there's an option that the terminal and insight server share a future, IMO.
users can have the insight machine use the terminal for structured data retrieval. this would be inefficient but legal.
one way the incumbent players can adapt - is to find some up-and-promising insight solution - and buy it. many reasons that this could play out.
1/ classic acqui-hire. instead of chasing AI talent and trying to create something they'd like to do - you get lots of them at once.
2/ existing clients of the insight server - might not be using your terminal because data budget did not allow it, but insights budget just might.
3/ existing customers of both - might have even less incentives to look for an alternative, even if price increases due to even higher costs producing faster alpha.
indeed, all these outcomes are possible, and in fact likely to play out as the market games itself to the new opportunities. I do think a cross platform outcome (incumbent terminal integrated with insight) is inevitable, but I'm interested more in the shift of power that could accompany it. It would take threading a needle but I think those that crack the insight engine harnessing user/firm context have a rare opportunity to upset the current power balance for the better. Hopefully this community can play a part in making that happen!
This reads like a philosophy of the market eating itself, The perennial tale of technology biting its own behind a bit like generative AI. Where is Alpha? When the internet has become a public relations and propaganda channel, what this reminds me of is startups like OpenAI need to harness communications like a cult. If it comes all about perceived value, and not fundamentals.
Very helpful framing and illustrations. This is a rare note worth printing out are reviewing over time. Am kind of in the middle of just this now. Pls DM if interested in discussing. Thanks
Indeed old terminals optimised for fast refresh and response first above all else. So a good amount of logic was needed on the client side. FDS and LSEG have web-only versions but feature parity cannot be reached. An interesting engineering problem yet to be solved.
To a certain extent, isn't this what Renaissance already does? And maybe even skipping the insight part of the funnel, quoting Robert Mercer...
“If somebody came up with a theory about how the phases of Venus influence markets, we would want a lot of evidence.” But he adds that “some signals that make no intuitive sense do indeed work.” Indeed, it is the nonintuitive signals that often prove the most lucrative for Renaissance. “The signals that we have been trading without interruption for fifteen years make no sense,” Mercer explains. “Otherwise someone else would have found them.”
And no argument about their ability to find Alpha over the long term (and charge an insane fee to provide it to their clients).
Indeed the man who solved markets was far ahead of the game in automating alpha. But sadly he didn't and couldn't make it accessible to the public. Systematic trading is relatively mature and was designed to capture alpha, systematic (rather assisted) investing has some way to go as it can be setup to facilitate the process than identify the signal, opening a much larger TAM without taking on market risk. What do you think?
So one lens is through the product as a supply/value chain. Why hasn't Bloomberg gotten into the algo or trading strategy business rather than stopping at the tools to implement it? Famously (and maybe apocryphally) Bloomberg product managers would ask their users about what they did immediately before (trigger) and immediately after (action) using the terminal and crept into expanded user problems/value prop. Why stop where they did? Bigger TAM yes, but would it somehow put their existing revenue at risk? Was it something their customers were unwilling to buy?
One potential answer is that finding alpha is an important differentiator for investors, and like Renaissance or claims in recent non-compete lawsuit on FOREX traders at Jane Street, it's a lucrative trade secret and not something you could or would want to outsource/buy from a source widely available to competitors. In other words, this is extremely valuable, a much larger TAM, but only when it is unique to the customer. This would favor homegrown/built solutions, or at least not addressable by a utility player like Bloomberg.
Another may be that there is an incremental value as a feature/tool set in the analysis and investment memo prep. If that's the case I'd expect to see that quickly developed/acquired by Bloomberg as part of their adjacent value strategy.
Indeed utility providers have to stay in their lane and it's an interesting question as to how to capture new TAM. No good option exists. Increase the size of data repositories as Drew suggests, and leave it to the quants to find the signal. This runs into ROI questions. The other is find the signals itself which is subject to alpha decay. I'd like to explore the topic in depth in a post soon.
The fact that the terminals are a fat client might, be their major longterm problem. Web first apps can do charts, tables, and code responses, not just text. I wish substack would let me post pictures.
Super solid take, thank you! Cheers!
Thanks Mav. Your substack def helping us finding insights in all the noise
Welcome TT, glad to hear, cheers!
insightful.
there's an option that the terminal and insight server share a future, IMO.
users can have the insight machine use the terminal for structured data retrieval. this would be inefficient but legal.
one way the incumbent players can adapt - is to find some up-and-promising insight solution - and buy it. many reasons that this could play out.
1/ classic acqui-hire. instead of chasing AI talent and trying to create something they'd like to do - you get lots of them at once.
2/ existing clients of the insight server - might not be using your terminal because data budget did not allow it, but insights budget just might.
3/ existing customers of both - might have even less incentives to look for an alternative, even if price increases due to even higher costs producing faster alpha.
etc.
indeed, all these outcomes are possible, and in fact likely to play out as the market games itself to the new opportunities. I do think a cross platform outcome (incumbent terminal integrated with insight) is inevitable, but I'm interested more in the shift of power that could accompany it. It would take threading a needle but I think those that crack the insight engine harnessing user/firm context have a rare opportunity to upset the current power balance for the better. Hopefully this community can play a part in making that happen!
"though easter eggs exist" – You can't leave us hanging without a foot note!
ah good catch Milan, I've updated the post with footnote 2.
Another excellent article.
Thanks Byron
The last two articles have showcased some great thinking. Keep it up!
Very rich and insightful. thank you!
Thanks Chetan
Very insightful, TT, thank you. For your stats…Made it to the end!
Thanks Sean, appreciate it :)
This reads like a philosophy of the market eating itself, The perennial tale of technology biting its own behind a bit like generative AI. Where is Alpha? When the internet has become a public relations and propaganda channel, what this reminds me of is startups like OpenAI need to harness communications like a cult. If it comes all about perceived value, and not fundamentals.
Hi Michael, indeed, it depends on which side of the fence is greener. Capturing or facilitating alpha.
Very helpful framing and illustrations. This is a rare note worth printing out are reviewing over time. Am kind of in the middle of just this now. Pls DM if interested in discussing. Thanks
Thanks Robert for the kind words. Look forward to your thoughts...
Amazing article. Thank you
Thanks Raj
Great take. I invite you to see what we are building with Enterprai.
Certainly, where can I find out more?
It’s a messy, complex, layer cake of jobs analysts do, that never ends. Nice work on the piece.
Messiness with no end in sight. ToDatta doing the hard work of making it easier...🫡
That is the plan, and I'm told we're on the right track.
Indeed old terminals optimised for fast refresh and response first above all else. So a good amount of logic was needed on the client side. FDS and LSEG have web-only versions but feature parity cannot be reached. An interesting engineering problem yet to be solved.
To a certain extent, isn't this what Renaissance already does? And maybe even skipping the insight part of the funnel, quoting Robert Mercer...
“If somebody came up with a theory about how the phases of Venus influence markets, we would want a lot of evidence.” But he adds that “some signals that make no intuitive sense do indeed work.” Indeed, it is the nonintuitive signals that often prove the most lucrative for Renaissance. “The signals that we have been trading without interruption for fifteen years make no sense,” Mercer explains. “Otherwise someone else would have found them.”
And no argument about their ability to find Alpha over the long term (and charge an insane fee to provide it to their clients).
Indeed the man who solved markets was far ahead of the game in automating alpha. But sadly he didn't and couldn't make it accessible to the public. Systematic trading is relatively mature and was designed to capture alpha, systematic (rather assisted) investing has some way to go as it can be setup to facilitate the process than identify the signal, opening a much larger TAM without taking on market risk. What do you think?
(reposting as a reply to this thread):
So one lens is through the product as a supply/value chain. Why hasn't Bloomberg gotten into the algo or trading strategy business rather than stopping at the tools to implement it? Famously (and maybe apocryphally) Bloomberg product managers would ask their users about what they did immediately before (trigger) and immediately after (action) using the terminal and crept into expanded user problems/value prop. Why stop where they did? Bigger TAM yes, but would it somehow put their existing revenue at risk? Was it something their customers were unwilling to buy?
One potential answer is that finding alpha is an important differentiator for investors, and like Renaissance or claims in recent non-compete lawsuit on FOREX traders at Jane Street, it's a lucrative trade secret and not something you could or would want to outsource/buy from a source widely available to competitors. In other words, this is extremely valuable, a much larger TAM, but only when it is unique to the customer. This would favor homegrown/built solutions, or at least not addressable by a utility player like Bloomberg.
Another may be that there is an incremental value as a feature/tool set in the analysis and investment memo prep. If that's the case I'd expect to see that quickly developed/acquired by Bloomberg as part of their adjacent value strategy.
Thoughts?
Indeed utility providers have to stay in their lane and it's an interesting question as to how to capture new TAM. No good option exists. Increase the size of data repositories as Drew suggests, and leave it to the quants to find the signal. This runs into ROI questions. The other is find the signals itself which is subject to alpha decay. I'd like to explore the topic in depth in a post soon.
It was all about p-values and a deep data repository
The fact that the terminals are a fat client might, be their major longterm problem. Web first apps can do charts, tables, and code responses, not just text. I wish substack would let me post pictures.