Alan Klement
@alanklement
Followers
5K
Following
22K
Media
618
Statuses
4K
Product Market Fit Coach. I will teach you how to create and predict product-market fit.
New York City
Joined January 2009
I just started a JTBD 2.0 and Go to Market Prototyping Community. Why Join? You can ask questions suggest ideas that we can discuss in depth without polluting your public timeline. (at least, it seems that way) https://t.co/4g8DRPgx00
1
0
14
The aim of product and market Experimentation should be: Predicting effects across time and conditions The aim should never be "statistical significance," which is about stating whether a particular difference exists. All data sets have differences, and findings differences
0
0
0
Common misconception about experimentation that people who sell A/B testing don't want you to know: The claim: "Statistical testing determines whether differences are probably not due to random chance” Is wrong and unhelpful. A p‑value quantifies how incompatible the
0
0
2
“Statistical significance of B over A thus conveys no knowledge, no basis for action.” — “On Probability as a Basis for Action,” The American Statistician (1975).
0
0
0
Worthy Read: AI Can Synthesize Data for You—But Should It? by Christina Wodtke "When AI does the synthesis, you skip an essential part of the process. You don't inhabit the data." 🤖 Want to build better product intuition? It's not about AI-generated reports—it's about getting
2
3
16
But what if you're shipping the wrong thing? Shipping the right thing slowly > shipping the wrong thing fast
Ship fast, ship small, ship frequently
0
0
5
Be careful about LLMs. They get things wrong. I had to correct it on entropy
0
0
1
I asked ChatGPT to critique my chapter. Looks like I'll keep my subscription!
1
0
4
Closure is necessary for health systems. A systems cannot maintain cohesion (or struggle to) when it's boundary conditions are in transience (good fences make good neighbors) cc @normonics
1
4
18
Don't make this mistake. By definition, a "black swan" cannot be predicted. Otherwise, it wouldn't be a black swan.
1
0
2
One reason why A/B tests are neither sensitive enough to detect effects, nor robust enough to predict impact. When dealing with complex systems, simple rules outperform statistical inference. E.g. "feature umbrellas when it rains; otherwise feature sunscreen"
0
0
0
Can someone please unplug the United States and plug it back in? We need to reset things.
1
0
6
Anyone care to comment? Interesting stuff here. Maybe Jobs theory can help these people...
As I've been mentoring PMs (and other functions) in my network recently, I'm hearing from a lot of people who are having a career crisis - folks have just hit a wall and aren't just looking for a new job, but a new approach to work 🧵
1
0
1
As I've been mentoring PMs (and other functions) in my network recently, I'm hearing from a lot of people who are having a career crisis - folks have just hit a wall and aren't just looking for a new job, but a new approach to work 🧵
3
4
34
Some solid criticisms of Ariely and Nudge—from 2012! https://t.co/fQ7vfnENU4
0
1
8
Interesting paper that gives a mathematical definition of emergence, and talks about other emergent-related phenomena cc @normonics
0
1
2
The new stand along application from @OpenAI is a great product move (at least for me): 1. Because I can launch it with a keyboard shortcut, it becomes more integrated with my workflow. 2. When I used it in the browser, I also used it along side Google's Gemini. They would
0
0
2