Sharing our learnings on data tooling

Free data tool knowledge incoming!

Side launch! Exploring data tools? We've spent a lot of time working with startups and building data infrastructure. What Data Stack is the Wirecutter for Data Tools - we're open sourcing our knowledge and research to help you through buying decisions.

---

Last week, I chatted about how difficult it can be to navigate which SaaS tools to buy and how our choices will only multiply in the coming years. Founders and leaders also must consider if they should build instead of buy. The preference, for many technology companies, is to build internally. This offers customizability and cost savings, but also layers on complexity, roadmap tradeoffs, and countless potential headaches down the road. If it’s not part of your core product and will be a distraction to engineers, buying may make a lot of sense. Further, if your team does not have expertise in this space (e.g., non-data engineers) then buying may make even more sense. Regardless of the decision, the goal is to be a mountain bike.

We found that with any startup we advised or any meetup we attended, the conversation almost always landed at: “What tools do you use?”. Given our backgrounds in the data space, we had strong opinions. We love chatting with startups about this, so we decided to formalize our knowledge with this survey tool: What Data Stack?

We ask a series of exploratory questions before writing up a short diagnosis of the data stack that fits your needs and sharing it with you. Our goal is to help you select the best product analytics, ETL, warehouse, and transformation tools that minimize monthly SaaS cost. In many cases, we may recommend no tools and instead a series of hacks to get the job done. Check it out!

Happy weekend,

Raman at Rhetoric

---

📚 What’s made me a better storyteller this week

Strong opinions about when it comes to weather apps—which are to be trusted and which simply reduce complex weather data to a simple emoji that’s usually wrong. Here’s the story as to why.

A fun use case that is actually a masterclass in training AI chatbots: Teaching ChatGPT to Speak my Son’s Invented Language.

Filed under “fascinating data problems that I hadn’t considered until now”: how do we make alarms (smoke alarms, car alarms, security alarms) just sensitive enough to only sound during highly-specific scenarios?