Benn it seems there is a general trend towards "raise less, build more" as a pendulum swing away from excessive finance. eg https://trohan.com/2023/08/20/raise-less-build-more/ not canonical source of the idea but just the top of the pile on my reading list
I wonder what the impacts of that look like for startups? Some guesses but would love your take here.
Better products - products like dbt had pretty good potential but pretty much didn’t materially change at all once they took VC funding, I wonder if this is common? Would they have done better to stay lean?
Less money in the ecosystem - a huge part of software success is predicated on selling to other well funded startups, creating a flourishing ecosystem that can then “escape” and sell beyond that. If companies take less funding, there will be less money to spend on other startups, meaning the ecosystem as a whole is less funded, less dynamic. Analogy would be that only the strongest “weediest” survive in this “garden”, rather than creating the environment where roses can thrive... (low effort visual). Weedy in the sense of uncreative or unoptimised. Would we see the organic and creative destruction that leads to some great innovations? (struggling to imagine one, ChatGPT?)
VC industry - I don't know. Stops working? They need companies to return the fund. This would kill them? Maybe not, maybe it leads to better outcomes. This is where my insight falls short.
I do think there's something to be said for raising less, and if I were to do things again, I'd be very deliberate about how I raised. But that's less about money making you lathargic, and more about keeping control of the the throttle. Once you take a big enough check, you have to go for broke. And I think a lot of startups commit to that path way sooner than they should.
(I will say, overall thrust aside, I think that post gets a lot wrong. I think the argument that VCs are aggressively pushing huge rounds on companies is mostly a caricature that only happens to a handful of very hot deals. Plus, it just dimisses offhand that companies that raise more money aren't more successful, which I doubt is actually true, and almost certainly isn't literally true, since successful companies will almost by definition raise more than those that fail.)
Still, assuming that trend is true...
- on products, I don't think we get better ones. Like, had dbt not raised money, what would they have done instead? They'd still have to make money, still be competing against an open source thing, etc. They might've retained more good will, but it's hard to imagine that the product would be better had it been built by a handful of engineers scraping by and consulting part time than by a full team dedicated to it full time.
On the ecosystem, I could see either way? Crypto seems like a good signal here. Lots of money in the system didn't actually create durable products; it just attacted a lot of people looking to get rich quick. Data probably has some of the same dynamics, where people got into it because they could get funding and the clout that came with it, not because of any real dedication to the problem. Maybe some of those blind squirrels find nuts, but for the most part, it just seems like a waste of money.
On VCs, this one seems the most straightforward - a lot of funds die. Which seems...right? Just as a lot of people started companies who weren't that interested in the problem, a lot of people also started funds who had no real business doing it. There's no way that money was all well spent.
Ok interesting, thanks! Very interesting points on the article. What is the VC angle for pushing this narrative? Thin out the crowd?
Product - I don't know! Dataform was really good, so much better than dbt Cloud at the time, they got pushed out by dbt, there was no air in the room. I Make the point here a bit. https://substack.com/@groupby1/note/c-39535008
Ecosystem - crypto does seem a great example. Maybe on the extreme end, so much noise clouds out the value creation but I imagine a good few years for that outcome to be proven. Maybe the "constraints lead to creativity" hold true here and my weed analogy doesn't work. Maybe algal bloom is the better one!
VC - yes this seems sensible, no Matt Levine quirky second order effects here!
Eh, I don't know if there's really some VC angle for pushing that narrative; honestly, it mostly just strikes me as lazy. It's a lot easier to write posts about how bad this extreme thing is than to talk about nuance. So my guess is the "angle" is really, "I want to become a VC influencer and I need to write content."
On product, I never used Dataform, so I'll take your word for it on the experience. That said, I don't think the money had much do with it. Looking back, the acquisition got announced in early Dec of 2020, less than a month after dbt announced their series B. Acquisitions take some time from first conversation to announcement, so the deal was probably in the works either when dbt had only raised $13m in their A. Even if they had wind of the B when they started talking with Google, the B was only $30m, which isn't small, but hardly enough to corner an entire market. So I'd argue that datafold either sold because dbt already had the community traction *without* the money, or datafold's product simply didn't have that much appeal. (There's a third argument here that they sold because dbt cornered the *VCs market*, either because VCs didn't want to bet against it, or something more collude-y, where VCs already had a horse in the race. The former seems possible; the latter, eh, probably less.)
On the ecosystem, in thinking a bit more about it, that seems like it depends mostly on the value of the underlying thing. Crypto was mostly a waste because the foundation didn't have much value. For data stuff, is that true? Perhaps more than we'd like to admit, but less so than crypto, I'd hope.
I'll ask around about Dataform, but I think it was probably a combination of all of the things you listed. I was close to it and there were many pretty sad customers who had to figure out a plan, and dbt Cloud being a pretty compromised alternative. Felt like PMF, or more so than dbt Cloud has ever really attained (IMO.. I tried both extensively)
Another way of thinking: If dbt wasn't buoyed along by their community, they'd have discovered what Dataform did way back then, that there isn't an effective business model to support this.
> Crypto was mostly a waste because the foundation didn't have much value. For data stuff, is that true? Perhaps more than we'd like to admit, but less so than crypto, I'd hope.
Big if true!
ps I would love your take on the cascade epilogue. I've largely taken their side of the argument ("The fracturing of BI" etc) but not found resonance in the slacks of the world.
On Dataform, cool. But also, like, it seems like there's something off about the idea that Dataform had PMF that dbt never did, when dbt is the one that dominated the market. Dataform may have been a more polished product, and had a crisper experience, or whatever, but something about dbt fit the market far, far better. I don't think it's just the community either, since most of the community came from people who really liked the product.
I have some thoughts about this Cascade narrative, and don't think I entirely agree (something like this was the rough plan for the post for the this week, but we'll see what actually happens). The short version is I think this sentence doesn't come close to holding up:
For the last two decades, the “business intelligence” space has been slowly trending from monolithic, vertically-integrated tools (Cognos, Microstrategy) to more lightweight, audience-specific apps that can work on a variety of data stores (Hex, Sigma, Metabase).
Hex, Sigma, and Metabase all make, what, about $50m total? That may turn into some mega- trend, but to assume that it will is just getting fooled by hype. In the type since those companies have been around, Looker's added ~$500m in revenue. Plus, PowerBI is actually younger than, say Metabase. There's not some obvious fragmentation trend here; there are people trying to fragment it, and having mild success that may well stall.
I don't think Dataform had general PMF to be clear, just a painful migration when they were acquired. They lacked broad community adoption of their open source core, and I think this was a tough narrative to compete directly with dbt here. Dataform retaining dbt compatibility would have been useful, but already then it lead to a poor UX and so they split.
SQL transformation with a small team, Dataform was just easier and quicker. I wonder if the strong stigma attached to Dataform within dbt community maybe was part of the story, something along the lines of this bit. It felt tribal.
Cascade, super interesting, definitely some hype within the fragmentation efforts indeed. Look forward to learning more in due course.
One obvious thought on fragmentation, from a user perspective it seems generally positive (I don't need to buy X when Y is a better fit), but implied is smaller subsets of the original market, hoping that a new market is found.
Mr. Worldwide says fish or cut bait, Stancil
Just me and George RR Martin, playing chicken with death with our masterpieces.
Benn it seems there is a general trend towards "raise less, build more" as a pendulum swing away from excessive finance. eg https://trohan.com/2023/08/20/raise-less-build-more/ not canonical source of the idea but just the top of the pile on my reading list
I wonder what the impacts of that look like for startups? Some guesses but would love your take here.
Better products - products like dbt had pretty good potential but pretty much didn’t materially change at all once they took VC funding, I wonder if this is common? Would they have done better to stay lean?
Less money in the ecosystem - a huge part of software success is predicated on selling to other well funded startups, creating a flourishing ecosystem that can then “escape” and sell beyond that. If companies take less funding, there will be less money to spend on other startups, meaning the ecosystem as a whole is less funded, less dynamic. Analogy would be that only the strongest “weediest” survive in this “garden”, rather than creating the environment where roses can thrive... (low effort visual). Weedy in the sense of uncreative or unoptimised. Would we see the organic and creative destruction that leads to some great innovations? (struggling to imagine one, ChatGPT?)
VC industry - I don't know. Stops working? They need companies to return the fund. This would kill them? Maybe not, maybe it leads to better outcomes. This is where my insight falls short.
I do think there's something to be said for raising less, and if I were to do things again, I'd be very deliberate about how I raised. But that's less about money making you lathargic, and more about keeping control of the the throttle. Once you take a big enough check, you have to go for broke. And I think a lot of startups commit to that path way sooner than they should.
(I will say, overall thrust aside, I think that post gets a lot wrong. I think the argument that VCs are aggressively pushing huge rounds on companies is mostly a caricature that only happens to a handful of very hot deals. Plus, it just dimisses offhand that companies that raise more money aren't more successful, which I doubt is actually true, and almost certainly isn't literally true, since successful companies will almost by definition raise more than those that fail.)
Still, assuming that trend is true...
- on products, I don't think we get better ones. Like, had dbt not raised money, what would they have done instead? They'd still have to make money, still be competing against an open source thing, etc. They might've retained more good will, but it's hard to imagine that the product would be better had it been built by a handful of engineers scraping by and consulting part time than by a full team dedicated to it full time.
On the ecosystem, I could see either way? Crypto seems like a good signal here. Lots of money in the system didn't actually create durable products; it just attacted a lot of people looking to get rich quick. Data probably has some of the same dynamics, where people got into it because they could get funding and the clout that came with it, not because of any real dedication to the problem. Maybe some of those blind squirrels find nuts, but for the most part, it just seems like a waste of money.
On VCs, this one seems the most straightforward - a lot of funds die. Which seems...right? Just as a lot of people started companies who weren't that interested in the problem, a lot of people also started funds who had no real business doing it. There's no way that money was all well spent.
Ok interesting, thanks! Very interesting points on the article. What is the VC angle for pushing this narrative? Thin out the crowd?
Product - I don't know! Dataform was really good, so much better than dbt Cloud at the time, they got pushed out by dbt, there was no air in the room. I Make the point here a bit. https://substack.com/@groupby1/note/c-39535008
Ecosystem - crypto does seem a great example. Maybe on the extreme end, so much noise clouds out the value creation but I imagine a good few years for that outcome to be proven. Maybe the "constraints lead to creativity" hold true here and my weed analogy doesn't work. Maybe algal bloom is the better one!
VC - yes this seems sensible, no Matt Levine quirky second order effects here!
Eh, I don't know if there's really some VC angle for pushing that narrative; honestly, it mostly just strikes me as lazy. It's a lot easier to write posts about how bad this extreme thing is than to talk about nuance. So my guess is the "angle" is really, "I want to become a VC influencer and I need to write content."
On product, I never used Dataform, so I'll take your word for it on the experience. That said, I don't think the money had much do with it. Looking back, the acquisition got announced in early Dec of 2020, less than a month after dbt announced their series B. Acquisitions take some time from first conversation to announcement, so the deal was probably in the works either when dbt had only raised $13m in their A. Even if they had wind of the B when they started talking with Google, the B was only $30m, which isn't small, but hardly enough to corner an entire market. So I'd argue that datafold either sold because dbt already had the community traction *without* the money, or datafold's product simply didn't have that much appeal. (There's a third argument here that they sold because dbt cornered the *VCs market*, either because VCs didn't want to bet against it, or something more collude-y, where VCs already had a horse in the race. The former seems possible; the latter, eh, probably less.)
On the ecosystem, in thinking a bit more about it, that seems like it depends mostly on the value of the underlying thing. Crypto was mostly a waste because the foundation didn't have much value. For data stuff, is that true? Perhaps more than we'd like to admit, but less so than crypto, I'd hope.
I'll ask around about Dataform, but I think it was probably a combination of all of the things you listed. I was close to it and there were many pretty sad customers who had to figure out a plan, and dbt Cloud being a pretty compromised alternative. Felt like PMF, or more so than dbt Cloud has ever really attained (IMO.. I tried both extensively)
Another way of thinking: If dbt wasn't buoyed along by their community, they'd have discovered what Dataform did way back then, that there isn't an effective business model to support this.
> Crypto was mostly a waste because the foundation didn't have much value. For data stuff, is that true? Perhaps more than we'd like to admit, but less so than crypto, I'd hope.
Big if true!
ps I would love your take on the cascade epilogue. I've largely taken their side of the argument ("The fracturing of BI" etc) but not found resonance in the slacks of the world.
https://www.cascade.io/epilogue/the-data-and-analytics-market
On Dataform, cool. But also, like, it seems like there's something off about the idea that Dataform had PMF that dbt never did, when dbt is the one that dominated the market. Dataform may have been a more polished product, and had a crisper experience, or whatever, but something about dbt fit the market far, far better. I don't think it's just the community either, since most of the community came from people who really liked the product.
I have some thoughts about this Cascade narrative, and don't think I entirely agree (something like this was the rough plan for the post for the this week, but we'll see what actually happens). The short version is I think this sentence doesn't come close to holding up:
For the last two decades, the “business intelligence” space has been slowly trending from monolithic, vertically-integrated tools (Cognos, Microstrategy) to more lightweight, audience-specific apps that can work on a variety of data stores (Hex, Sigma, Metabase).
Hex, Sigma, and Metabase all make, what, about $50m total? That may turn into some mega- trend, but to assume that it will is just getting fooled by hype. In the type since those companies have been around, Looker's added ~$500m in revenue. Plus, PowerBI is actually younger than, say Metabase. There's not some obvious fragmentation trend here; there are people trying to fragment it, and having mild success that may well stall.
I don't think Dataform had general PMF to be clear, just a painful migration when they were acquired. They lacked broad community adoption of their open source core, and I think this was a tough narrative to compete directly with dbt here. Dataform retaining dbt compatibility would have been useful, but already then it lead to a poor UX and so they split.
SQL transformation with a small team, Dataform was just easier and quicker. I wonder if the strong stigma attached to Dataform within dbt community maybe was part of the story, something along the lines of this bit. It felt tribal.
https://groupby1.mattarderne.com/p/dataform-and-dbt#:~:text=preferences%20or%20opportunities.-,Sentiment,-generally%20went%20something
Cascade, super interesting, definitely some hype within the fragmentation efforts indeed. Look forward to learning more in due course.
One obvious thought on fragmentation, from a user perspective it seems generally positive (I don't need to buy X when Y is a better fit), but implied is smaller subsets of the original market, hoping that a new market is found.