Given the priority about getting as many customers as you can, what role does marketing play? Should it have a higher and earlier profile to drive brand awareness and customer acquisition?

aripap It is a good counter point.  In a way it completely depends on the market environment.  Right now the world is focused on growth and in this environment early revenues are less material than early users.

Interesting post.  However, I would argue that enterprise SaaS doesn’t exhibit the same winner-take-all dynamics that consumer does.  None of the recent IPOs (with possible exception of Veeva which is vertically focused) are really dominating their markets.  
Also, I think it’s important to price enterprise SaaS sufficiently high to signal quality.

There are two kinds of b2b saas products, the cheap (per seat) and cheerful products like e-signature or online storage and the more complex ones that require some setup and integration. If you are selling the latter, a too low price point will scare away many of your best prospects.

markorgan This is a great comment.  I spoke a number of times with Joe Payne after you left Eloqua, and they had the same problem I described above about declining prices on renewal business because Marketo’s entry strategy was to win at any price.  In a market with a surplus of capital it is often better to be a fast follower.  I agree that network effects are great if you can get them.

CFederman Charlie, I will answer the second question first.  Our model is predicated on the idea that if you plot the expected exits of our investments at the time of a Series A that they should have a similar distribution as other venture-backed companies at the same stage of investment.  We call this the base rate.
The base rate in our industry over the past few years has been an average successful outcome of $120MM (via M&A), and a median outcome in the 50s.  At the time of our first investment, when revenues are small and the risk is large, that seems like a fair set of assumptions.
I would like to think that we do a better job picking winners and growing companies than the industry average, but it is unrealistic to assume that every company we fund at the A round will be a 99th percentile outcome.
A lot of VCs are putting big dollars at risk, at high valuations, in very early stage companies.  We would normally call these deals $3-4MM Series A rounds, but the dollars coming in look like a $15-$20MM growth financing. When VCs make that sort of investment in pre-revenue, pre-traction companies they are making the implicit assumption that the return profile will somehow defy the odds over time.  That strategy has not done so well historically.  It is one thing we try to avoid.
Once companies get past the Series A and Series B rounds and have demonstrated revenue then they get into a different set of outcomes.  It is the same as life expectancy – if you are 90 years old your life expectancy is higher by virtue of the fact that you have already passed the average.  We are much more comfortable writing bigger checks at this point and adjust our expected outcomes proportionately.  We have funded companies at $300MM pre-money valuations before and still made a 3x+.
On to your first question, which is harder to answer.  It is true that our successful companies will consume a lot more capital than we initially anticipated when we set off to build greycroft.  The good news is that there are a lot of investors who want to write growth equity checks for later stage companies.  That means we don’t have to reserve all the way to the end.
Our model suggests that the additional dollars are better put to work by making more A and B investments than by being a “lifecycle VC” and reserving 10MM for a potential growth round five years from now for every 1MM series A investment we make today. Plus it is hard to forecast these rounds in an early stage fund, so we would likely end up with either too much reserved or too little reserved.
The bad news is that we will likely get diluted in later rounds as companies raise more money. Ideally the valuations are high enough that it won’t create a significant damper on our returns.  It is also an opportunity for us some day if we want to do a growth vehicle because we have accumulated a lot of pro-rata rights.
I hope this all helps.

dcoppins I think a handful of factors determine what the final market share allocation will be among the top players in a specific software category.  The big factors that come to mind are switching costs, barriers to entry, customization, speed of on-boarding, amt of required integration, and cost of support.  The best companies are trying to maximize the first two factors above and minimize the next four.
SAAS products today have almost zero marginal cost, while on-premise software products had hardware costs back in the day.  There is probably an economics PhD dissertation in here somewhere…