Analytics tools

Analytics tools answer one of two questions, and very few answer both well. Product analytics tells you what people did inside your app — funnels, retention, feature adoption. Web and marketing analytics tells you how they arrived — channels, campaigns, landing pages. Buying a product analytics tool to answer an acquisition question is the most common and most expensive mistake in this category.

A third group, warehouse-native tools, runs queries directly against your own database instead of storing a copy. They cost more to set up and less to trust, because there is only one version of the numbers.

What to check before you commit

• How events are priced. Per-event pricing punishes exactly the instrumentation depth that makes the tool useful. Per-seat pricing punishes sharing the data. Pick the failure mode you can live with.
• Whether you can get your raw data out. Export or warehouse sync is what makes a switch possible later. Without it, every year of history is a lock-in.
• Identity resolution. Check how the tool stitches anonymous visitors to signed-in users. Get this wrong and your funnels quietly overcount.
• Consent and regional rules. If you serve the EU or UK, confirm the tool supports consent mode and a lawful data-transfer basis before you instrument anything.

What to watch for

Dashboards are easy to demo and hard to maintain. The tools that stay useful are the ones where defining a new metric takes minutes and does not require an engineer. Before buying, try to answer a question you actually have during the trial — not the one the sales demo was built around. If tracking is blocked by ad blockers for a meaningful share of your traffic, factor that undercount in before comparing any two vendors' numbers.