Image Image Image Image Image




VC Math – Or why you’ll never get funding…

Yet another cleaned up post from a startup thread on LinkedIn.  This one is about VC math and TAM’s.

One thing that’s very much missing in most entrepreneur’s understanding of the fundraising process is the notion of total addressable market (TAM) and it’s relationship to fund size. Basically, it’s the understanding that fund size vs number of partners drives the available investment size.

An overview of how this works

Since partners can only typically manage at most 10 companies simultaneously, 2 partners at a $100million fund will invest in a maximum of 20 companies. Usually, the first five years of a fund are devoted to new investment, and the last five to supporting existing portfolio companies. Money is usually allocated 50/50 to new vs support, so this hypothetical fund would have about $50m to invest in 20 companies. $50m divided by 20 gives you roughly what the investment amount would be about $2.5m.

The flip side of this are rough calculations about TAM and exits. If the TAM of a company is $100 million and you assume they can capture 10% of the market in 10 years, then their revenue will be around $10m, with a valuation of somewhere between 3x to 15x revenue depending on the vertical. At the lower end of that, it would be a company worth $30m if everything works out.

Going back to fund size, if the fund is targeting a 10x return, then investing $2.5m in a company that might be worth $30m at exit is just about right, assuming no other dilution. Of course, this leaves out a lot of details and is uber-simplistic, but I can guarantee that every VC you talk to is doing these sorts of rough calculations in their head.

However, what’s really important about all of this is understand your company’s TAM and doing your homework about size and lifecycle of funds investing in your vertical. If your TAM is $100m and you approach NEA ($2b raised for their latest fund), then you will likely never get funding. A lot of entrepreneurs don’t seem to be willing to do the legwork in researching investors/funds/etc, but that’s one of the keys to success. I would add that the VC tool of choice for doing research is VentureSource – yes, it’s not cheap, but if you are looking for funding, it’s a goldmine of information on valuation and fund status

This is a very common mistake made by a lot of entrepreneurs, even experienced ones. I’ve given a presentation at a bunch of conferences about this very topic and it’s always shocking how many people don’t understand these dynamics.



Startup growth metrics for Silicon Valley

Last December, I was in an interesting meeting where the Bureau of Labor Statistics (BLS) presented a bunch of interesting stats several days before they were due to be released.

The most interesting statistic, however, was one that was generated specifically for this particular meeting and is not part of the regular dataset.  It was statistics about company growth in Silicon Valley, and it revealed the lack of companies growing to significant size.   For me, it just reinforced my gut feeling that the VC model was broken and that the real target should be lower exits, not monster hits.

First some definitions, in BLS terms:

  • Small company – less than 50 employees
  • Medium company – 50 to 250 employees
  • Large company – 250+ employees

I would note that the BLS looks at payroll to determine employees, so contractors don’t count.

The BLS had put together data about companies growing from small to medium to large over the 8 year period from 2000-2008, it was summarized in the following table:
[table id=1 /]
What’s surprising about this is that only 8 companies became ‘large companies’ in 8 years. That’s one a year, and if you think of company size as a proxy for revenue, then it’s pretty stunning.

P.S. Sorry for the crappy table, turns out tables in WP are a huge pain in the ass….  The vertical numbers represent the state of affairs in 2000, whereas the horizontal numbers represent the state in 2008, so reading from left to right gives you the changes from 2000 to 2008…



What I look for in a pitch

Over the years, I’ve gotten many requests from people wanting help raising funds.   Most of the time, I find that they have failed to do their homework and that even the most basic elements are missing from their pitch.   I recently posted in a forum describing what I would look for in a pitch, so I thought I’d re-post it here, a bit cleaned up and expanded.

What I look for in a pitch

A pitch is several stories.  It’s the story about how you will successfully run your company if you get money and it’s much like a sales pitch.   But it’s not just your company’s story, it is also the story of you and the people on your team, the industry you are targeting, the story of your customers and finally, the ending chapter of your company.   Like every good story, these should have arcs that meet at some logic point.   And, like most good stories, it needs to be cogent, logical and progress along path.   All that said, what are the tangibles that I would look for in a pitch?  Well, here are some:

  • Entrepreneurs should know everything about their business/market and I consider it a bad sign if I know even just a little more
  • Know who I am, why you are talking to me and what your expectations are from me
  • Have done their homework about investors, funds and criteria.
  • Polished presentations & financials
  • Business plans in the traditional sense are a waste of everyone’s time – I don’t really care if you have one
  • A good, clear slide deck that lays out the future of the company (the story, see above)
  • Detailed financials that are in sync with the deck
  • Polished, articulate delivery of both the deck and financials (you need to know this stuff cold)
  • A clear understanding of real and potential competitors, with a view of competitive differentiators
  • A clear and complete executive summary that someone can grok in less than 10 min.
  • A good website
  • Finally, there should be some sort of working prototype or demo, esp. if it’s software.

There are tons of online and offline resources for learning how to do all of this and there really is no excuse for not doing it.  I would also point out that I don’t expect everything to be in the deck, just a good, compelling story about the future of your business, industry and customers.  That said, I expect you or your team to be able to answer every question I ask, in detail if appropriate.  And if you don’t know, don’t make it up.

So there it is, a basic outline of what should be in the pitch.   I would point out that there is a lot of other detail about what should be in there and that you should understand what your target investors are looking for when putting together your pitch.



Choosing a VPS hosting provider

As part of my ongoing effort to reduce technology administration, I’ve also been looking at fully managed hosting.   Right now, I have a dedicated server at SoftLayer, which has been a fantastic hosting provider.   But, I have to manage the server, which means keeping up with all the updates, handling backups, etc.  It’s not a lot of work, but it’s something I’d rather not do.   Also, a dedicated server is way overkill for what I need, I think CPU utilization is at 0.01% or something.  And virtualized hosting technologies have come a long way in the last 3 years, making fully-managed dedicated-like machines widely available.

So, I set out to find a fully managed server, preferably provisioned as a virtual server since that would probably give me better backup.  There are a lot of routes to this, including cloud services.   It turns out that cloud services are actually substantially more expensive that what I pay for a dedicated box and few offer fully managed systems even at that high cost.   So I was basically left with traditional virtual private server (VPS) offerings.   In this space, there are basically two models – Xen or Virtuozzo, aka. OpenVZ (at least on Linux…).  Xen has the advantage of providing dedicated resources (memory/CPU) to the VPS, where as Virtuozzo has guarantees and bursting, sort of like cooperative multi-tasking.  Generally what this means is that Xen has fewer VPSs on a single server than Virtuozzo as there is no possibility to share resources between VPSs.  In practice, it also means that Xen VPS hosting is also quite a bit more expensive than Virtuozzo.

So, with that in mind, I set off to find some hosting providers.    There are a LOT of hosting providers, and the real difficulty is identifying those which have relatively good performance, service quality, price and features.  Here are the ones I seriously considered:

Most of these were extensively researched by reading forum posts (in WebHostingTalk and quite a few various other forums) about various providers.   These providers consistently come up as recommended and, in the case of ServInt, WiredTree, KnownHosts, JaguarPC and LiquidWeb, actively participate in the discussions.  It was also particularly instructive to see how hosting providers worked to resolve client problems.   I also looked at how long a company had been around and what their finances might be like, and did some network tests to check connectivity.   All this research was done over a span of about two months, with the final decision to be done before Xmas 2009. There are plenty of cheaper providers (these are about $50/mo for their cheapest offerings), but you get what you pay for and it’s hard to make decent margins at the cheapest prices.

Finally, I also considered EC2 using this image but it’s unmanaged, so it was a non-starter.

I would note that if you are a business with any sort of traffic, while a VPS is fine (and should be considered a minimum), you really should have your own dedicated system, particularly if it contains sensitive client information.   My recommendations are either – which I have used for 3 companies over the last 10 years (and I never once has a failure); and SoftLayer – which I have used for the last 3 years for my personal system. The absolute cheapest dedicated hosting I have seen is VolumeDrive, but I have never used them, so can’t really speak about it.

In the end, I would up on a VPS from ServInt (which is where you are reading this from…).  They had the best ROI and features combination.   The very close runner up was WiredTree.   Really, you can’t go wrong with either, and I choose ServInt only because they offered Plesk’s Power Panel as a management UI, which, although I dislike most of these GUI tools, I like better than C-panel.

I’m almost done migrating sites from SoftLayer to ServInt and it’s been pretty easy.   After that, I’ll have to deal with email migration (ugh).



Hosted Email – Update #1

Well, after 24 hours of trying out, I’m left frustrated by their jumble of different systems with no attempts to aggregate them into a single ‘portal’, particularly the diversity of different admin systems.   OTOH, the mail service is quite fast, although I have not tested inbound mail much.    They called me this morning to see how things were going (I bitched about their lack of a unified UI), which was a nice touch.    I also asked about BES pricing, and it seems that this might turn out more expensive than I had hoped and with quite a bit of risk in migrating from BIS to BES at the carrier level.  It’s too complicated to explain, suffice it to say that I have a legacy cell plan that is generous with data and any changes might make that go away…    There is also the risk of paying lots of $$$ for BES, which I might not actually use for very long if an Android phone comes out with a keyboard, CDMA and GSM….

Sooo, I’m left re-considering my options.   Suddenly, Google’s offerings seem quite attractive, esp. since they have a Blackberry OTA sync client that’s free.  Hmmm.   I might also have to have a second look at The Message Center as they offer a BES-independent BB sync of some sort.

What a pain.  Really and truly.   I need to make decision soon so I can migrate and stabilize all this before I go on extended travels…



Things that live on forever….

There have been quite a few memes in the past decade about the stupid things you do online living forever. Well, it’s not just the stupid things, you see. Just for kicks, I went back and looked at the website of the very first organization I built a complete, commercial web-presence for, back in 1995.

The organization in question was the US Travel and Tourism Administration (USTTA), a division of the US Dept. of Commerce that keeps track of statistics about tourism and generally tries to support the industry (which, BTW is either the first or second largest US export, depending on how it’s counted). I was working for a government contractor, Ellsworth Associates (which has since been sold several times over) and was the head webmaster, a huge deal in 1995.

My very first job was to build a web presence for USTTA, which we would up calling TInet (in the fashion of the day…). So I set about designing a look and feel that would work well.

Fast forward to 2010, almost 15 years, and, guess what, TInet looks almost exactly the same…. Check it out:

Fantastic, how retro.  I guess my 15 year old design was hugely advanced for it’s time.   Wonder if they are still using the e-commerce (shopping cart back then) code I wrote in Perl all those years ago…..  Hopefully it still runs on Linux.

P.S. Apparently there is no more shopping cart and it’s now running on IIS.  So much for progress.



Hosted email – The task of finding a provider…

Well, for the last month or so, I’ve been investigating several migrations to more managed hosted services so I don’t need to spend so much time doing admin.  Email is rather difficult as I host several power users who have pretty exacting requirements and I wanted to get more capabilities.   A lot of people suggested Google for Domains, and I did consider that pretty seriously.  For $50/user/year, it’s pretty decent, but I decided to try hosted Zimbra instead.

I looked at the following hosting companies:

  • – website is garish, but it offers everything I wanted, although sales is NOT responsive
  • – ditto, but sales was very responsive
  • – lots of good recomendations, sales very responsive, no upfront pricing
  • – I like supporting local SMBs, but they never got back to me about pricing
  • – Has an interesting suite integration with Sugar, but it’s overkill for me and the minimums are expensive
  • – good price vs features, pushed all the right buttons, but this seems an Exchange centric shop trying to branch out, which is a risk for me

In the end, partially because of my wanting BES and ActiveSync plus a fair amount of storage and a large number of domain aliases, 01 and Mindcentric where the only two real contenders.  Others offered a similar combo, but the ROI was not so good when adding up all the extras, like domain aliases or extra storage. I’m currently trying out a free account for one month at, we’ll see how it goes.  At $50/year, it’s competitive with Google Apps, although with a lot less storage.

I really think that Zimbra should have a more detailed partner offering for hosting.  Certainly their rankings help, but with so many providers to choose from, it would be good to have a community filter.   Of course, that’s not really in their best interest.  And, BTW, what happened to Yahoo offering hosted Zimbra?  Seems like a natural fit….

I did briefly consider hosted Exchange, but then I remembered how hard it usually is for me to wrap my brain around Microsoft’s terminology and implementation logic.  It sounded like a perfect recipe for chaos.

[note: I got a nice email from SpecialAI about my post and clarified some things as a result.  It's good to see that at least one company is paying attention to what people are saying about it....]



TEDxSV – Not inspiring

I’ve heard a lot about TED over the last 10 or so years.  It’s always an event I’ve wanted to attend but have never been able to justify the cost.  Lots of brilliant people I know rave about it, so I was happy to see the TEDx series of local events emerge.   I missed TEDxSF San Francisco, as, like many other people, I only heard about it after it happened, but I was lucky enough to go to TEDxSV.

Or so I thought.

I couldn’t make the whole thing as it was an all day event and I had overlapping obligations.  Instead, I did my best to catch bits and pieces of the stream and show up for the afternoon portion.   Disappointing is an understatement.  I’ve watched many great TED talks online, and only one of the TEDxSV talks came close to any of those (Sekou Andrews – watch it here starts at about 22min).   Most of the talks seemed to be talking down to the audience, as in ‘you are lay people, let me preach to you’.   Not inspiring, not like Itay Talgam describing how performances are about telling many stories, including the story of the audience (see Itay Talgam’s presentation at TED here).   Most of the TEDxSV talks had only one story and it was about the presenter.

To bad, it could have been great, but perhaps this is what happens when you try to replicate something in multiple places.



Thunderbird 3 on Mac OSX – Not so great

I’ve been a longtime Thunderbird user, and I thought I would upgrade to V3 today.  Bad idea.   Although the UI was slightly updated (notably through the use of tabs), everything else was either the same or broken.   The mail compose toolbar wound up overwriting the From field section (with From, Subject, etc all in lower case) and nothing seemed quite right.  It was faster, but I decided that reverting back to the status quo ante was the fastest route to productivity.

It seems, however, the Mozilla makes it as hard as possible to download previous versions.   I never actually found the download for Thunderbird 2 and instead went looking for it in my ‘Downloads’ folder.   I keep it for occasions such as this when people decide that old versions are not good for you….

Finally, since Lightning did not work with Thunderbird 3, I decided to use Sunbird (a standalone calendering app) to extract the Lighting data.   Since I was doing this after installing TB3, I went hunting for the file containing my calendar data.   Not only was this the most obscure file in the most unobvious place, but Sunbird was incapable of importing.   Compounded with my basic problems with the buggy TB3, reverting back to TB2 was the thing to do.  After I’d found my download backup, it was relatively easy to do, although the whole adventure cost me the better part of 2 hours.

On another note, I have huge amounts of stored email.   It’s always been an issue, and it’s one of the reasons I use TB over Apple’s just can’t handle large mailboxes (e.g. 3gigs 70k+ messages).    Searching across all this is somewhat of a problem, however, and my solution was to run a local search engine on exported mbox files.   However, I’ve found a better solution, MailStewart, a Mac-only application that will archive all your mail and attachments, index them and make the whole think searchable.  It’s cheap and works great.



The problem with biometrics

I’ve always said that the security of biometrics is really problematic.  Once someone figures out how to hack a biometric device, it’s impossible for users to ‘change the password’.  Unlike most systems that provide authentication (e.g. passwords, SecurID tokens, drivers licenses, etc), you can’t just throw out the ‘token’ (e.g. your fingerprints or eyes) and replace it with a new one.

There have been numerous attempts to hack fingerprints in particular, such as gelatin overlays, cutting off the actual persons fingers, etc., but it seems that the ultimate hack has now come about:


A Chinese woman managed to enter Japan illegally by having plastic surgery to alter her fingerprints, thus fooling immigration controls, police claim.

Apparently, this is quite a widespread thing and it should really worry people.   Not because of those bypassing the system, but because it’s only a matter of time before fingerprints are ‘cloned’ and innocent people are improperly accused/denied/arrested.  It’s compounded by the fact that laws and law enforcement sees fingerprints as one of the stronger ‘proofs’ of someone’s presence in a particular place.

How long before fake irises or even DNA?  Who knows, but it turns out that fingerprints are no more secure than any other form of identification.