A Smart Bear
 

Articles

Long-form articles, easy to read on any device, including print.

The practical application of "Rocks, Pebbles, Sand"
You’ve probably seen this analogy before, but many real-world complications get in the way of actually doing it. It’s not just about how big tasks are, but about using different frameworks to prioritize each, and following certain guidelines about how to construct sprints.
Fermi ROI: Fixing the ROI rubric
“Maximum value in minimum time.” Sounds good in theory, but traditional rubrics surreptitiously fail to produce the best answers, and fail to create explanations that help others understand why they’re the best answers. This system works.
Binstack: Making a maximal multi-dimensional decision
Rubrics are often used to select the best option in a multi-dimensional decision space. However, they often do not clearly identify a winner, nor do they result in an explanation of the decision that is easily communicated to others, especially those whose favorite option was scored close to the winner, but was not selected. Binstack is a solution.
The Elephant in the room: The myth of exponential hypergrowth
Fast-growing startups are frequently described as “exponential,” especially when the product is “viral.” Turns out, this is incorrect, even for Facebook and Slack. If you have an incorrect model, you don’t understand growth, which means you can’t control it, nor predict it. Here is a different model to understand how companies actually grow.
The "Talk vs Walk" framework
This exercise we invented at WP Engine is surprisingly useful in engaging both Marketing and Product, generating actions for both sides that make products more desirable and competitive.
The Iterative-Hypothesis customer development method
A simple but effective system for generating insights about how your potential customers think, what they need, and what they’ll buy. This method has been used both to reject startup ideas and to validate WP Engine before it had any customers (it is now a Unicorn).
Distributed Logical Time
Properly ordering events in time is notoriously difficult in distributed systems. This algorithm is a simple, decentralized, scalable, constant-memory mechanism for independent replicas to record events in time, such that “happened-before” is preserved in almost all cases.