I gained several insights after reading the article [Translation] The Dangers of Google Analytics-Powered Startups | POSTD, so I’ll record them here for reference.

This doesn't mean startups shouldn't rely on quantitative data from sources like Google Analytics. Rather, they need to be more sophisticated in understanding that insights come in many forms, and resist the urge to base all reporting and performance measurement on the most readily importable channel into a spreadsheet.

The more you learn about tools, the more you’ll understand their strengths and weaknesses. While it’s great to use tools yourself, it’s probably even more effective when tools are used on you. Wouldn’t it be nice if your boss set goals based on Google Analytics and you could explain why the numbers don’t look as good as they “should”? I think so.

The point in the article that “sampling makes most consumer reports very inaccurate” really resonated with me. For example, if you’re analyzing purchase paths on an e-commerce site using only sampled data, you might make critical judgment errors.

While Google Analytics is easy to implement and very convenient, what concerns me is that the data collection methods are a black box. If you want more accurate data analysis, I recommend taking an approach of implementing your own tracking system.

That’s all from the Gemba.