Thanks for being here!
This is the Alternative Data Weekly for Friday, June 23, 2023.
Announcements:
I’ll be at Snowflake Summit in Vegas June 26-27-28. Let me know if you’ll be there. Monday night 26th I will be at the Low-Key Data Happy Hour sponsored by Portable, Metaplane, RudderStack, Brooklyn Data Co., Striim, DataGalaxy
If you’ll be in SF at DataBricks instead, make sure you check out the Bobsled event June 28th.
Theme that emerged in this week’s email is … data for AI training is analogous to shovels for gold miners.
QUOTES
“Without a good data product, you won’t be a successful company in the long run.”
News Articles
Podcasts
Cool Charts
New Products (coming soon)
Final Thoughts (NAAB - Newsletter As A Business)
#1 – Axios published Thousands of Reddit communities go dark in protest of API changes. June 2023.
My Take: This is a fluid situation and I am writing this early in the week, so things may have changed. Data pricing is coming to the forefront. Anyone in the data biz knows pricing is tricky. In the course of Reddit’s day-to-day, they generate lots of valuable data. Historically, businesses have been able to freely access this data via API to build products. Going forward, developers wishing to build products for Reddit and/or do pure research will continue to access the data for free. BUT those businesses using Reddit data for commercial purposes will have to pay. A LOT. Twitter did something similar earlier this year.
The most frequently cited Reddit example is Apollo. Apollo created an app that is a better/different way to interface with Reddit. Apollo charges $10/year for the app. Until now, Apollo was able to access Reddit’s data for free. Reddit is now charging a data fee that will cost Apollo more than they can likely pass along to their customers. What happens next? Either Apollo passes along the increased cost to customers (see elastic demand) … or they cease operations.
I think Reddit is trying to capture the highest paying clients up front … someone must be willing to pay, or they wouldn’t be charging it. Once the high-end is established, Reddit can then modify the data offering in such a way (delays, less robust access, etc) that they find product-market-fit at the lower end of the pricing scale.
This case will be studied in business schools for years to come. Worth reading Byrne Hobart’s thoughts on the Reddit situation.
#2 – Rohit Choudhary published How to Build Great Data Products. June 2023.
My Take: From the author: “The goal is to uncover whether the data available to you can deliver insights that are not available anywhere else in the market”. Most interesting to me is that more data is not always better data. Having strong domain knowledge & an understanding of the customer’s pain points is key. This allows you to keep it simple. Customers have short attentions spans; they have to understand what you are communicating immediately. The right data (& only the right data) delivered at the right time to the right people solves the problem.
#3 – Oxylab’s Aleksandras Šulženko published How Governments Can Use Alternative Data for Policymaking. June 2023.
My Take: This is a back to the basics Alt Data article, but with a view towards how governments can better use alternative data sources. Of most interest to me was the idea that governments can Nowcast economic conditions (modern day Sim City). Given how governments operate, they may be the last to put good data to work, but can arguably create the most benefit once they are doing it.
BONUS: Vin Vashishta’s Generative AI Needs Context And Business Acumen To Succeed. June 2023. “…but if the goal is wider adoption, ignoring the way businesses operate will become a barrier.”
BONUS 2: Pierluigi Vinciguerra published The state of web scraping and AI. June 2023. “We’re at the beginning of a new gold rush (AI) where those who sell shovels (web data) can enable gold discovery.”
BONUS 3: The Verge’s interview with Reddit CEO Steve Huffman. June 2023. “Giving a 100-percent free subsidy to competitors is not a good business strategy.”
What else I am reading:
Jonas Laeben’s Unlocking the Power of AI: Why You Should Educate Yourself and Cultivate a Critical Mind. June 2023.
Scott Straub’s How to Do Good With Your Data and Harness Analytics to Serve Underserved Populations. June 2023.
Vin Vashishta’s The Reddit Blackout: A Cautionary Tale About Making Money With Data. June 2023.
Jason DeRise’s The Data Score published NVIDIA: How could alternative data be used to assess its long-term potential?. June 2023.
Data observability platform Bigeye announces acquisition of Data Advantage Group. June 2023.
#1 – Lenny’s Podcast Lessons on building product sense, navigating AI, optimizing the first mile, and making it through the messy middle | Scott Belsky (Adobe, Behance). May 2023.
My Take: A bit of a different podcast here … this podcast is focused on building products, and not specifically data products. But with all the “data as a product” themes we’ve seen in recent weeks, I wanted to dig-in on a building a (data) product.
Interviewee Scott Belsky is described as having great “product sense” given all the successful products he has built. Most interesting to me was the idea that he described the importance of having empathy (13:30) for the customer … know the pain your customer is experiencing. Make the product feel like magic & offer occasional positive surprises.
It is worth listening to AI discussion starting around minute 26:00. Scott describes how AI will be able to do repetitive, mind-numbing tasks and free up talent to get into the flow-state & focus on more impactful tasks.
Highlights (62-minute run time):
Minute 00:00 – Scott’s background
Minute 04:50 – Why Scott shifted roles at Adobe
Minute 08:29 – Advice for PMs looking to build product sense
Minute 10:43 – The first mile
Minute 13:18 – How to develop more empathy
Minute 16:33 – How to build consumer products that work (nimbleness & knowing customers behavior)
Minute 20:42 – Scott’s philosophy that you should “only do half the things you want to do”
Minute 26:15 – Scott’s optimism about how the world will look in five years with AI
Minute 29:44 – How AI will impact product teams (“collapse the stack”)
Minute 32:55 – How the PM role will change as a result of AI
Minute 35:09 – How Adobe is leveraging AI tools
Minute 36:59 – What the term “golden gut” means
Minute 38:15 – Advice for PMs to stay ahead of the new AI trends (“need to play with this new tech…”)
Minute 41:02 – How to start writing more (I like this!) … Scott’s Implications newsletter
Minute 41:49 – The messy middle (Scott’s book)
Minute 47:03 – What Scott looks for as an angel investors
Minute 50:16 – Why resourcefulness will take you further than resources
Minute 52:41 – Adobe’s current priorities and the path ahead
Minute 54:58 – Lightning round
Source: The state of web scraping and AI
Web scraping momentum (from Google Trends):
Coming soon.
Contact me if you’d like to highlight something in this space.
Source: My brain.
NAAB: Newsletters As A Business
I started writing this weekly back in October 2020. I’ve since moved from MailChimp to SubStack. I’ve changed formats a few times, but feel like the weekly has found a groove and we have been rewarded with new subscribers & consistently strong readership (~50% open rate; 15-20% click on at least one link).
This weekly newsletter surpassed 1,000 subscribers on June 15th. It only took me 138 consecutive weeks of publishing (sarcasm).
When I let the world know via LinkedIn and Twitter, I subsequently added more subs than any previous 10 day period.
I also had a few firms reach out about sponsorships.
In the next few weeks, The Alt Data Weekly will add another section with “new products” (see above) that will highlight new offerings in this space. We will be offering the opportunity to sponsor this weekly as well.
I am not exactly sure what this will look like, but if your company has interest in reaching a very targeted group of data vendors, data buyers, and data practitioners, let me know and we can work something out.
Thanks!
…and counting