Supply And Demand
I saw this chart on Semil‘s blog this morning:
What is shows is that as the amount of money raised (and deployed) in seed funds has grown over the last ten years, the ability of the companies that received those seed investments to raise a follow-on Series A round has declined (massively).
That trend is what you would expect, of course. Supply outstrips demand at some point.
But from where I sit, I am having trouble with the magnitude of these numbers.
First of all, I don’t think the “conversion” from Seed to Series A was ever in the 80% range. I think it is generally around 50% and moves around that number a fair bit. But I can’t imagine a time when 80% of seed funded companies go on to raise a Series A.
I also don’t think it is now sub 30%. Maybe sub 40%. Maybe not. But I’m having a hard time believing that less than 3 in 10 seed-funded companies go on to raise a Series A.
What I think has happened is that there is now a significant “grey area” that has developed in the middle of Seed and Series A. We have “post seed”, or “seed two” rounds. We have “early As”.
So the data isn’t clean and it is harder to track from type of round to type of round.
I also think a lot of the seeds that were being done back in 2006 were non-institutional and harder to track. As the seed fund market has exploded in the last ten years, more of the seed rounds are including at least one institution and are now getting tracked in a way they were not in 2010.
So, are more companies getting seed funded? Yes.
Is a lower percentage of them going on to get Series A rounds? Yes.
Has that percentage gone from north of 80% to south of 30% in ten years? No way.
But, to the question of “is it harder to raise a Series A?”, I think the answer is “it depends.”
There is more Series A money out there too, but it has not grown as quickly as seed money.
It is certainly harder to raise a Series A than a Seed. But that has been true for some time.
As a beginning seed stage angel investor, I found Crunchbase to be super helpful in understanding valuations and deal flow for early stage companies. However, it does seem a lot of the data there is skewed towards institutional involvement as individuals don’t report their investments.
Peter Chung:Your reply appears like an advertising in disguise. Are you pushing Crunch base on us?Captain Obvious!#UNEQUIVOCALLYUNAPOLOGETICALLYINDEPENDENT
FRED:Could is actually be the seed round is raising funding that may not require an series A until an acquisition or private sale or ICO wiping out the need for any VC funding that could alter overall numbers that traditionally were relied upon to access those numbers of funding rounds.Captain Obvious!#UNEQUIVOCALLYUNAPOLOGETICALLYINDEPENDENT
“But from where I sit, I am having trouble with the magnitude of these numbers.”Investment Cowboys such as Fred don’t like to see closing of the frontier. Put another way, maturation of the “dot.com”investment market is leading to lower profits for the little guy riding his horse because the “big money” is consolidating the sector. See, for example, how a16z.com is pursuing a “textbook case” of vertical and horizontal integration.
I agree 100% with that
Survival bias at work here: the older the data gets, the more it tends to be skewed towards companies that did reach next financing stage and have more info available on them. And on top of it, the longer a cohort, the more likely that it reached the next stage. At Dealroom.co (together with two European VCs) we are looking into a very similar analysis for Europe as we speak. Released soon.
the fog of war.early. europe?
I’d like to see 2 more variables on that time horizon: average size of these deals and elapsed time between Seed and A. That would give more depth to the analysis.
i don’t like the straightness of that black line. it upsets the visual perception of that’s going on (it’s not an ‘honest’ representation).
Might be useful to break out the seed funding amount by (1) – Number of companies seed funded and (b) – Average size of each seed deal. The y1 axis is showing the product of the two variables, while the y2 axis is applying the ratio to the first variable.Besides the expected demand-supply outcome highlighted in the post, is this also driven by whether deeper pockets are required now than previously to build out a company ? As in…you are more likely as an investor to have to break out “larger per deal amounts” at the Series A stage, for a lesser proportion of companies. The ‘winner-take-most’ effect moving up the funding chain.
Somewhat alarming chart. I like your analysis, Fred. Does anyone have any similar data points on how often a Series B follows A? I’m curious to better understand how often round B and future rounds follow.
I’m not going to debate the chart. But history does rhyme. Every time we have a long expansion, seed money gets put in a too expensive valuations and too great of a rate.Then the tears start.Because when you raise at seed you are funding a dream. When follow on investors come in they are funding numbers. Nothing like numbers to fuck up a dream.Being a seed investor is BRUTAL. If things go well follow on’s will want to cram you down. That is why you need pro-rata rights.But when it is not certain, pain train. New investors will cram you down and you are faced with taking a total loss or a cram down.Of course you also have total flame outs.Every cycle people learn this and get washed out.
this is true. valuations have gone up in the last ten years. I found Brad Feld and Jason Mendelson’s book and online class Venture Deals to be incredibly productive for both early stage VCs and entrepreneurs. Corporate Finance is a strategy just like marketing. As Charlie O Donnell pointed out in his blog, investing at higher valuations at seed really impacts the final IRR. That’s sort of a “duh” statistic but really impactful when you think about the entire funding lifecycle of the company. I don’t think some entrepreneurs and early stage investors really understand it. For example, if you are an institutional investor, why are you writing smallish size checks? It’s about how much equity you own. Not every company is going to be Uber.
There was a well-discussed Kaufmann paper more than 6 years back about the sector as a whole – “We have met the enemy and he is us”.https://papers.ssrn.com/sol…But the increasing monies and valuations have continued. The math of lower returns is very simple, but understanding the math does not change the behavior.
As Charlie O Donnell pointed out in his blog, investing at higher valuations at seed really impacts the final IRR…(snip)…I don’t think some entrepreneurs and early stage investors really understand it.This could be in part because of the way they view business and whether they see it as numbers or as ‘a good idea and a good person will hustle and it might work I will take a chance’. More seat of the pants and not ‘return’.There is a person who used to comment on this blog who I was speaking to in person a few years ago and he said to me ‘maybe I will sell 10% of my business for $300k what do you think about that?’. (Everyone here that comments frequently knows of this person btw.). Not sure if he was pitching me or simply asking what I thought. What I did think though was not anything about any financial formulas etc just whether I thought the business was good (it was operating and I had tried the product) and whether this person could pull it off or not. Nothing else. I would suspect that a great deal of seed funding rolls the same way. Sure you are gambling but in the end it’s part investing and part entertainment throwing around small sums of money. Do you think that when Ron Conway wrote checks in the early days (from what I hear) he thought anything but ‘good guys good idea I will gamble on this’?
Big difference between being an angel-like Conway was in his early days-and being a seed fund with LPs. When you are an angel, you can do whatever the hell you want for whatever reason you want. When you are a seed fund with outside LPs, you need to generate return for them.
Very true. This reminds me of the Zenith tag line ‘the quality goes in before the name goes on’.  Reason being is that when you are dealing with larger numbers and more deals it is not practical to fly any other way. In other words if, at a small scale the strategy (which Conway apparently is returning to) works, why wouldn’t you try to do the same at a large scale? Because you can’t. So you need some other proxy. Numbers which means quality goes down. You also need to be able to (and this is really important) justify what you are doing ‘I am working for the money I am making for you see what I am doing it’s not seat of the pants!’ <– Marketing and we can call it ‘proof of work’. Dog and pony show. https://youtu.be/uYPO3PFTf0…
“BRUTAL”.Can be.So, it’s about MONEY.For an entrepreneur, they can get money from (1) investors or (2) customers.The US, border to border, villages to the largest cities, is awash in many thousands of companies, each 100% owned by an entrepreneur who never took equity investment. Really, those thousands of entrepreneurs have found it easier to get money from customers than from investors,I know a lot of successful entrepreneurs: Commonly they use severe adjectives but not as severe as “brutal”.Looks like, net, a lot of entrepreneurs are having an easier time making money than the seed investors.There’s a question about the graph: If take the graph at face value, the seed investments range from a little under $1000 to a little over $7000.Ah, come ON, guys! In my neighborhood, the guys who mow grass show up with a late model pickup truck worth ballpark $30,000, with a trailer worth, what, maybe $2000, with one or two very capable riding lawn mowers worth about $15,000 each, plus other equipment. So, before they leave for the grass to be mowed, they have capex of nearly $75,000. And I’m sure they didn’t get VC equity funding.That’s just grass mowing. Can get comparable numbers, often much more, e.g., for a McDonald’s franchise, all along Main Street, USA.So, what’s with the $1000 seed funding? That counts kindergarten entrepreneur front yard lemonade stands?
How can anyone trust a chart where you can’t even see the methodology or vet the data source??? In what way is this to be taken seriously or as anything more than a basis for a discussion here (which is what is happening).Grove Street Advisors, a well-known LP, to share this graph they produced using their own data along with Pitchbook dataAnd what does that mean ‘using their own data along with Pitchbook data?’. How exactly did their data differ from Pitchbook? How was it combined?If you do a google image search you will see it’s now being passed around as meaningful in some way. At least you have correctly rained on the parade here. In a nice way. That is the larger lesson. If you are going to read something anywhere think a bit about what you are reading and stop assuming what I call ‘the assumption of legitimacy’ that what you are reading is correct. At least if you have to make decisions based on it.I am sure no investor would ever make an investment decision seed or otherwise based on something presented this way.The theme of course makes sense. But trusting w/o questioning is also the cornerstone of a great deal of social engineering, marketing, advertising, relationships and exactly how people get into trouble in life (not asking questions or assuming someone else has already). <— Really.
That is precisely why I said I won’t debate the chart. 67.9% of statistics are made up. (of course if anybody doesn’t get the joke, I made that up) This chart is made up, but not directionally wrong.
FWIW I don’t think Grove Street would 1/ intentionally misrepresent findings and 2/ I’m sure they have some disclaimers. If we zoom out, to me it’s less about the actual number and more about the direction. Many of the Top Tier VC funds in California have kept the absolute number of their Series As per year relatively consistent despite the increase in company formation.
In no way do I think that Grove Street would do this intentionally either. That is what surprises me even more about the process they used to come up with the chart. Which I thought prior to even reading what Fred said.I don’t think that it makes sense to use a chart to support a general thesis about direction with numbers that can be attacked in the way that I have so easily. As an example at a trial in a courtroom any junior defense attorney would totally be all over this chart if some expert were up on the stand. Sure this is not a court room and there is no trial. And that expert would never have even gotten to that point.I have been sharing this graph with friends over the last two weeks, and they all feel itOne thing I will note is this and I hope you take it as a compliment Semil. If you are a ‘somebody’ (and you are a somebody and so is Fred) I think you have to be more careful in terms of what you present than, say, me in the comments or JLM talking about history. And you will note that even I preface much of what I say with all sorts of disclaimers etc. Interesting I take what JLM says about history as being very correct but then again I don’t based investment decisions on that info either.
Yes, I should have added a bit more of a disclaimer on my site when I posted… part of it is that I haven’t been paying attention to this world until 2011 so the earlier data I didn’t have a baseline to interpret. Thanks.
Yup.Let’s see: Due to the magic of computing, the Internet, Web sites, Disqus, etc. right away I found athttps://avc.com/2018/05/a-c…my post, of 12 days ago commenting on a Goldman-Sachs graph, with in partFourth, to take any such graph at all seriously we MUST have clear references to primary sources where we can look up and confirm the data. This graph has no such references.What a shovel full. I can’t beat back against all the insects, vermin, sewage out there and, instead, have my own work to do.For your But trusting w/o questioning is also the cornerstone of a great deal of social engineering, marketing, advertising, relationships and exactly how people get into trouble in life (not asking questions or assuming someone else has already). <— Really.Yup.
Some people make all the stuff up all the time.Some people make some of the stuff up all the time Almost all people make some of the stuff up some of the time.A shocking truth that often escapes people is that almost everyone makes some of the stuff up even it it looks like the data is well sourced; everyone thinks someone else has vetted it and asked the tough questions; it is convenient to ignore underlying fuzziness of data if the conclusions confirm the narrative they want to push.Have deep skepticism of all data and infer directionality at best only if correlated with multiple other independent sources.
My understanding is that Pitch Book relies on self reporting. Like hedge fund indicies, there is a heavy bias towards companies who have something positive to report. I agree with Fred’s criticism that there are grey areas of funding that were out into a Seed or A bucket; however, I think this chart may understand the gap b/t Seed and A.
Hi everyone. Fred, thanks for posting & reigniting the discussion. The chart is from Grove Street Advisors, who took Pitchbook data to make this. I published with their permission. For me, this was an interesting graph for a few reasons:1/ I am looking at the zoom-out. The tidal wave of seed funding is amazing, even if the Series A conversion % line would flatten out or be higher.2/ Of course, it’s impossible to get great data b/c private sector data like this isn’t reported, and what about SAFEs or notes, or what is a Series A these days anyway?But… I would argue the following — take the Top 20 VC firms in the world. Now, from 2010 forward to today, I would argue the overwhelming majority of them are doing anywhere between 10-20 Series A deals per year, consistently. I have verified data from two Tier 1 Valley funds that I am not in a position to share publicly, but they are between 10-20 every year. If we extrapolate just to those Top 20 firms, I wonder how similar. Directionally, if feels so….We will never know for sure. And there are two ways to interpret this. Option A is to say there may be more new company experiments and seed funding, but only a finite set that ever get to Series A readiness — or Option B, there is room for newer funds to take on more Series As or for the current funds to increase Series A capacity.
Hasn’t the median size/value of Series A deals gone up materially ?
Why focus on only the top 10-20? There is a lot more money out there than just those firms
Yes, this is the debate I ended my initial comment with. There are two interpretations, depending on your POV. Some see this as an opportunity for new capital sources; others only pay attention to the top cohort of funds. What I hear in private from many experienced LPs is that they look to these top funds as a signal for Series A. Yes, they could be wrong. Another way to put it is — does one adhere to a “classic” definition of what it means, or is there a new definition for Series A that has emerged? I don’t know the answer.
People who can’t think for themselves and care about what others are doing are destined to underperform
I don’t think that is entirely fair in this particular context. For LPs, most of them have their hands in many things, and looking at earlier investors as a sign of quality might actually make sense, at least historically — for instance, current breakouts like Discord, Bird, Coinbase, Limebike, Roblox, Wag, OpenDoor, Duo, Carta etc all have top firms somewhere on the cap table. Of course, those firms don’t always get it right (and not including crypto here) but it feels too early to tell which way things will go.
We’ll said. Do you mind if I quote you on that?There is an unsaid ”unless you are extremely lucky.”
Top 10 to 20? Exactly what is the definition of ‘top 20’?. And is it firms or the actual VC’s?https://www.nytimes.com/int…https://pitchbook.com/news/…https://www.cbinsights.com/…https://www.nanalyze.com/20…https://www.streetwiserepor…
As I read the chart, I see that there was a lot less money floating around in the early days. So is it possible that the seed rounds raised in 2006-2010 were of higher quality due to the difficulty level in raising seed money at the time vs. it being easier in the later years, as more money became available. So, as more money flows into seed, the quality of seed-stage companies decreases (due to a some seed stage investors being more new to the game). The A rounds remain consistent (as they tend to attract more experienced investors), which leads to a lowered conversion rates due to more companies at the top of the funnel failing to meet the basic requirements of what a series-A company has historically been required to meet.I wonder if there is an equivalent chart for series A/series B in the dotcom and I suspect it may rhyme with this.
I see two things that you didn’t discuss. First, seeds are generally easier to raise for a variety of reasons but imho having mainly to do with the overhead associated with dealing with VC firms. The fact that the seed rounds are getting larger exacerbates that. That also speaks to the comment about second seeds and pre-series A etc. Entrepreneurs will go to the path of least resistance.Second, I think its an indication of the lack of viability of startups coming up over the last 10 years or so. I’d like to see the breakout of the types of companies being funded. My guess would be is mostly web sites that are variations on what has gone before. The web is getting harder to make money on in the same way that package software is now basically a thing of the past. If the seed rounds are getting larger, the expectations for revenue at the end of that round are going up as well. Most of these (mainly silly) websites don’t and probably can never make any money. By the time you get to the end of a big seed round, that going to be pretty obvious and once the rigorous VC due diligence kicks in a lot are not going to make it.
> The web is getting harder to make money on in the same way that package software is now basically a thing of the past.Even if true on average, that is not very meaningful for one entrepreneur with information in addition to just the simple random sample for the estimate of the expectation of the ROI.How much money can be made with a Web site can depend a lot on what’s on the Web site! There’s no law that says that an entrepreneurs’ efforts have to be just… variations on what has gone before …Or, a Web site likely uses the x86 instruction set, maybe Microsoft’s software, HTTP, HTML, CSS, maybe Redis and SQL Server, etc. Think of those as commodity glass bottles that do not limit the value of the contents, either just bottled spring water or what the Warriors spent $400,000 on to drink after their victory! Or a bottle might contain very valuable pharmaceuticals.package software is now basically a thing of the pastLast week, on the computer I’ve built, I “authenticated” Microsoft Office 2003 twice, Microsoft Windows XP 32 bit Professional SP3, and Microsoft Windows 7 64 bit Professional SP1, RealPlayer, PhotoDraw from Office 2000, Web browsers from Google, Mozilla, and Brave, printing software from Brother, and the Microsoft .NET Framework 4, etc.Uh, there are some big advantages in Office 2000 and Office 2003!! E.g., Office 2000 has PhotoDraw V2, and it’s darned nice to have and apparently the last such from Microsoft. I did the logo for my startup with PhotoDraw V2. It’s my way to create GIF and PNG files. I just did a scale model of my office as part of rearranging — I get to slide the pieces around in PhotoDraw, easier than moving the real physical pieces!Office 2003 has Outlook 2003 which is okay: I have fully detailed notes, click by click, on how to configure the thing, how to control what PST files and where that it uses (e.g., not in the default location of a hidden file system directory deep in the directory tree of the boot drive, how to get around the problems with “contacts” and “addresses”, etc. For later versions of Outlook, the problems with contacts and addresses are even worse.Long ago I wrote my own POP3 e-mail software, recently didn’t have access to Outlook, used my old software again, and discovered something recent and bad: Now nearly all e-mail software sends SMTP e-mail with the content as HTML, some quoted printable thing with some goofy character set tricks, lots of MIME (multi-media internet mail extensions) parts, etc. The actual number of bytes sent is huge, many times larger than any sender actually typed in. Since in Outlook I read all e-mail just as “plain text”, I didn’t see all that extra complexity, but with my own e-mail I had to cut through, translate, etc. all that cruft. Outlook 2003 does well with that cruft handling; the Outlook 2003 code does a LOT of good work getting to “plain text”.Yes, one reason for less “package” software is that now people download from the Internet. Internet users are awash in offers to download software for $50-$1000 or so. E.g., now a download of Sketchup costs $695. The recent, high end versions of Nero, RealPlayer, etc. all cost. I’m sure Adobe still has lots of high end software that is expensive. SAS used to be expensive; I doubt if it is free now!There’s lots of software for sale, but now commonly it’s downloaded instead of shipped in shrink wrapped boxes!Computing is not quite the really simple garden you seem to assume.
https://techcrunch.com/2018… Just your average daily $14 billion dollar Series C
More like a military budget
Fred (and/or others): Please define “Seed” and “Series A”. Insights please!Per Fred’s “Valuation Inflation” post a week or so ago, the average pre-money valuation for a seed round is now $10-15mm in the 2017 time frame and the average amount raised in seed rounds is $4mm+. That sounds like a Series A round to me (defined by size and valuation) yet it is called a “Seed” round. Why?
I don’t think this graph shows anything contrary to your anecdotal evidence, Fred. The trend from 43% to 36% to 29% conversion to Series A hides the absolute numbers. The denominator (or number of seed deals) is going up, perhaps massively, given the increase in the seed funds. So it is possible that the same number of (or if we want to be aggressive, maybe more) seed deals are converting to Series A, but the huge increase in the denominator causes the percentage to go down. It’s just math…..or am I missing something?
@sigmaalgebra:disqus raised an excellent point regarding the numbers appearing to be super low for seed investment (in the context of what is being discussed). How can the number be that low $3k who is writing checks for $3k other than someone’s Uncle. Even if the scale is off (in 1000’s) it doesn’t make any sense at all. What’s going on here @semilsha ?
This looks the great preface page to a new fund….The Security Token Series A Fund.
This could also be because people have better understanding of what will scale and what will not. So they dont go for further rounds. We did not for this reason.
>Has that percentage gone from north of 80% to south of 30% in ten years? No way.I wonder if your view on this is biased due to the success of USV companies? I hear of a lot of companies not finding fit, partnerships dissolving, etc… The more common one that I feel I am seeing more often these days is that there seems to be a lot more “seed investors” out there these days funding projects that should have never been funded. Also, the decline in “app innovation” has been declining.I would love to see the stats on this from strictly a USV perspective. I would hypothesize that USV portfolio companies would not fit the stats that Semil presented.
It seems like seed rounds used to be $500K-$1M. Now they are $3-4M. Series A rounds used to be $5M, now they are $10M+.
The concept of “startup” has also blurred. Some tech startups are more ‘tech enabled [traditional] businesses’ than startups in the venture funded, high-growth dominate-a-market sense.This change is for a number of reasons, but include the democratisation of tech en general, the ease of which to start a company (lower tech barrier to entry, regulatory changes and new online tools to make starting and operating a company easier), the encouragement by governments and other institutions for people to become ‘startup’ entrepreneurs, as well as the distorting impact of angel / early stage money which often comes from inexperienced angel investors who prefer profitability and stability, over growth and risk.One example is EIS/SEIS schemes in the UK. Although on balance this has been a hugely positive initiative and continues to be, as it has increased the funding available to ‘get companies going’ enormously, often the strategic advice given by Angels providing this money, and sometimes the Founders taking the money, is not ‘high growth’ advice nor is that adopted as a strategy. And even if it is the strategy, the startup either picks a too small a market, or never reaches the velocity, in order to attract Series-A funding.In summary then I believe the number of Series-A funded companies has gone down; but the real question is how well was the initial data set screen to include or exclude businesses which were never startups in the silicon valley sense, and were instead ‘entrepreneurial tech enabled businesses’ which happened rightly or wrongly to raise angel / seed money of some description.As Fred says, the data is messy and there is much more variance in the way startups are funded, and reach an institutional or Series-A round (whatever that means) these days.
I did not look at the methodology behind the data, but it may also be a taxonomy issue. The consensus around what a seed round is has shifted over the years. I think there are many more seed rounds these days that would have been classified as A rounds 10 years ago.
The sharp decline in Seed to Series A conversion could be the result of increasing seed investments by less experienced investors, leading to deteriorating quality of seed deals.