Alternative Data Weekly #241

Theme: ROI Needs Some Context & Context Needs Some ROI.

John Farrall

Jun 06, 2025

Special thanks to our sponsor Babbl Labs. Check out their offering!

New: Extracting Alpha and Insights from YouTube Financial Video Content.

QUOTES

“LLMs are like pets — they need to be fed properly” - Mehdio

News
Pods
Charts
Final Thoughts (Coaching Impact)

My Take: Some weeks I struggle to find articles worth sharing — this was not one of those weeks. In fact, there were almost too many good articles to choose from.

A theme emerging is the push to deliver ROI and the rising importance of “context”. ROI is easy to define; context, less so. It’s one of those things that will feel obvious in hindsight, once someone nails the experience. When they do, ROI won’t be far behind.

#1– Monte Carlo published a The Total Economic Impact Of Monte Carlo’s Data + AI Observability Platform. May 2025.

My Take: This commissioned Forrester report focused on the ROI of deployed Monte Carlo’s platform. Not surprisingly, the report found in favor of engaging Monte Carlo, but the framework is of general interest, combining quantified benefits, like direct time savings, downtime reductions, direct reduction in cloud costs, with non-measurable benefits (unrelated, but related article from Stephen Haptonstahl Not everything valuable is measurable. June 2025).

#2 – Lior Barak published The Success Metrics Layer. June 2025.

My Take: This article addresses how to measure “success”. There is an element of K.I.S.S. (Keep It Simple Stupid). Preferably binary metrics (i.e. Y/N).
Spend time defining “success” up front, keeping in mind that what looks like success to one group, might look very different for another group (ex. Finance & Marketing might define success differently).

#3 – Swaroop Jagadish and Shirshanka Das from Datahub published Unlocking the Future of AI and Data: DataHub’s $35M Series B Journey. May 2025.

My Take: The article shares the idea of technical context, operational context, business & social context. When I think of context, I think of the machine/model/application … whatever … to understand what I want. And then deliver to me only that information. I just want it to be easy and obviously valuable.
A simple example would be delivering to me a summary of my day that is unique to me. From a professional knowledge worker’s perspective, this would include only the most relevant information from the vast amount of data resources to which I have access.
This includes the things I know I need. But this should also include the information I didn’t even know existed that is helpful to me (for example, ahead of a meeting, deliver me a summary & link to meeting notes from a colleague that met with the same person a year ago … notes likely buried deep within the CRM that I never would have even looked for).

BONUS: Robeco published The data advantage: How web scraping and NLP give investors a decision-making edge. May 2025. “In the future, alternative data, machine learning, and NLP will enhance collaboration by improving both quant models and fundamental research, thereby strengthening the firm’s offering. Asset managers that can adapt and leverage the growing power of data and AI techniques will see differentiated advantages.”

What else I’m reading:

Jason Saltzman published Data is the fuel for AI. May 2025.

Jj published The Copilot Delusion. May 2025.

Gene Rapoport, Sanjin Bicanic, and Muyiwa Talabi of Bain published Survey: Generative AI’s Uptake Is Unprecedented Despite Roadblocks. May 2025.

Striim published Modern Data Streaming for Real-Time Artificial Intelligence. May 2025.

Joe Reis published AI Denialists are the Data Industry’s Flat Earthers. May 2025.

Matt Ober published FactSet Chooses a New CEO. June 2025.

SafeGraph published ICSC 2025 Recap: The Data Arms Race Is On. June 2025..

Alex Izydorczyk published GenAI for Hedge Funds Startups. June 2025.

Atheron Lab published AI Startup Fallout: Tracking Big Tech’s Disruptive Impact. June 2025.

Nick Newport published AI innovations in Investment Management. June 2025.

Multiple Authors published Market Signals from Social Media. May 2025 (h/t Michiel Maaskant)

Mehdio published Making Cursor smarter (and up to date). May 2025.

Ollie Forstythe published Adopting AI Within The Workplace. May 2025.

Source: A16z’s podcast published Where Will Value Accrue in AI: Martin Casado & Sarah Wang. May 2025.

My Take: Interesting interview with two people investing in AI companies. Some takeaways:
- 1- The market is larger and growing more quickly than we expected … and we were very bullish!
- 2- Be careful where you place bets, as the wipe-out potential is as big as the hypergrowth potential.
- 3- Heat cannot be confused with momentum (I’m not sure what that means exactly?).
- 4- Productivity gains are much greater this year vs last year.
The biggest concerns seem to be on retention. AI seems to have had a big positive impact on bootstrapping growth … but what will ultimately be the retention?
13:15 – “tangible ROI focus” & that is driving the growth. We see this everywhere … focus on specific ROI.

Highlights (23-minute run time)

Minute 02:00 – interview starts; what is the state of play in AI today?
Minute 03:30 – what is the scale we are talking about?
Minute 04:20 – looking “across the stack”
Minute 05:30 – investment moving from infrastructure to apps
Minute 07:45 – AI vs SaaS
Minute 09:45 – Defensibility
Minute 11:45 – Cursor discussion (how are they doing this?)
Minute 15:40 – “pro-sumer” market has us focusing more on retention
Minute 16:50 – big wipeouts discussion (the more money raised early…the more pressure to show results)
Minute 18:15 – discussion if China impact (largely positive)
Minute 19:00 – discussion of models (gotta be careful here … premium teams that cn get the capital)
Minute 20:00 – spicy takes
Minute 20:50 – key messages

Source: 2025 Dupont AI Survey. May 2025.

Sidenote: had a conversation with my college-aged daughter (Eliza Farrall) who is in her first “real world” internship. I said she should use ChatGPT to help work through a problem, and she thought that was cheating. I said, “It is cheating at school. It is expected at work.”

BONUS: KPMG published 2025 Banking Survey: Technology. May 2025.