Alternative Data Weekly #198
August 9, 2024
Thanks for being here!
Announcement(s):
Thanks to our sponsor Amplify Data!
One size doesn't fit all - Personalize your data feeds with Amplify.
Theme that emerged in this week’s email is … the right data is key, start with the problem to be solved rather than taking your solution to find a problem; creating clarity is a skill.
QUOTE
“There will be models that we settle in many, many models, but for different use cases. And we think that data is going to be the key differentiating factor.“ – Doug Peterson, CEO of S&P
News Articles
Podcasts
Cool Charts
Final Thoughts (Actually using AI)
#1 – Several contributors from McKinsey published From raw data to real profits: A primer for building a thriving data business. July 2024.
My Take: I spent a few years talking to dozens of companies about the monetization opportunity with their proprietary data. Many were willing to listen, but interest waned when the value was typically lower than expected, the effort was greater than expected, and the compliance concerns came to light.
New technologies make it A LOT easier to prep the data, new markets will (potentially) increase the market size/value, and as selling data becomes more common, compliance people will get comfortable(ish).
#2 – Financial Data Transparency Act Joint Data Standards. July 2024. Plus subsequent response from CUSIP: CUSIP Global Services Statement on Proposed Data Standards for the Financial Data Transparency Act. August 2024.
My Take: I am no expert in this area (h/t To Tim Baker for highlighting this), but find it fascinating as I read and learn more. These types of standards are important and often taken-for-granted. The incumbents controlling the data have great businesses. It sounds like FIGI (the proposed open-source replacement of CUSIP) has shortcomings, but if the regulatory bodies are getting behind it, momentum will build. Will be interesting to watch.
#3 – Alex Boden’s Thoughts on Artificial Intelligence in Data & Analytics. August 2024.
My Take: Rather than treat AI as a hammer looking for a nail, it is better to understand current workflows and apply AI in a way that enhances or augments (rather than replaces) the worker.
BONUS: Doug Peterson of S&P S&P Global Inc. (SPGI) Stifel 2024 Cross Sector Insight Conference (Transcript). August 2024. “There will be models that we settle in many, many models, but for different use cases. And we think that data is going to be the key differentiating factor.“
What else I am reading:
Interactive version of Matt Turck’s MAD diagram. March 2024.
Timothy Lee’s Predictions of AI doom are too much like Hollywood movie plots. August 2024.
Matt Ober’s Should Factset go Horizontal? August 2024.
Matt Ober’s AI is a race to zero, what does that mean for data?. August 2024.
Jason Derise’s Psychological Safety and Learning in the Face of IT Crises: Insights from the CrowdStrike Outage. August 2024.
Adam Braff’s Marry me? Not yet, she replied Swiftly. August 2024.
Isabella Bousquette of WSJ published Morgan Stanley Moves Forward on Homegrown AI. July 2024.
Source: Lenny’s podcast interviews DoorDash’s Jessica Lachs. July 2024.
“Clarity above all else” – Lenny at 1:05:35
My Take: Really enjoyed this interview. Jessica built the data team at DoorDash (interesting non-traditional background … minute 29:00). Interesting frame of thought is that the data team is (should be) an impact driving function … not just service function. Answering the “why”, not just the “so what”.
Analytics embedded in business team vs central. Jessica likes central data team. Good discussion here. I tend to agree, but find myself easily persuaded.
One of many things I found interesting was the idea of being intentional about setting aside time to think and brainstorm and come up with new ideas & think long-term. This is harder than you’d think to do.
This idea of creating culture of extreme ownership … set the expectation. Solve a problem vs this is my job & I only do this one thing. We at ModuleQ call this a “jobs to be done” environment.
Find short-term metrics that drives a long-term initiative. Simple metrics … focus on only what matters? See Lenny’s quote above.
Highlights (79-minute run time):
Minute 01:50 – interview starts
Minute 05:00 – Central vs embedded data teams (reporting lines, not necessarily goals)
Minute 10:00 – thought partners with partner teams in product, mktg, etc.
Minute 11:00 – benefits of central team (talent, consistent metrics & methodologies, culture/brand)
Minute 15:00 – being proactive; the benefits and challenges
Minute 24:00 – talent and the benefit of curiosity
Minute 29:00 – Jessica’s background
Minute 34:00 – the early days of DoorDash
Minute 40:00 – how did early team create culture of extreme ownership
Minute 44:00 – keep things simple, what metrics are you focusing on and what long-term goals are they supporting?
Minute 60:00 – managing global team (more similar than different)
Minute 63:00 – how are they using AI?
Minute 68:50 – Lightning round (book recommending: The Rose Code, Kate Quinn & Libby App)
Source: Seattle Data Guy’s The Art of Execution: Making Things Happen as a Data Leader. July 2024.
My Take: Love this idea of being clear. It is so difficult to clearly articulate what you are trying to accomplish. Particularly in areas like data/AI where things are changing so rapidly. Focus on high impact & quick wins…and the ability to clearly communicate the what & the why is an underrated and undervalued skill.
Source: McKinsey’s From raw data to real profits: A primer for building a thriving data business. July 2024.
My Take: Creating new data, analytics, and AI platforms business is a popular idea.
Source: Jeremy Merrill & Rachel Lermna of The Washington Post published What do people really ask chatbots? It’s a lot of sex and homework. August 2024.
We are just getting started with how people use GenAI.
In my former role I used ChatGPT nearly every day to help my terrible SQL programming skills become average.
While I have not used ChatGPT to build this weekly newsletter, I have used it to help draft long notes, summarize articles … I have tried to find a way to consistently use GenAI in my daily life to perhaps automate a rote practice, but I have yet to do so.
That said, I do use ModuleQ’s UnpromptedAI to prep for meetings and stay abreast of topics of interest (that is how I find some many relevant articles relating to Alt Data).
I am sure the economy will go through an “AI downturn” as many AI use cases are not quite ready for primetime…but unlike Crypto (which is still largely looking for “use case” beyond being a trading vehicle & volatile “store of value” … can you be both volatile & a store of value?) … BTW, I still love crypto …. AI most definitely has real world use cases that will make knowledge worker’s lives clearly better.
