Alternative Data Weekly #240
Thanks for being here!
The theme that emerged in this week’s email is … Big Data —> Big Context?
Special thanks to our sponsor Babbl Labs. Check out their offering!
New: Extracting Alpha and Insights from YouTube Financial Video Content
QUOTES
“Data and AI are no longer support functions—they're now core to business strategy and execution. AI is optimizing global supply chains, predicting medical outcomes, and designing next-generation products before a single prototype is built." - David Steier, CMU
News Articles
Podcasts
Cool Charts
Final Thoughts (Low-Hanging Fruit)
#1 – Pedram Navid published All I want in AI is some context and a chat window. May 2025.
My Take: #contextisking. The promise of AI is that it, using the MASSIVE amount of data to which we have access, will make our lives better. But this only happens in the context of our lives. Protocols like MCP (see below) promise this sort of understanding of our very own individual, very personal, context.
#2 – JaserBK published Your AI Product Doesn’t Need a Data Scientist. May 2025.
My Take: This is the first in a series of coming articles. It takes a look at how & where Data Scientists can add value with AI products. A big key is understanding the business, and having domain knowledge & understanding the context of a question and the workflow being impacted by an AI product.
The person who can be the bridge between technical & business is an increasingly important role.
#3 – Jason Derise published Two Minutes Ahead of the Future. May 2025.
My Take: Speed of signal delivery makes a difference. But if not done well (data quality, etc), speed just helps you make mistakes faster. “It’s just a faster way to lose capital with more confidence”.
PS I had a similar experience with the Cavs - Pacers Game 2, where I left early to beat the rush. I thought this was smart … 30 seconds to go, up by 5, the Cavs with the ball. The two-minute walk to the car was enough time to catch the local radio guys bemoaning the loss.
BONUS: Sambeet Parija from Aetheron Lab published MCP for Dummies: Why Everyone’s Talking About It. May 2025. “MCP is a big step toward making AI usable, not just impressive. It’s a practical foundation for connecting AI to the rest of your world. You don’t need to understand the code behind it, but if you care about how AI will show up in your work and life, this is one of the protocols that will quietly power it.”
What else I am reading:
Salesforce Signs Definitive Agreement to Acquire Informatica
Adrian Krebs published AI Agents: We Need Less Hype and More Reliability. May 2025.
Alan Reed published Why Reliable Trading Hours Data is Necessary. May 2025.
2025 Dupont survey What does the public think about AI?
Homeland Security and Governmental Affairs Committee (HSGAC) Majority Staff Report Hedge funds’ use of artificial intelligence in trading . June 2024.
Randy Bean published Carnegie Mellon University Is Training A Next Generation Of Data & AI Leaders. May 2025.
Joe Wilkins published Professors Staffed a Fake Company Entirely With AI Agents, and You'll Never Guess What Happened. May 2025.
Distinctive Insights published Intelligence Vault: Vendor Zoom – Investment Research AI Methods. May 2025.
Saeed Amen published Neudata New York conference 2025. May 2025.
Peter Baumann published Things not to do as a Data Influencer on LinkedIn. May 2025. (I think I have steered clear of breaking the rules!)
Source: Trading Insights: All About Alternative Data. Elouise Goulder of JP Morgan’s Making Sense podcast interviews Mark Fleming-Williams of CFM.
My Take: Mark Fleming-Williams also hosts the Alternative Data Podcast. The interview provides interesting insights into the world of buying alternative data. At CFM, Mark has created processes to maximize the data ingestion workflow (investigation, testing, legal, negotiation).
AI is changing everything … the way data is collected and made valuable makes alternative data more accessible to a wider range of potential buyers.
Attributes of data:
Long history (3-5 years minimum).
Wide coverage (large number of tickers).
Point-in-time coverage (the data has not been changed since the day it was created).
A few other notes:
We are buying after testing for a few months ... this is de-risking the transaction for buyer...no such thing as “free trial” … buying investing A LOT of time and energy into evaluation.
Renewal rate of quant funds is 95%, but getting “on the train” first is really hard.
Highlights (23-minute run time)
Minute 01:30 – interview starts & Mark Fleming-Williams background
Minute 03:00 – the role of Head of Data Sourcing (team of 3)
Minute 06:45 – are free trials the “norm”?
Minute 09:30 – what are the attributes that are deemed critical
Minute 14:30 – history of the “Alternative Data” space
Minute 17:30 – AI is changing everything
Minute 19:30 – Mark’s best guess as to what is coming with AI (ex., easier to create a scraper)
Source: Monda’s The State of Data Delivery 202
Latest trends in data delivery & sharing from global data providers.
Source: Dan Goldberg published Selling data and analytics to Wall Street. May 2025.
Low Hanging Fruit
How I Built This: Substack: Chris Best and Hamish McKenzie).
Sometimes it is better to go after the low-hanging fruit. Half-baked solution that solves one problem exceptionally well. The Substack guys identified a specific problem (creators didn’t own their audiences & it was hard to get paid directly). The first version “was like curing cancer” for a single big creator. This first customer moved 100k subscribers over in a week.
This was V1.
It was two more years of iterating before they had Substack as we know it today, but the key was finding “low-hanging fruit” for a single customer…and then using that framework to build something for the masses.
Check out the cartoons from the Marketoonist.
Thanks for including my article. On the plus side for your story, you didn’t have to see them blow the lead first hand and then sit in traffic for a longer time. I think opposite would be worse, thinking it’s a sure loss and leaving but then missing a dramatic win in person :)