I have become fixated, if not quite obsessed, with thinking about what will happen when the venture spigot gets shut off. I having been writing a lot about this (and here). There is a lot of uncertainty out there on this subject, but I have boiled it down to three things about which I am fairly certain:

1. At some point, it will become much harder to raise a venture fund and thus much harder to get funded by a venture investor.
2. This is normal. The rest of Capitalism calls it a business cycle, and it does not have to end with a “Bursting Bubble”, or a post-cataclysmic landscape.
3. It is very hard to predict who will suffer most in the downturn. There are lots of interdependencies which means that problems in one area can cause much bigger problems in another.

Beyond these, it is all open to debate.

I have been going back and forth on this subject with a few people lately. And there has been a growing chatter in the venture community on this subject, largely spurred by Benedict Evans’ piece (my response here). Fred Wilson at Union Square Ventures had an interesting post on the subject as well. His views are very similar to mine – capital flows matter. In fact that was one of the first posts on this site.

Of course, the fact that post is over three years old should make it clear that predicting the timing of downturns is impossible.

Nonetheless, I think it is worth exploring more, to prepare for that inevitable downturn. The world of finance has been grappling with this subject for a long time. (Some would say the whole basis for finance is the preparation for downturns.) And a key concept in managing for risk is the idea of diversification aka not putting all eggs in one basket. For instance, do not put all retirement assets in one stock, because any number of random events could wipe out that company. Better to buy multiple stocks (or a mutual fund), and better still to buy stocks and bonds and real estate etc.

At its heart, this is the idea behind a venture capital fund – lots of companies in the fund may go bust, but no single disaster can wipe out the fund, giving time for the other companies to find their inner unicorn.

For companies, however, diversification is much harder. We do not like to invest in companies that do too many things – they lose focus.

So it seems to me that understanding the risk landscape for the Someday Downturn is to understand where companies are exposed to risk. I tend to think of this as being ‘levered’ to something. Start-ups are levered bets on some form of technology. They are betting big on a particular outcome.

Of course, there are all kinds of risks out there. So the question is determining which risks will be exposed when venture funding becomes harder. Evans argues that companies today are much less dependent on venture funding because they are more capital efficient and thus much closer to breakeven. I think it depends on what the company is trying to achieve.

Take Uber as an example. They have raised a lot of money, but if the venture spigot turns off tomorrow, they already have a massive business. And it is geographically hedged – meaning that even if the Bay Area economy tanks their business in other parts of the world will likely not be affected. By contrast, there are a lot of developer-focused companies out there: dev tools, ad networks, etc. When it becomes hard for app developers to raise venture money, they will cut back on all of these things.

I think this area will be one of the big victims of Rule #3 from above – the Law of Unexpected Consequences. For instance, one of the more interesting areas today is ‘Big Data’. There are a lot of really interesting companies out there building Big Data applications, platforms and analytics. However, dig a little deeper, and it quickly becomes clear that the biggest consumers of big datasets, and thus Big Data tools, are ad networks. We could easily see some very promising Big Data companies disappear because they are much more levered to these sorts of vulnerable sub-sectors.

There are also businesses which do not scale in a linear fashion. For instance, enterprise software companies need to build up a large sales force, find reference customers and close very complex deals. These companies need to go a couple years wholly dependent on venture funding in order to reach critical mass. Admittedly, conditions are much easier today than they were even five years ago. There is a lot more flexibility in enterprise software adoption patterns. Nonetheless, it should be clear that losing venture funding is going to be harder for some types of companies than others.

Let me conclude by repeating that I do not want to come across as alarmist. I have no idea when the business cycle will turn. And it is clear that many parts of the industry will be able to escape unscathed. Nonetheless, I will stand by my statements above. Things will change. It may or may not be a catastrophe. But there will be a lot of unexpected casualties along the way.