High ROI Data Science

Share this post
The Foundations Of Excellent Mondays
vinvashishta.substack.com

The Foundations Of Excellent Mondays

Vin Vashishta
Jul 25
2
Share this post
The Foundations Of Excellent Mondays
vinvashishta.substack.com

I have gotten excellent questions this week from several subscribers. I am working to quickly turn those into posts while also maintaining the core series content. The likes have also been rising, which tells me the improvements are starting to bear fruit. Thank you all so much for subscribing and putting effort into improving the quality of content.

If you've been here since the earliest days of this newsletter, you might remember the first set of improvements: grammar and post flow. I started using an editor and Grammarly to clean up the typos and mistakes. I shortened most posts to between 1200-1800 words, making them easier to read. Those were some pretty dull fixes for a newsletter that covers some innovative concepts.

A recent subscriber suggestion was to build a series around challenges for first-time leaders. I started a series about advanced leadership concepts that was not as well received as this. I wrote another post in the first-time leadership series for this week about dealing with imposter syndrome. Again, not a leading-edge concept.

Another early change I made was to start at a lower level for the causal, decision science, ontologies and applied innovation series. I went too far forward without covering the basics. I discussed forward-looking applications without explaining how to get there. I used advanced use cases but neglected to explain the more straightforward, everyday applications.

Even my series on decentralized platforms starts with boring use cases for NFTs like managing metadata associated with physical assets. I have worked on some of the most advanced business applications of data science but rarely use the most advanced deep learning models. I have used natural language for advanced use cases but am just looking at implementing my first LLM right now. I am leaning towards recommending a simpler approach, so I may push that further into the future.

I have a 70/30 rule about fundamentals vs. innovation and bleeding edge that is obscured by my content's focus. My favorite baseball team, the San Francisco Giants, shows what happens when those numbers are flipped. They focus 70% on innovation but only 30% on fundamentals.

Gabe Kapler is the team's manager. Last year, the team won 107 games which is a massive achievement. Much of their success was due to an innovative approach to hitting. Kapler has taught the team to extract as many pitches as possible at each at bat. That requires hitters to be disciplined and only swing at high-quality pitches.

Starting pitchers throw more pitches and leave the game sooner as a result. The opposing team is out on the field longer and prone to more errors as they fatigue. The Giant's pitchers work quickly. They use very little time between pitches and are on the field for as little time as possible.

He has completely transformed how every player works when they come up to bat and how pitchers work. A credit to his leadership comes from every player's buy-in to this new system. They trust a new approach in a sport that is slow to embrace change. It is innovative and effective.

Last year, the Giants had an older team of veterans. This year, the team is younger. Kapler has spent 70% of his time with young players, training them on the new system. Why? They make errors on basic plays. They miss throws to get outs. They fail to make simple plays and catches. The young players have not learned to play baseball. They spent all their training time learning to innovate.

Farhan Zaidi is the Giant's President, and he too is an innovator. He is responsible for acquiring new players and building the team. His approach is to bring in top players on short-term contracts. It's smart because it limits the risk of a player who doesn't fit the team or someone who takes a significant dive in productivity. With a 5 year contract, the team can be stuck with a mistake for several years. With a 2-year contract, that risk is managed.

This year, the team lost one of the greatest baseball players of all time. Buster Posey was a team leader and the core of a great team. People came to the team just to play with Buster. Brandon Belt, Brandon Crawford, and Madison Bumgardner were also core team members who won 3 World Series titles together.

The new approach to player contracts lost Madison to another team because the Giants were unwilling to pay him what he deserved. Buster retired, and his replacement is very new to the major league. He is not ready to fill a Hall of Famer's shoes. The Brandons remain, but they are not enough of a core to build a team around.

Building short-term, low-risk is a good strategy, but teams are built around a core. Leadership must be willing to take chances on players they can build a team around. Great teams are built with long and short-term approaches.

The changes required to turn the team around are feasible. All is not lost, but this season probably is. Will Gabe and Farhan see the flaws in their approaches and make the right improvements? If they bring me in as their new VP of Baseball Analytics, of course they will. ***Call me Farhan***

The greatest innovations fail without execution. Companies that execute in the short term without a long-term growth plan also fail. 70/30. 70% fundamentals and 30% innovation is a critical balance.

How do we maintain a vision for the future while managing the fundamentals? Stay close to users and the data team. I build communities and coalitions around the long-term vision. I keep people engaged by delivering incrementally and continuously improving.

Fundamentals are 70% of innovation. The data pipeline is critical to everything else. I can spout on about building ontologies, but if the team cannot deliver data to models for training, who cares!? Data gathering, storage, and serving are mundane, critical capabilities. We need to get so good at the data phase that most of it is automated, and the process is highly reliable.

Causal methods are the future, but they are meaningless if the team cannot deliver a model to production and support it once it is there. The machine learning platform, integration, release management, and continuous improvement are essential to innovation.

Technical strategists partner with senior leadership and the business to help transform the business and operating models. They define the technology model and implement all 3, but none of that matters if the business lacks capable senior leadership. The best data science team fails without excellent business leadership.

Expert knowledge starts the process of building ontologies and applying causal methods. The data team cannot access expert knowledge if the business lacks domain experts. The innovation rocket engines are stuck on Earth without a guidance system or steering mechanism. Spending all our budget on pilots and pinching pennies on the navigator leads to excellent flying that misses the moon.

I am known for my vision but hired for my fundamentals. We all are. DALL-E2 gets all the likes, but I use regression as the starting point for many causal-based projects. Deep learning comes later, and those models have limited utility. They serve the process of discovery and indicate directions for study. The highest value process is vanilla experimental design.

My 9-figure project series started with a basic topic, incremental innovation. MySpace was the disruptor. Facebook was the fast follower that brought incremental innovation to social media. Meta is the disruptor, and I am looking for the fast follower. Hopefully, they are too, or they'll follow IBM and Microsoft's footsteps of losing their market leadership position to fast followers.

How did Microsoft regain dominance? They got closer to users, focused on building internal execution capabilities, and spent 30% on innovation.

Companies will overreact and pull back too much as they enter our current downturn. Budgets will be 100% fundamentals. They will slim down and excel at executing on their core business…of last year. They will spend the next 18-24 months building what they needed in 2021.

When the downturn ends, that stagnation will put them out of business. Surviving the downturn will be their undoing. The 30% innovation forces me to build today to support tomorrow. Investing in long-term growth is just as critical as focusing on fundamentals.

Transformation is continuous, and the waves are much closer together. The old paradigm of building what's necessary now and forklifting later is unsustainably expensive. The fundamentals must be made to scale and generalize to support the next 5 years' worth of transformations. Vision and innovation are part of today's fundamentals and execution.

The community is critical to improvement. The closer I am to you, the better my fundamentals become. Your feedback is essential, and a single suggestion can benefit everyone significantly. This, too, is overlooked in the innovation cycle. Ecosystems build, learn, and improve faster. That is the core disruption of Web3 and why I am building now to support the new paradigm.

Over the next year, our community will expand into an ecosystem. This core will always be essential. We will always be the 70% who drive innovation. You are the present and future. I do not scale, but this community will. I look forward to seeing the amazing things you will all build and where you take the foundations.

Vin

Share this post
The Foundations Of Excellent Mondays
vinvashishta.substack.com
Comments

Create your profile

0 subscriptions will be displayed on your profile (edit)

Skip for now

Only paid subscribers can comment on this post

Already a paid subscriber? Sign in

Check your email

For your security, we need to re-authenticate you.

Click the link we sent to , or click here to sign in.

TopNewCommunity

No posts

Ready for more?

© 2022 Vin Vashishta
Privacy ∙ Terms ∙ Collection notice
Publish on Substack Get the app
Substack is the home for great writing