(B) I have been recently thinking how to propose KPIs for an infrastructure-oriented product versus KPIs for a consumer-oriented product, and how they might differ.

A Key Performance Indicator (KPI) is ideally an easy-to-measure indicator of progress toward a product goal. KPIs should have targets, have a specific timeframe, and be related to a given product goal. A KPI helps you understand how a particular product, service, or feature is performing over a given period. KPIs that analyze the past performance and health of the product are lagging indicators, while KPIs that analyze the future performance and health of the product are leading indicators.

In the end, a KPI should be used as a compass to navigate when leading the market, anticipating or reacting to user behavior changes or new competitive threats. It helps to understand whether the product team is taking the right path towards its strategic goals.

The following are widely known and used KPIs and metrics for a consumer-based product, feature, or service…

KPIs to forecast the business success of a product:

• Monthly recurring revenue (MRR):
  • MRR = Revenues gained (from new users + user upgrades) – Revenues lost (from churned + user downgrades)
• Average revenue per user (ARPU):
  • ARPU = MRR % Total number of users
• Product gross margin:
  • Product selling price – Product and acquisition costs (without indirect fixed costs)
• Customer lifetime value (CLTV):
  • ARPU * Average user lifetime
• Customer acquisition cost (CAC):
  • Marketing and channel spendings for a given period % total of user acquired during that period

KPis and metrics to monitor user engagement:

• Monthly/daily active users (MAU/DAU):
  • Number of users active over a certain period of time
• Average number of sessions per user
• Average session duration per user
• Average number of user actions per session
• Bounce rate:
  • Percentage of users who visited only one Web site or mobile app page
• Retention rate:
  • (Number of users at the end of the period – Number of new users)/ Number of users at the beginning of the period
• Churn rate:
  • Number of users lost % Total number of users

KPIs to measure user satisfaction:

• Net promoter score (NPS):
  • Promoters (rank product 9 and 10 on a scale of 0 to 10) – Detractors (rank product from 0 to 6 on a scale of 0 to 10)
• Customer satisfaction score (CSCAT):
  • Average ranking (on a scale of 1-3, 1-5, 1-10) per user

For an infrastructure (or platform) product or service, there are also a number of metrics that are widely used, and in particular to guarantee service level agreements (SLAs) such as:

High availability:

• The system must be available e.g. uptime for 24 x 7 for 365/366 days
• Measure: x =  (n – y) * 100 / n (n = total number of minutes in a month – y = total number of minutes the system is unavailable)
• Marketing for every system: ‘five 9s’ (99.999 percent) availability (but never achieved by anyone!

Resiliency:

• Acceptable level of service in the face of system faults and performance
• Often implemented in good system design: where are the expected system failures and how to provide the service around them
• Examples:
  • The map-reduce algorithm @ Google – processing for searches are distributed over a large number of servers – if one server fails – the system can still provide the search to a user (but require planning more servers than needed)
  • The Internet: if one service provider link fails, packets can be re-routed to another link (but require planning more links than needed)

Scale:

• Number of users that the system must support
• Generally stated as a goal to support a number of monthly active users (MAU): 10 million, 100 million, 1 billion
• The latest generation of system auto-scales: for instance, Kubernetes optimize the workload between the resources to meet the demand

Latency:

• End-to-end response time between user service request and fulfilled request provided to the user
• Generally accepted metrics:
  • Less than 100 milliseconds: great user experience
  • Between 100ms and 200 milliseconds: good/acceptable user experience
  • Over 200 milliseconds: degraded user experience

Security:

• Now usually specified as a number of data breaches
• Number of user accounts that hackers did access

Cost:

• The only metrics that is usually always measured!

Both consumer-based and infrastructure-based KPIs have the same goals – ensuring that the product or service delights the end-users! KPIs for consumer products or services aim to make sure that the service fulfills the direct needs of the users, while KPIs for infrastructure services aim to make sure that the service enables other consumer or infrastructure services to operate. In the end, the main difference between the two is the data sources. Consumer-based KPIs generally are developed mostly from user interactions within the service itself, while infrastructure-based KPIs are developed from sources within the infrastructure itself for instance:

• KPI for product adoption: increasing the number of services using daily the platform by 50% over the next 6 months
• KPI for product availability: the total hours of outage per month for the platform shall be less than 4 hours
• KPI for product costs: the platform shall operate for less than $50,000 a month.

Note: The picture above are some California native flowers.

Copyright © 2005-2020 by Serge-Paul Carrasco. All rights reserved.
Contact Us: asvinsider at gmail dot com.