A good Product Manager acts more like a GM than just a product owner. Accounting for the bottom line metrics of the business, they should keep an active pulse on the state of the business as well as monitor any signals of interest. The coffee dashboard has been the most effective resource in any company I’ve worked at.
Throughout my career, I have found the experience I learned at startups were far more valuable than what I learned at larger companies. This is because the Product Manager was more of a GM in their role and had a broader set of responsibilities that required them not only to go deep, but broad. When you’re managing several products and affecting the bottom line of the business, it’s important to have a snapshot of the business as it impacts your domain: enter the coffee dashboard.
The first thing you should do in the morning when you wake up and open your laptop, is to first look at the coffee dashboard. This level sets your day and gives you problematic areas in your product/business to surface. As you view it, enjoy your cup of coffee and spend at least half an hour intuiting the metrics.
What metrics should you look at? It depends on what your product surface area is, but my coffee dashboards have always been the following:
- Important business metrics (financial, conversion, retention, etc)
- Core metrics tied to your top/important features
- This should surface not only happy path conversions, but explicit metrics on churn or error rates
- Create anomaly alerts for any metrics that have an anomalous drop/spike in the data
The metrics should be as close to real-time as possible, or refresh data up to an hour, but no more. There is no email or Slack message that is more important than first getting a handle on the business and your domain. The coffee dashboard should contain only the set of metrics that you care about — it shouldn’t list 50 charts in which you skip 30 of them. Be selective about which charts make the cut and ensure the controls you have in place allow you to filter and slice the metrics in ways that can help you triage any potential issues.
You can build a version of this in Looker, Amplitude/Mixpanel, or BI tools — but the importance lies in ensuring it is broadly accessible by members of your team so everyone knows how the product domain is keeping score of the feature impact, as well as a snapshot of the state of its health.
PMs who fail to do this don’t actually own their area — they should be the first to spot any issues as well as any opportunities. This is also a good opportunity to use these signals to get to the root of the problem on your own before looping in an engineer or product analyst. Good luck with instrumenting yours!