Preach. If success were just a matter of looking at numbers, we’d all be successful.
The data sets themselves can tell different stories. You mentioned a big one, something game designers know intimately, you can’t just build for the vocal minority.
yup, the vocal minority is always a danger. And tbh, my experience is that you can't really make them happy (nor will they leave!). It's the silent majority that you often want to cater to
Or you can make them happy... at the cost of losing the majority. Like, Paradox knows this and stridently refuses to implement things that would make their grand strategy games too hard, despite forums and mod workshops _bustling_ with those.
However, this might actually be more data-driven than it seems. Mod workshops give them a chance to see how many of the mod-using crowd (which is still not perfect but bigger and potentially more representative than the vocal minority crowd) want the hardships (because you can offset this against, say, people who just add cosmetic mods).
Brilliant! Delighted to see a VC leader acknowledge innate intelligence.
The Gut is the primordial brain and our connection to Divine Intelligence not just Artifical Intelligence. Once tech understands that there is more to intelligence than crunching equations, and that our biggest inspirations come from outside the box – when we're not at our desks, like Einstein on the train platform and Newton under the apple tree – a whole new world will open for our industry. Divine Intelligence is the true cloud, from which too many of us in our industry have become disconnected, or are microdosing in efforts to reconnect to while blocking our natural abilities. There is much in this field that can be measured, and an entire industry based on DI waiting to be explored that can be even more powerful than AI.
Thank you for always offering brilliant and inspiring information and ideas @andrewchen.
Yes Mona! Visionaries rely on their ability to intuit what's on the horizon...which is why many of them get fired from traditional corporate jobs. They use their innate knowledge and the ability to connect historical data dots in order to anticipate customer behavior + solutions to big hairy problems. I love this stack by @andrewchen - he calls us back to relying on our experienced Gut + our powerful imagination to help create and innovate for the future. You and Andrew have a lot in common!!
The data is just bits and bytes. It's always been what we choose do with it that make the difference - our edge is in the quality of decision making - and for that, being data-informed trumps data-driven more often than not!
So much to learn in the data driven platforms. watching you I’s and making sure you t’s are crossed. In this age of technology, with so much content, it hard.
Could not agree more. I'm currently drafting a new post (aka: self-therapizing) about the dangers of being purely data-driven in a zero-to-one consumer startup. Fresh wounds & hard lessons.
As a data analyst who worked for major companies, I can say that part of the problem is that the data product development process is really slow - it might be in line with the slow decision making process in corporations, but it's definitely too slow for startups. The time between discovering a business need for data and delivering a data product needs to be shortened.
This could possibly be solved by full scale deployment of self-service analytics, to the point that building data pipelines and dashboards becomes a basic business skill like MS Office. However, the technology is not there yet, and gaps in functionality must be filled with carefully crafted business processes, which makes it tricky to implement in organizations.
Data is a trailing indicator. As are so many other aspects of being a product manager, including user interviews, NPS, even ARR/Churn. Good PMs balance vision and future with day-to-day requirements, tech debt, architectural needs and investments, and defect resolution. Forever I have felt like there is room for product management to be art as well as science, with gut instinct / sniff test / occams razor / etc being an important reality. Great product managers are born not made, because of that ability to act on instinct.
Word. The best startup ideas are a result of a strong intuition about the customer/problem/market dynamic/technology solve. This is what drives 0-1 step function change in the market (and challenge to an incumbent). Somehow these same companies forget that as they mature into a bigger company.
Data driven = Incremental, optimization, generally right,
Great read! I agree in many of the points. It would be interesting to hear your take on how to get buy-in from stakeholders when there is not enough data (And when waiting for data would not be the best alternative)
Great post keep them coming. I just finished reading Geoffrey Moore's "Infinite staircase" and this post really resonated. Ultimately we are story telling creatures and we have to tell those stories based on the information we have available.
Truth. Having done a few 0 -> 1's now (both exits & failures), speed combined with intuition is the only way. And you'll find yourself constantly defending this approach.
Also worth mentioning that neither Facebook or Google was a 0 to 1. Both drafted off companies who came before them and were essentially optimizing existing solutions from day one. In Google's case, not only were there other search engines, but even their business model was a copy & tweak. Overture had educated the market (100k paying customers) and distributed sponsored search onto every other search engine before Google launched AdWords.
This meant that rather than create from scratch, their product development process began as an optimization. This probably explains why that culture exists (and works so well once you're already a big gorilla).
What data? Moat? Data that the providers benefit from. That the LLMs need to train on to learn from and monetize. For who exactly? What's the ideal atomic network data? I say it is niche circles that can be replicated 10000X
First you were on whatsapp, then telegram, now its the therealandrewchen. Just email me at haba1@optimum.net. Oh, BTW, I just got connected to the chief AI scientist at Luma AI that a16z funded. I have the app use case these newGen AI models need to enable the masses to create on mobile. Like multimodal tokenization.
Preach. If success were just a matter of looking at numbers, we’d all be successful.
The data sets themselves can tell different stories. You mentioned a big one, something game designers know intimately, you can’t just build for the vocal minority.
yup, the vocal minority is always a danger. And tbh, my experience is that you can't really make them happy (nor will they leave!). It's the silent majority that you often want to cater to
Or you can make them happy... at the cost of losing the majority. Like, Paradox knows this and stridently refuses to implement things that would make their grand strategy games too hard, despite forums and mod workshops _bustling_ with those.
However, this might actually be more data-driven than it seems. Mod workshops give them a chance to see how many of the mod-using crowd (which is still not perfect but bigger and potentially more representative than the vocal minority crowd) want the hardships (because you can offset this against, say, people who just add cosmetic mods).
Brilliant! Delighted to see a VC leader acknowledge innate intelligence.
The Gut is the primordial brain and our connection to Divine Intelligence not just Artifical Intelligence. Once tech understands that there is more to intelligence than crunching equations, and that our biggest inspirations come from outside the box – when we're not at our desks, like Einstein on the train platform and Newton under the apple tree – a whole new world will open for our industry. Divine Intelligence is the true cloud, from which too many of us in our industry have become disconnected, or are microdosing in efforts to reconnect to while blocking our natural abilities. There is much in this field that can be measured, and an entire industry based on DI waiting to be explored that can be even more powerful than AI.
Thank you for always offering brilliant and inspiring information and ideas @andrewchen.
Yes Mona! Visionaries rely on their ability to intuit what's on the horizon...which is why many of them get fired from traditional corporate jobs. They use their innate knowledge and the ability to connect historical data dots in order to anticipate customer behavior + solutions to big hairy problems. I love this stack by @andrewchen - he calls us back to relying on our experienced Gut + our powerful imagination to help create and innovate for the future. You and Andrew have a lot in common!!
The data is just bits and bytes. It's always been what we choose do with it that make the difference - our edge is in the quality of decision making - and for that, being data-informed trumps data-driven more often than not!
So much to learn in the data driven platforms. watching you I’s and making sure you t’s are crossed. In this age of technology, with so much content, it hard.
sure. I love profitability
Could not agree more. I'm currently drafting a new post (aka: self-therapizing) about the dangers of being purely data-driven in a zero-to-one consumer startup. Fresh wounds & hard lessons.
As a data analyst who worked for major companies, I can say that part of the problem is that the data product development process is really slow - it might be in line with the slow decision making process in corporations, but it's definitely too slow for startups. The time between discovering a business need for data and delivering a data product needs to be shortened.
This could possibly be solved by full scale deployment of self-service analytics, to the point that building data pipelines and dashboards becomes a basic business skill like MS Office. However, the technology is not there yet, and gaps in functionality must be filled with carefully crafted business processes, which makes it tricky to implement in organizations.
Thanks Andrew, great read. You could sum up “data informed” with Daniel Kahneman’s recommendation to delay intuition until you collect more data.
It’s particularly true for VC, where expert intuition doesn’t work. I demonstrated it in this article:
https://thevcfactory.com/venture-capital-investment-decisions-intuition-or-data/
Terrific read. Thank you for the inspiration- much food for thought.
Data is a trailing indicator. As are so many other aspects of being a product manager, including user interviews, NPS, even ARR/Churn. Good PMs balance vision and future with day-to-day requirements, tech debt, architectural needs and investments, and defect resolution. Forever I have felt like there is room for product management to be art as well as science, with gut instinct / sniff test / occams razor / etc being an important reality. Great product managers are born not made, because of that ability to act on instinct.
Your point about the “fog of war” in product decisions is spot-on. We often get caught up in data without realizing its limitations.
Word. The best startup ideas are a result of a strong intuition about the customer/problem/market dynamic/technology solve. This is what drives 0-1 step function change in the market (and challenge to an incumbent). Somehow these same companies forget that as they mature into a bigger company.
Data driven = Incremental, optimization, generally right,
Intuition - Step function, could be wrong
Awesome stuff!
Great read! I agree in many of the points. It would be interesting to hear your take on how to get buy-in from stakeholders when there is not enough data (And when waiting for data would not be the best alternative)
Great post keep them coming. I just finished reading Geoffrey Moore's "Infinite staircase" and this post really resonated. Ultimately we are story telling creatures and we have to tell those stories based on the information we have available.
Truth. Having done a few 0 -> 1's now (both exits & failures), speed combined with intuition is the only way. And you'll find yourself constantly defending this approach.
Also worth mentioning that neither Facebook or Google was a 0 to 1. Both drafted off companies who came before them and were essentially optimizing existing solutions from day one. In Google's case, not only were there other search engines, but even their business model was a copy & tweak. Overture had educated the market (100k paying customers) and distributed sponsored search onto every other search engine before Google launched AdWords.
This meant that rather than create from scratch, their product development process began as an optimization. This probably explains why that culture exists (and works so well once you're already a big gorilla).
What data? Moat? Data that the providers benefit from. That the LLMs need to train on to learn from and monetize. For who exactly? What's the ideal atomic network data? I say it is niche circles that can be replicated 10000X
First you were on whatsapp, then telegram, now its the therealandrewchen. Just email me at haba1@optimum.net. Oh, BTW, I just got connected to the chief AI scientist at Luma AI that a16z funded. I have the app use case these newGen AI models need to enable the masses to create on mobile. Like multimodal tokenization.