All related (13)
Tasha Alfano
Staff Product Manager, Libraries and SDKs, TwilioFebruary 10

The worst KPIs to commit to are the ones you can’t commit to at all. We can set targets and metrics and make dashboards, but that’s exactly what they are - targets. I recommend looking at past performance and trends within the data and setting a realistic yet aspirational target to work towards. After that, begin iterating on your target. Revisit the KPI, analyze, adjust, and communicate your findings.

Paresh Vakhariya
Director of Product Management, AtlassianMay 8

Some of the worst KPI's in my opinion are:

  • KPI's that cannot be measured correctly
  • KPI's that do not give a sense for the goal you are tracking. You can use the AARRR (Acquisition, Activation, Retention, Revenue, Referral) framework to understand the best metrics you can choose to align with your outcome/goal.
  • KPI's that are not achievable in a desired timeframe. Yes there could be exceptions here but generally these are not the best ones in my opinion.
  • Any KPI's that do not really tell you the health of the business unless a holistic picture is presented. e.g. number of app installs is not that meaningful without retention and engagement metrics
Virgilia Kaur Pruthi (she/her)
Principal PM Manager / Product Leader, Microsoft | Formerly AmazonJanuary 31

This could really range based upon the company, your users, your target goals, where you are in your business lifecyle, etc.

The most basic ones are: acquisition, activation, retention, revenue, referral

You could also be measuring customer lifetime value

In regards to the worst KPIs, honestly those that cannot be discretely measured and tracked over a specific time period. Vanity metrics (e.g. the number of views from a marketing article or number of shares of a post) really add no value.

Jacqueline Porter
Director, Product, GitLabJuly 12

The worst KPIs are vanity metrics that have no ties back to actual adoption or business metrics. I once had a product manager commit to hitting a number of emails a notification system was supposed to send in a 30-day period. Without context, this seems like a great metric to track for volume, except the total count of emails tells you nothing about how many people are getting value, if they are getting value, if they are recurring users, or if the emails are contributing to user satisfaction. This can certainly be a metric in the toolkit, but not a KPI for a product line. 

Nico Rattazzi
VP of Product, ZumperFebruary 20

KPIs around delight unless this is your key product differentiator (which is proven to be compelling to customers). Focus on building an intuitive and effective product experience that users would want to recommend to their friends/colleagues. Focusing on the final pieces of polish such as interactions, delight, animations, etc are fluff until you're really providing value to your customers. This is why keeping your KPI or success metrics concise and essential will allow you to provide the most impact to customers.

Farheen Noorie
Director of Product Management, Growth and Monetization, ZendeskApril 18
  1. Rates: To me without absolute numbers, rates may paint a false picture. Let me explain with an example. Lets say you have a trial experience for your product and you are responsible for the cart experience and thereby conversion rates which is measured by number of paid customers/number of trialers. I would suggest that instead of rates the north star metrics should be a combination of number of paid customers as well as Average Deal Size (ADS) per paid customer. A conversion rate is a good number to track but may lead to wrong hypothesis when you see not normal trends in either your numerator or demonimator. In this example, conversion rate will go down if the number of trials significantly increases or decreases. If you are responsible for rates make sure that you own both the denominator and the numerator of that equation in order to truly able to influence a positive change
  2. Fuzzy metrics: As the say "What cant be measured, cant be managed". Metrics that are not explicit like Customer Happiness and NPS are not possible to measure. I'd go a step further to challenge why should one even measure those metrics and are there any core metrics that can be measure successfully. An example would be instead of measuring Customer Happiness, lets measure Customer Expansion because happy customers expand.