>So, when you put in a new feature, don't measure "how many people used the new feature", measure "whether people who had access to the new feature were more likely to pay"
I like that part. Jim Collins ("Good to Great") recommended to have one metric that "drives your economic engine". For a startup, you may not know what your economic engine is yet (still be pivoting an all that) but at least be measuring for a metric that relates to your economic, not popular, appeal.
In most early web startups, the biggest challenge (and the most common reason for death) is acquiring and retaining customers. User metrics are completely valid as your primary metrics early on.
I think you're getting at a very interesting conflict here that the author of the article attempted to explain. While user metrics can gauge the success of a web startup, that's often all they can do. Those metrics will (potentially) tell the company if they're improving or declining, and if they should be taking action.
However, it will not tell the company what actions to take. They're nowhere near as useful as a targeted, action specific metric in determining what to do. If my website's # of visitors drops 50% over a week, there might be something wrong, but I don't know what action to take without more information. Figuring out what engages users on the site and what keeps them coming back and measuring those things would be way more useful than simply knowing how many users are visiting.
If you track traffic sources with high granularity and get lots of traffic from ads or viral channels, you can get actionable data from traffic metrics.
Overall I agree - actionable metrics are the most important. I just take issue with the idea that only revenue related metrics are important.
Not sure I agree. I'd say "measure both". If people use it, it clearly has value to them, and that's very interesting whether or not they're more likely to pay. There's a big difference between "useless" and "useful but not something people are willing to pay for with the current pricing/upselling scheme".
I was expecting an article on how to measure performance in a startup - an interesting topic in itself given the fluid nature of the environment which makes it difficult to measure many important aspects - but the article seems to conflate web stats and metrics and does not offer much that is interesting or actionable.
Interestingly, the word metric is used 55 times in the article.
Perhaps you missed the point. Using metrics is important to decide what actions to take, or, more accurately, measuring whether a taken action lead to an important outcome.
It's no good just measuring visits or even conversions. What is important is measuring whether action a increased visits/conversions better than action b. The fluid nature of a startup means that most people probably don't know - there's no corporate history to lean on - and thus making careful use of metrics even more important.
> When not to use metrics... If your traffic is too low for metrics to be meaningful, then measuring metrics is largely a waste of time, and making decisions based on statistical noise is potentially harmful.
This is very very true. However, this should not deter you from making sure you have all the data tracking in place very very very early on. It's a HUGE pain to add metrics tracking later when you're growing like crazy.
I like that part. Jim Collins ("Good to Great") recommended to have one metric that "drives your economic engine". For a startup, you may not know what your economic engine is yet (still be pivoting an all that) but at least be measuring for a metric that relates to your economic, not popular, appeal.