In case you missed it, Bill Gurley made some comments in an interview with the WSJ about how burn rates of current venture-backed companies have moved back to levels not seen since ’99. The idea of burn rates that high rings alarm bells, as there are few things more dangerous to a startup than burning through cash too fast. Many more people have echoed Gurley’s sentiment, including the likes of Fred Wilson and Marc Andreessen:
4/When the market turns, and it will turn, we will find out who has been swimming without trunks on: many high burn rate co's will VAPORIZE.
And it looks like they may be on to something based on our analysis of U.S.-based software companies, chosen since these are the companies mainly in question, encompassing 14,000 rounds (request a demo to see them yourself). PitchBook uses net burn rate (typically in a monthly time horizon) as defined by the amount of capital spent by a company that is not covered by revenue or debt. EXPENSES – REVENUE – DEBT DRAWN = NET BURN RATE.
$500,000 monthly expense – $300,000 monthly revenue – $0 new debt = $200,000 monthly net burn rate
This net burn rate can also be calculated by looking at the capital raising of a company. Since companies don’t typically raise money unless they have to, we can calculate burn rate without revenue and expenses by applying the equation: CAPITAL RAISED / MONTHS BETWEEN ROUNDS = NET BURN RATE. To keep this somewhat complex analysis simple where we can, we make the assumption that companies spend their last penny just as they raise their next round. There are a number of individual exceptions, and in reality, it is probably more like a 90% spend, so our numbers will be slightly on the conservative side as a result.
In this burn rate equation, there are two variables: capital raised and months until next fundraising. Let’s take a look at the time between rounds first.
Looking at the above chart, you will see that the time between rounds has held steady for roughly the last 10 years and is actually currently longer than what it was in the 90s. In the breakout of the last five years, you can see there, too, that the funding timeline is holding. Based on these charts, there is little indication that higher burns are forcing companies to raise capital sooner than they have had to over the last decade.
If burn rates are in fact increasing to levels seen in the heady days of ’99, we would expect to see the amount of capital raised per round to have jumped, which, in fact, we do and we don’t.
Median VC Financing Size of U.S. Software Companies
As you can see, on an inflation-adjusted basis, only companies with D+ rounds are actually raising more money than what was raised at the height of the tech bubble. So if companies are raising less or equal to what was raised during the ’99-era at longer time intervals, then burn rates cannot be at the level that they reached in those days. However, you will see that later-stage software companies in the last four years have been raising more than they were between 2003 and 2010, meaning burns have in fact increased since time between rounds has held steady.
The Smoking Gun
Bill, Fred, Marc and all the others, though, are not looking at the last five years; they are looking at today. Applying our equation of CAPITAL RAISED / TIME BETWEEN ROUNDS = NET BURN RATE we can calculate the median net burn rate across each series for just the last couple of years in order to find evidence of actual increases in net burn rates.
Burn Rate of U.S. Software Companies
Finally, in this data, we see the actual smoking gun of proof. Specifically, the net burn rate has increased for Series B and later rounds to the highest levels since the height of the tech bubble. Series D companies are currently burning at a rate of $1.82 million per month, C companies at $1.04 million per month and B companies at $660,000 per month. These represent hefty increases from 2013 of 69%, 66% and 22%, respectively. While it hasn’t been specified, the concerns being voiced are primarily focused on the more mature venture-backed companies. That seems to make sense as early-stage (seed and A) companies are not yet increasing net burn rates at the same pace, and we actually see series A companies burning less per month in 2014 than they have the last couple of years.
So there you have it—Gurley’s, Wilson’s and Andreessen’s assessment that burn rates are increasing is in fact backed up by the data. Although burn rates in the U.S. software industry are not yet what they were during the height of the 90s tech bubble, the trajectory has clearly changed. Is this yet another indicator of a bubble? Or is the need for higher burns a new reality as a result of the competitive power law that people are increasingly realizing exists in today’s software industries? Regardless, we will be closely tracking this and regularly benchmarking burns in order to measure the trend and impact.
*Updated Noon 9/29 to clarify burn rate calc used as the net burn rate. See this great post from Mark Suster for a detailed breakdown on gross v net burns.