According to my usual data sources, 2014 was not nearly as good a year in seed funding as I would have expected from reading news reports. I was fully prepared to see “bubbly” data. However:
- From angels, seed stage funding plunged nearly 50%, from $11.2B to $6.0B.
- From VCs, seed stage funding dropped about 25%, from $940M to $740M.
(While VC early stage jumped over 60% and expansion stage spiked over 110%.)
- But median seed valuations increased 20%, from $2.5M to $3.0M.
The run up of a bubble does not typically include sharp volume drops, but the price rise may indicate something interesting is going on.
See here, here, here, here, here, here, and here for previous posts in this thread. My data sources are the Center for Venture Research for angel data, the NVCA for VC data, and my personal tracking spreadsheet for “super angel” funds not part of the NVCA. I use the HALO report for pricing data.
Out of concern that one possible explanation for individual angel investments dropping is a shift to angel funds, I reconstructed my angel fund tracking spreadsheet from scratch. I was worried that my list of funds was too haphazard. So I pulled a longer, hopefully more complete, list of micro VC firms from CB Insights. I then removed ones that are members of the NVCA, whose investments should already be included in that data. I also went through fund CrunchBase listings and Web sites, filtering out those who invest primarily outside the US or not at the seed stage. I noted the reason for any such exclusions in my latest spreadsheet. This updated source contains 74 funds totaling $3.2B, while my 2013 source contained 29 funds totaling $1.2B—a substantial increase in coverage. There were also 23 funds on the new list for whom I could not find dollar amounts. However, I assume each fund with dollar amounts is completely deployed in the current year so the total should still be a gracious estimate given that even very fast funds actually deploy over two or three years.
Despite the addition of $2B in covered super angel investments, the 2014 graphs were still sobering with a 25% total volume drop year-over-year:
The big question is how do you get a pronounced volume drop and a pronounced price increase? First, one of the data sources could have a problem. The obvious candidate here is the CVR angel data because it accounts for most of the total volume I track, the methods aren’t documented, and this is its biggest one-year drop ever. According to the latest report, total angel investment volume was only down 2.8%. But the proportion of seed and early investment plummeted from 45% to 25%. I can think of several reasons for a potential measurement inconsistencies here. Note that the 2012 measurement was 35% so it has a lot volatility.
Similarly, the CB Insights pricing data could be the result of an anomaly, as it also experienced its biggest one-year move. In contrast to CB Insights’ 20% jump in valuations, the CVR valuation data showed a slight decrease across all stages (but the CVR valuation data is reported inconsistently, so I have avoided using it in the past). Of course, the CB Insights and CVR data collection methods could somehow result in systematically different samples that explain the conflicting data.
There’s another explanation that I find tantalizing, if only because it would confirm my hypothesis that founder opportunity cost drives the earliest company valuations. Think of seed stage companies as “supplying” investments and investors “consuming” them. Econ 101 says that a simultaneous decrease in volume and increase in price implies that the supply curve has shifted left. It’s like a freeze wiping out a significant fraction of the orange crop. People buy less at higher prices because there’s a shortage. But why would this happen now in the seed stage startup market?
I have a guess: the macroeconomy recovered. Unemployment eased from 7.5% in June 2013 to 6.1% in June 2014. Moreover, according to Indeed.com, software engineering salaries jumped 20% in 4Q2013. All of a sudden, the opportunity cost of founding a seed stage startup went up dramatically. That could definitely have an effect on formation rates, which would show up first in the seed stage funding data. Another “supply side” explanation would be that founders simply need less money to advance their ideas past the seed stage. The total number of angel-funded deals actually went up 3.8% according to the CVR report.
At RSCM, we’ve certainly seen no shortage of quality opportunities at low valuations. If anything, we are deluged. Of course, we purposely focus on smaller deals so we wouldn’t expect to see any shortage if lower capital requirements were the underlying cause. Our experience is also consistent with a data collection anomaly. I’d love to get my hands on a good dataset for accelerator program application volume. That might allow us to distinguish between declines in formation rates and capital requirements.
Bottom line: I’m still very skeptical that there is a seed stage bubble.
So there are finally some signs if life in seed-stage technology funding:
- In nominal terms, we roughly equaled the global peak from 2005.
- Seed stage valuations have remained flat since 2011.
- Adjusting for the size of the economy and our wealth, the level is still down.
Kind of hard to call this situation a “bubble”. But I can live with calling it a “recovery”.
Once again, see here, here, here, here, here, and here for previous posts in this thread. My data sources are the Center for Venture Research for angel data, the NVCA for VC data, and my personal tracking spreadsheet for “super angel” funds not part of the NVCA. For super angel investment, I worry most about detecting new chunks of money, not necessarily measuring the “true” level. I use the HALO report for pricing data, which goes back to 2011.
Here are the graphs (spreadsheet here):
It looks like all the components are recovering, though traditional VC somewhat more slowly and super angel somewhat more quickly. The question is still, “Bubble or no bubble?”
Let’s look at prices, as I did in my 1H2013 post. According to the full year 2013 HALO report, the median seed valuation is still $2.5M… just like 2012… and 2011. The 75th percentile valuation is up slightly in 2013, from $3.7M to $4.2M. But the 25th percentile valuation is down a hair from $1.5M to $1.4M. According to the methodology described in the report, this data includes angel group deals before Series A. So what I think is happening is that some companies that might have gone for a VC round in the past are doing a larger angel round instead. If you check out my spreadsheet, you can see that check sizes for what the NVCA calls “seed” have taken another swing up, probably pushing some early startups out of that market. So no obvious pricing pressure.
Moreover, I think the following graphs make a bubble quite unlikely. I’ve been waiting for years to pull these out. The first one “deflates” the seed investment levels by adjusting for GDP. Thus it measures how seed investment has changed relative to total economic output. The second one deflates seed investment levels by adjusting for the level of the S&P500 index (on July 1 of the given year). Thus it measures how seed investment has changed relative to the total stock of wealth.
Compared to our economic output and total wealth, seed-stage investment seems like it still has a significant amount of headroom. I’m actually pretty sure I could build a darned accurate forecasting model based mostly on the S&P. Given that the index is up roughly 25% from July 2013 to July 2014, my eyeball estimate is that 1H2014’s numbers will show us somewhere around a $16B annual rate.
Well, it’s been almost three years since I started watching for quantitative evidence of a “bubble” in seed-stage technology funding. I feel like a broken record saying there’s still no sign. Here are the highlights:
- 1H2013 volume is 30% below the 2005 global maximum
- 1H2013 volume is 10% below the 2011 local maximum
- Seed stage valuations have been flat since 2011
You simply don’t have a bubble when volume is down and prices are flat!
To review the history of my seed bubble watch, see here, here, here, here, and here. Recall that I use the Center for Venture Research’s angel data, the NVCA’s VC data, and my personal list of “super angel” funds not part of the NVCA. The volume calculation methodology is not designed to produce the most accurate estimate of the true number of seed-stage dollars. Rather, I want it maximally sensitive to sudden influxes in new seed money. I use the HALO report for pricing data, which started coming out in 2011.
That said, here are the graphs (spreadsheet here):
The story continues to be that traditional VCs have become increasingly irrelevant as their seed dollars have dropped 60% from 2009 to 1H2013 and their share of all seed dollars has plunged from 22% to 7.5%.
Angel’s position has gradually eroded from 2011 to 2013, with their share decreasing from 88% to 77%. Super angels and seed funds have gained in share during that time, jumping from 3.0% to 15%. My guess is that trend will continue unless the individual angel pool increases via new platforms like AngelList. In any case, the new breed of funds is not growing fast enough yet to make up for decreases from other sources.
[Edit 8pm: Somehow this paragraph got deleted from my draft.] There also appears to be no pricing pressure at the seed stage. According to the 2012 and 2Q2013 HALO reports, the median seed-stage pre-money valuation has remained $2.5M since 2011. Moreover, the 25th and 75th percentile valuations have actually decreased, making it hard to argue that there is some hidden dynamic masking a buildup in prices.
Interestingly, the HALO report shows a continued drop in California’s share of angel group activity. From 21.0% in 2011, to 18.1% in 2012, to 17.3% in 1H2013. I’ll take this as continued confirmation that RSCM is right that some of the best values are outside the Bay Area.
It will be interesting to see what the data shows for 2H2013 and 1H2014. With the S&P reaching new highs throughout 4Q2013, institutions should increase their allocations to alternative investment funds and angels should feel like they have more wealth to invest in startups. Assuming the public markets don’t experience a sudden drop in the beginning of 2014, of course.
While my goal is to eventually apply the Market Space model to large enterprises, I’m going to begin with startups. Obviously, my work at RSCM makes startup close to my heart. And most large enterprises were new entrants at some point, so analyzing the birth of firms seems like it should lay some crucial groundwork. (For previous posts in this series, see here: one, two, three, four.)
Looking at the search for profitable products as a Multi-Armed Bandit (MAB) problem illuminates the general complexity of the firm’s challenge (see previous posts in this series: one, two, three). But in terms of analyzing specific firm behaviors, I think it’s important to acknowledge that we don’t have a pure MAB here. It seems pretty clear there’s more causal structure in Market Space.