Alternative Data Weekly #243
Theme: Data Wars - The Rise of The Proprietary
Special thanks to our sponsor Babbl Labs.
Winners of the BattleFin NY New Data Provider Showcase Competition 2025.
New: Extracting Alpha and Insights from YouTube Financial Video Content
QUOTES
“As AI becomes more and more specialized, thousands of these proprietary deals will need to exist (sometimes augmented with human-generated data). While the issue to date has been illiquidity of this information & difficulty to license it at scale, as the data shortage for AI becomes more acute more and more of these deals are being structured.” Travis May
News
Pods
Charts
Final Thoughts (from Checks to Streams … Market Research Disruption)
#1– Travis May published Meta’s Bet on Scale AI Is Just the Beginning of the AI Data Wars. June 2025.
My Take: “Data Wars”. I like the way that sounds. There has been a massive increase in demand for data. Unfortunately, that increase in demand has not seemed to result in an increase in data price because the growth of data supply exceeds data demand.
There is an argument that the value of readily available data (either public data or easily bought) will go to zero. Value will accrue to the “massive amount of human knowledge in the form of proprietary, non-crawlable information.”
How does this type of specialized data get most effectively monetized? This remains to be seen.
#2 – Shaili Guru published Can You Prove Your AI Moat?. June 2025.
My Take: The #1 way to prove your moat = proprietary data. Defined as “exclusive, flowing, usable data that competitors cannot easily replicate.”
See the above article for the growing value of proprietary data.
I like the framing that, in addition to the current metrics, the question is how value compounds over time, so your moat is constantly growing with each iteration (flywheel). “If your AI stack is getting more innovative, stickier, harder to replace, and more trusted with every cycle, you’re on the right path.“
#3 – Analytico published How to Thrive as a Data Analyst in the Age of AI. June 2025.
My Take: I like the author’s optimistic view of the data analyst role. Find a few tasks to automate (Chapter 3), freeing up time to focus on higher-value tasks like building domain expertise and defining success metrics. spend your time doing more valuable & engaging work.
“When you automate the tedious, you create space to master the meaningful. That's the shift.”
What else I am reading:
Gabriel Daros published In a world first, Brazilians will soon be able to sell their digital data. June 2025.
Saed Alshaer & Mohdsalam Alhomran of PWC published Data monetisation and beyond: Redefining the economics of data. 2025.
Letter from Aiera on Our Series B. June 2025.
Katie Parrott published Every CEO Is Writing the Same AI Memo. Here’s What They’re Really Saying. June 2025.
Nomad published How Investment Firms Unify Alternative Data Management. June 2025.
Elor Arieli published When the Model Isn’t the Problem: How Data Gaps Undermine AI Systems. June 2025.
Source: The Analytics Engineering Podcast interviewed Lonne Jaffe The history and future of the data ecosystem. June 2025.
My Take: This is a good historical background on the data ecosystems on how we got where we are, and why. Plus, any podcast that mentions Jevons paradox twice, is likely a good one.
Of most interest (to me) was the conversation around “sticky” legacy systems (minute 29:00).
Why do customers stay with outdated software?
Switching costs
Hard to re-write jobs & tasks onto new platform
My take: if it ain’t broke, don’t fix it
Does AI reduce switching costs?
52:00 – Pitchbook mentioned; workflow is going to change … all the data the knowledge workers can access should be accessed and utilized to get the best result (we do this at ModuleQ!).
Highlights (54-minute run time)
Minute 01:45 – interview starts; Lonne’s background
Minute 05:00 – Sorting history
Minute 07:00 – M&A as an innovation tool
Minute 10:00 – Lonne’s academic background & juggling skills
Minute 12:00 – some history starting from the 1970’s
Minute 20:00 – Jevon’s paradox in data & compute
Minute 25:00 – jobs to be done … the rise of data prep
Minute 29:00 – sticky old software (old tools are actually good, proven, durable)
Minute 38:00 – data integration AI features; “medium tier” companies are winning here
Minute 41:00 – High elasticity of demand (TAM expands as prices drop)
Minute 43:00 – Structured data access for non-technical users is finally viable.
Minute 50:00 – “Birthright” tools
Source: Nvidia published State of AI in Financial Services: 2025 Trends. 2025.
Bonus: ESMA published LEVERAGING LARGE LANGUAGE MODELS IN FINANCE: Pathways to Responsible Adoption. June 2025.
Everyone says within 5 years….
Zack Cohen & Seema Amble published Faster, Smarter, Cheaper: AI Is Reinventing Market Research. June 2025.
Market research is going to change.
AI is certainly changing the way information is distributed and consumed.
But the real opportunity is in how information is collected.
Where we once relied on episodic surveys (broad & shallow) or expert interviews (narrow & deep), AI enables continuous, intelligent data collection that is both broad and deep.
Historically, you had two options when collecting first-party intelligence.
Broad & shallow survey (ie, survey hundreds or thousands of people asking 1-2 questions each).
Narrow & deep “channel checks” (ie, interviewing a small number of experts)
AI delivers the opportunity to do both.
Q: How?
A: AI Personas.
AI Twins. These allow you to create a persona that you can constantly quiz. For example, you can create the persona of a “42-year-old mom who lives in Cleveland and shops at Kohls, Target, Kroger, & Whole Foods” Adding the additional color of income, zip code, marital status, etc. This persona can be trained with a host of information and quizzed constantly (broad, constant, & now with the ability to go deeper with questions).
AI experts. This persona can interview dozens of nichey B2B sellers and share the information back to the community. Similar to a Boardy-type service, except the AI-expert-persona can be trained to know everything about even the smallest niches & dynamically know the next best question to ask. Experts can asynchronously have these conversations (deep & now with the ability to go broader).
The collection gets 10x better (broader & deeper) and 10x cheaper (less human involvement).
The real promise of AI.
Am I nuts? Who is doing this?