data apps.

After all, maybe in the battle between Census and Hightouch, the actual loser is Workato?

Let's activate that data

Tool of the week: lots of tools to build BI apps

Speaking of building data apps – tools and frameworks for developing analytical apps are on the rise. Here are a few ones with one-liners of how I understand them:

1. Hex – lets you build data apps collaboratively with an awesome notebook experience
2. Dash – from Plotly with enterprise
3. Cube – API + metrics layer for building data apps, bring-your-own-viz
4. Streamlit – sort of like Dash, maybe the future data OS. Snowflake paid $800M for it a month ago – that's a fact.
5. Datapane – sort of like Streamlit but emphasizes building "rich apps" that can be cached and embedded without a live connection to a DB.
6. Evidence – SQL + Markdown → Beautiful Reports, nuff said.

Before You Go

‍