❌ Don’t do this! Post-seed, pre-A series is one of the most critical times in your startup’s life and you need to make sure you understand what is happening with marketing, on your website, or in your mobile app.
Here are some tips & tricks that work well in order to build a scalable foundation for analytics while not breaking the bank:
- DO ✅: make sure you start to understand your business’ north stars. This seems vague and unspecific, but we see most teams fail with analytics if there is no good understanding of what they actually want to achieve with it.
- DO ✅: set up a simple self-serve reporting tool — they will not scale well, but they are sufficient to get you going for now. We like Amplitude as it has a good event capture layer, and its data can easily be exported — but we have also had good experiences using Mixpanel or even plain Firebase with Google Data Studio reports.
- DO ✅: once you have that reporting tool, empower your growth strategy with simple metrics that you can track weekly. Don’t re-invent the wheel here, you probably don’t need a fully custom retention metric just yet! Clean, understandable (QA’able…) and self-serve does it.
- DO ✅: at the same time, think about simple event tracking to track common user journeys. This sets you up for the future data platform as events are ephemeral — if you don’t capture them now, there is no way to recover them! Do not overcomplicate this; simple page views or screen views (with proper identity tagging) is sufficient. Pricing tiers for tools like Amplitude or Rudderstack are quite attractive for early-stage companies and they are easy to implement, so we recommend to use at least one of those tools.
- DO NOT ❌: However, do not overinvest in your data platform yet! You should start capturing events in a tool like Amplitude (that has great visualisation options for straightforward insights like usage, retention, churn etc) but don’t start modelling the data yourself just yet. Your business is still changing fast, and chances are any modelling you do now will be irrelevant in six months.
- DO NOT ❌: It is also too early to hire any data specialist yet! While for some companies a data analyst might have immediate use, there is almost certainly no reason yet for hiring data/analytics engineers — let alone a full-fledged machine learning engineer. Unless it is a critical part of your product of course :)
At Tasman we have helped out dozens of growing companies in Seed stage making the choices above — contact us for a free consultation!