What is early stage?

Being out doing the venture capital dog and pony show now, there are a lot of things I'd tell you were pretty frustrating, despite the fact that our process is going pretty well.  When we do close our round, I'll post a lot more about it.  However, the one thing I would have to say has been the most bewildering has been the divergance among venture firms definition of what early stage is, and what they're willing to bet on.

Virtually all of the VCs we went to wanted to see that we had built a product, which we have, but then the question came to traction--and it wasn't at all clear what kind of traction was enough and what it proved.

When I was at Union Square Ventures, we used to say that we wouldn't invest in a consumer facing web application without any consumers.  In fact, to quote Fred, "I coined that phrase."  :)  The question, though, was what exactly that proved?

In my mind, it eliminated execution risk--that you wanted to back a team who was able to complete a functional product.  We didn't want to take the technological risk of whether or not the stuff worked--and that having a team who could get product out the door was a good filter.

What it did not prove, to me, and what a lot of firms seem to think, was that early traction somehow proves adoption and sustainable consumer interest.

This thinking falls down for two reasons.  First of all, there are lots of web applications that fail to get beyond the initial "TechCrunch bump", or as Josh Kopelman puts it, the 53,651.  That was the number of TechCrunch RSS subs back in May of '06.  The interesting thing is that number is almost twice the number of users del.icio.us had when we started talking about an investment.  You never know if that initial ferver of the digerati will transfer over to a mainstream crowd.  Even at 100,000 users, that could still flatten out and never get past the geeks, and mainstream adoption is absolutely critical to success. 

Look at Notchup.  Everyone signed up for it back in the beginning of the year, pushing it past 150,000 monthly users.  I mean, who wouldn't want to get paid to interview?  The only problem was that it wasn't sustainable, and traffic has nearly completely disappeared.

You could say that startups should then have at least six months of before being funded, but then you get into the question of who's supposed to be funding that path.  From conception to launch, it probably takes most companies 9-12 months.  Tack on another six months, and now you're talking almost 18 months of life--a lot to squeeze out of an angel round, no?  Perhaps angel rounds now need to be at least 500k to get a company anywhere useful, because just building a product won't be seen as enough.

This emphasis on initial traction also fails to take into consideration iteration and the idea that most companies will not get it right the first time.  Sure, there are a select few companies that get it right from the start, but that's not the case with most.  Take Flock, for example.

Now, I don't know what they consinder success, but certainly they're doing better now, and seemingly sustainably so, than they were when they first released.  Who knows, perhaps another big product release could push their monthly traffic to a half a million uniques.  Either way, they were able to have the flexibility to go back to the drawing board and figure out how to improve their product.  We saw that with the guys from InfoNGen, who nearly went back to square one to iterate on their information product in the financial services space, making it much more lightweight and feedbased.  Without the right funding, they wouldn't have been able to do that.  What helped them was that they had done it before (iterating on a product to make it successful) so that was part of their previous M.O.

Products will, and should, change over time based on the lessons learned from getting users in the door.  How many users do you need to learn those lessons?  At the moment we've had 3500 people take the Path 101 personality test, without really marketing it at all.  Everyday, we learn a lot more about what people expect from the results, and what makes it easier for them to take and complete the test.  We've seen completion rates double just based adjustments made after the first few hundred people took it--and we're running that same process with other features that we're testing to smaller audiences. 

What we know we really need, though, is a good user experience expert and front end developer to help fix things, and that takes financial resources.  That in itself is an interest bet, though.  You could imagine that there are sites, if made easier, would take off with a better experience, but without being able to envision what those experiences look like, it's difficult to really count on that. 

This theory also discounts the value of marketing and business development.  A central focus of our distribution strategy is to syndicate our career tools out to partner sites--like professional socities, alumni groups, etc.  A good marketing and business development person can make that happen--but it's just a matter time and good old fashion pounding the pavement.  However, if you're supposed to have deals done, product complete in partnership quality form, users in the door, can you really even call it early stage anymore?

We didn't even talk about business model.  I had one early stage firm want us to be 12 months away from revenue--and not just revenue driven by putting ads up on the site.  Can you imagine if those same demands were put on Google, Twitter, or LinkedIn?  They never would have gotten anywhere, because they never would have been given enough runway to iterate on the product and get to a critical mass of users to make revenue possible. 

Twelve months from revenue generation hardly seems "early stage" to me for a consumer application--because ultimately, in a consumer app, revenue is going to be driven by someone else.  That means that not only do you have to build your consumer application, but then you need to build the monetization engine, too--almost a completely new application for your paying business users who, for example, might pay for data created by the site, or in our case, to search out the right candidates for employment.

Again, all of this stuff is obviously critical and I'm not meaning to be dismissive about it, but startups our out there trying to figure out how much blood they're supposed to squeeze from the stone of an angel round.  Should they expect to raise multiple angel rounds, or just really large angel rounds?  

The interesting thing is that the advent of the larger (500k+) angel round may be shooting some venture firms in the foot.  Companies raising around 750k might be able to get to the point of building something worth acquiring by then--not for huge money, but certainly for enough to send everyone home happy because of the lack of capital sitting on top of them. 

The exception to all these rules is when a VC firm writes a multimillion dollar check to a repeat entpreneur to go after a particular space without even having a product yet--like when Benchmark funded Zillow.  The founders of Zillow were behind Expedia and so it was basically assumed that team could find success if properly resourced, in a completely different industry.  It certainly seems plausible, but then it makes me think of a friend of mine whose first startup was a blowout success, whose team went on to create a product in a different market.  What she's realizing is that she didn't really learn a lot the first time around--that when they built a product that caused the phones to ring off the hook and revenues to start pouring in, no one really questioned it, and so no one really learned much about why or how they were successful.  Those lessons would have to be learned to enable repeatability.  Her team is struggling now with their second go at this.  I think if I were a VC, I might try my hand at finding good teams who have failed once before--but then we dive into what makes a good team if not good results, and that's a whole other question and a whole other blog post!