Brad Feld

Month: March 2005

Paul Graham – founder of Viaweb (bought by Yahoo! in the late 1990s) and author of Hackers & Painters – has two new excellent essays up.  While I don’t know Paul, I do know of him, and I think his writing is great.

  • How To Start A Startup.  Paul thinks you need three things: (1) good people, (2) make stuff people want, and (3) spend as little money as possible.  All the right stuff with fun examples, lots of Paul’s trademark bluntness, and plenty of provocative questions.
  • A Unified Theory of VC Suckage.  I won’t even try to defend my VC brethern since Paul’s theory is sound in many ways.  He admits that he’s met a few VC’s that he likes, so there must be something messed up in the universe somewhere.

He’s finally put an RSS feed up on his site so you can have his thoughts delivered directly to your computer.


Software revenue recognition can be incredibly complex in today’s world of SaaS, perpetual vs. subscription licensing, SEC SAB 101 (and others), new FASB pronouncements, and the legacy of bad (or fraudulent) revenue recognition policies at companies like Computer Associates.

I believe the right approach for an early stage software company is to come up with a revenue recognition policy, vet it with your outside auditors, document it, and then stick to it no matter what.  The only changes should come from direction from the outside auditors at the end of the year during the audit (or – quarterly during the audit if you are a public company).

So – it’s a great pleasure when I receive a copy of an email from a CEO to his entire management team who is obsessed with doing things right (in this case I’m the lead board member on the audit committee).

Folks – let me be crystal clear on this. We have our approved revenue recognition policy and we aren’t to deviate from it without proper authorization. Please do not ask X (CFO) to do anything differently than what is in this document. Additionally, if you see anything being done incorrectly please escalate to myself or Brad (or if anybody perceives me pushing for something incorrect, please alert Brad immediately).  Thank you.

Unambiguous, clear, unwavering, and inclusive of the board / audit committee.  As Q105 comes to a close, make sure you’ve got your revenue recognition policy in place, everyone understands it, and it’s unambiguous how to behave.


There’s nothing like a solid week of vacation with no phone, email, or blogs to get the writing juices rolling again. Of course, now that I’m through my email, I only have 8200 blog posts to read to catch up – thank god for jet lag – wait, what am I saying?

In our term sheet series, Jason Mendelson and I have been focusing first on “the terms that really matter.” We are down to the last one – the pay-to-play provision. At the turn of the century, a pay-to-play provision was rarely seen. After the bubble burst in 2001, it became ubiquitous. Interesting, this is a term that most companies and their investors can agree on if they approach it from the right perspective.

In a pay-to-play provision, an investor must keep “paying” (participating pro ratably in future financings) in order to keep “playing”(not have his preferred stock converted to common stock) in the company. Sample language follows:

Pay-to-Play: In the event of a Qualified Financing (as defined below), shares of Series A Preferred held by any Investor which is offered the right to participate but does not participate fully in such financing by purchasing at least its pro rata portion as calculated above under “Right of First Refusal” below will be converted into Common Stock.


[(Version 2, which is not quite as aggressive): If any holder of Series A Preferred Stock fails to participate in the next Qualified Financing, (as defined below), on a pro rata basis (according to its total equity ownership immediately before such financing) of their Series A Preferred investment, then such holder will have the Series A Preferred Stock it owns converted into Common Stock of the Company. If such holder participates in the next Qualified Financing but not to the full extent of its pro rata share, then only a percentage of its Series A Preferred Stock will be converted into Common Stock (under the same terms as in the preceding sentence), with such percentage being equal to the percent of its pro rata contribution that it failed to contribute.]


A Qualified Financing is the next round of financing after the Series A financing by the Company that is approved by the Board of Directors who determine in good faith that such portion must be purchased pro rata among the stockholders of the Company subject to this provision. Such determination will be made regardless of whether the price is higher or lower than any series of Preferred Stock.


When determining the number of shares held by an Investor or whether this “Pay-to-Play” provision has been satisfied, all shares held by or purchased in the Qualified Financing by affiliated investment funds shall be aggregated. An Investor shall be entitled to assign its rights to participate in this financing and future financings to its affiliated funds and to investors in the Investor and/or its affiliated funds, including funds which are not current stockholders of the Company.”

We believe this is good for the company and its investors as it causes the investors “stand up” and agree to support the company during its lifecycle at the time of the investment. If they do not, the stock they have is converted from preferred to common and they lose the rights associated with the preferred stock. When our co-investors push back on this term, we ask: “Why? Are you not going to fund the company in the future if other investors agree to?” Remember, this is not a lifetime guarantee of investment, rather if other prior investors decide to invest in future rounds in the company, there will be a strong incentive for all of the prior investors to invest or subject themselves to total or partial conversion of their holdings to common stock. A pay-to-play term insures that all the investors agree in advance to the “rules of engagement” concerning participating in future financings.

The pay-to-play provision impacts the economics of the deal by reducing liquidation preferences for the non-participating investors. It also impacts the control of the deal, as it reshuffles the future preferred shareholder base by insuring only the committed investors continue to have preferred stock (and the corresponding rights).

When companies are doing well, the pay-to-play provision is often waived, as a new investor wants to take a large part of the new round. This is a good problem for a company to have, as it typically means there is an up-round financing, existing investors can help drive company-friendly terms in the new round, and the investor syndicate increases in strength by virtue of new capital (and – presumably – another helpful co-investor) in the deal.


I’ve been a Spencer F. Katt fan for as long as I can remember.  I even have a classic Spencer Katt t-shirt and a coffee mug somewhere for rumors I sent in to PC Week (back when eWeek was PC Week and rumors were about Novell – well before Al Gore invented the Internet).  Spencer just started a blog – hopefully he’ll start splicing in his Kattoon’s.


It has been a while since I put up a term sheet post so I thought I’d tackle a hard one today. While it’s fun to tease lawyers about math (and – actually – about anything), my co-author on this series Jason Mendelson (a lawyer) often reminds me that lawyers can do basic arithmetic (and occasionally have to resort to algebra). The anti-dilution provision demonstrates this point.

Traditionally, the anti-dilution provision is used to protect investors in the event a company issues equity at a lower valuation then in previous financing rounds. There are two varieties: weighted average anti-dilution and ratchet based anti-dilution. Standard language is as follows:

Anti-dilution Provisions: The conversion price of the Series A Preferred will be subject to a [full ratchet / broad-based / narrow-based weighted average] adjustment to reduce dilution in the event that the Company issues additional equity securities (other than shares (i) reserved as employee shares described under the Company’s option pool,, (ii) shares issued for consideration other than cash pursuant to a merger, consolidation, acquisition, or similar business combination approved by the Board; (iii) shares issued pursuant to any equipment loan or leasing arrangement, real property leasing arrangement or debt financing from a bank or similar financial institution approved by the Board; and (iv) shares with respect to which the holders of a majority of the outstanding Series A Preferred waive their anti-dilution rights) at a purchase price less than the applicable conversion price. In the event of an issuance of stock involving tranches or other multiple closings, the antidilution adjustment shall be calculated as if all stock was issued at the first closing. The conversion price will also be subject to proportional adjustment for stock splits, stock dividends, combinations, recapitalizations and the like.

Full ratchet means that if the company issues shares at a price lower than the Series A, then the Series A price is effectively reduced to the price of the new issuance. One can get creative and do “partial ratchets” (such as “half ratchets” or “two-thirds ratchets”) which are a less harsh, but rarely seen.

While full ratchets came into vogue in the 2001 – 2003 time frame when down-rounds were all the rage, the most common anti-dilution provision is based on the weighted average concept, which takes into account the magnitude of the lower-priced issuance, not just the actual valuation. In a “full ratchet world” if the company sold one share of its stock to someone for a price lower than the Series A, all of the Series A stock would be repriced to the issuance price. In a “weighted average world,” the number of shares issued at the reduced price are considered in the repricing of the Series A. Mathematically (and this is where the lawyers get to show off their math skills – although you’ll notice there are no exponents or summation signs anywhere) it works like this (note that despite the fact one is buying preferred stock, the calculations are always done in as-if-converted to common stock basis):

NCP = OCP * ((CSO + CSP) / (CSO + CSAP))

Where:

  • NCP = new conversion price
  • OCP = old conversion price
  • CSO = common stock outstanding
  • CSP = common stock purchasable with consideration received by company (i.e. “what the buyer should have bought if it hadn’t been a ‘down round’ issuance”)
  • CSAP = common stock actually purchased in subsequent issuance (i.e., “what the buyer actually bought”)

Recognize that we are determining a “new conversion price” for the Series A Preferred . We are not actually issuing more shares (you can do it this way, but it’s a silly and unnecessarily complicated approach that merely increases the amount the lawyers can bill the company for the financing). Consequently, “anti-dilution provisions” generate a “conversion price adjustment” and the phrases are often used interchangeably.

Got it? I find it’s best to leave the math to the lawyers.

You might note the term “broad-based” in describing weighted average anti-dilution. What makes the provision a broad-based versus narrow-based is the definition of “common stock outstanding” (CSO). A broad-based weighted average provision includes both the company’s common stock outstanding (including all common stock issuable upon conversion of its preferred stock) as well as the number of shares of common stock which could be obtained by converting all other options, rights, and securities (including employee options). A narrow-based provision will not include these other convertible securities and limit the calculation to only currently outstanding securities. The number of shares and how you count them matter – make sure you are agreeing on the same definition (you’ll often find different lawyers arguing over what to include or not include in the definitions – again – this is another common legal fee inflation technique).

In our example language, we’ve included a section which is generally referred to as “anti-dilution carve outs” (the section (other than shares (i) … (iv)). These are the standard exceptions for share granted at lower prices for which anti-dilution does not kick in. Obviously – from a company (and entrepreneur) perspective – more exceptions are better – and most investors will accept these carve-outs without much argument.

One particular item to note is the last carve out: (iv) shares with respect to which the holders of a majority of the outstanding Series A Preferred waive their anti-dilution rights. This is a carve out that started appearing recently which we have found to be very helpful in deals where a majority of the Series A investors agree to further fund a company in a follow-on financing, but the price will be lower than the original Series A. In this example, several minority investors signaled they were not planning to invest in the new round, as they would have preferred to “sit back” and increase their ownership stake via the anti-dilution provision. Having the larger investors (the majority of the class) “step up” and vote to carve the financing out of the anti-dilution terms was a huge bonus for the company common holders and employees who would have suffered the dilution of additional anti-dilution from investors who were not continuing to participate in financing the company. This approach encourages the minority investors to participate in the round in order to protect themselves from dilution.

Occasionally, anti-dilution will be absent in a Series A term sheet. Investors love precedent (e.g. the new investor says “I want what the last guy got, plus more”). In many cases anti-dilution provisions hurt Series A investors more than prior investors if you assume the Series A price is the low watermark for the company. For instance, if the Series A price is $1.00, the Series B price is $5.00, and the Series C price is $3.00, then the Series B is benefited by an anti-dilution provision at the expense of the Series A. However, our experience is that anti-dilution is usually requested despite this as Series B investors will most likely always ask for it and – since they do – the Series A proactively asks for it anyway.

In addition to economic impacts, anti-dilution provisions can have control impacts. First, the existence of an anti-dilution provision incents the company to issue new rounds of stock at higher valuations because of the ramifications of anti-dilution protection to the common stock holders. In some cases, a company may pass on taking an additional investment at a lower valuation (although practically speaking, this only happens when a company has other alternatives to the financing). Second, a recent phenomenon is to tie anti-dilution calculations to milestones the investors have set for the company resulting in a conversion price adjustment in the case that the company does not meet certain revenue, product development or other business milestones. In this situation, the anti-dilution adjustments occur automatically if the company does not meet in its objectives, unless this is waived by the investor after the fact. This creates a powerful incentive for the company to accomplish its investor-determined goals. We tend to avoid this approach, as blindly hitting pre-determined (at the time of financing) product and sales milestones is not always best for the long-term development of a company, especially if these goals end up creating a diverging set of goals between management and the investors as the business evolves.

Anti-dilution provisions are almost always part of a financing, so understanding the nuances and knowing which aspects to negotiate is an important part of the entrepreneur’s toolkit. We advise you not to get hung up in trying to eliminate anti-dilution provisions – rather focus on (a) minimizing their impact and (b) building value in your company after the financing so they don’t ever come into play.


Five years ago, Amy told me that in order to stay married (to her), I needed to commit to going away for a week every three months for a vacation.  A real vacation.  No computer, no email, no cell phone (and now no blogging / RSS).  Five years later, we fondly refer to our “quarterly complete disconnect” vacations as “Qx vacation” (e.g. this one was Q1).

Amy has dreamed of living in Paris as long as I’ve known her and even though I’d much rather spend a week in Rome, we spent Q1 vacation in Paris helping her get settled in to a six week “intensive Paris experience” where her goal is to really learn how to speak French.  She’s been writing about her first week on her blog – if you have any interest in an American writer’s experience in Paris in 2005, check out her Postcards from Paris section.

I spent the week sleeping, reading, running, sleeping, playing with my magnificent wife (who I’ll miss a lot over the next month), eating, sleeping, and staying as far away from my computer as I could.  I turned it on for the first time today and – shockingly – it hadn’t changed much.

Over Thanksgiving (Q4 vacation – get the picture) – I maintained my typical vacation reading pace of a book a day.  I kept it up again and chewed down seven books this week – all of them worthwhile.  Following is a quick summary of each

I started the week with Long Distance by Bill McKibben.  McKibben is an awesome writer and a hugely entertaining person.  At 37, he decided to pursue a “year of the body” (instead of what he’d been doing previously – which was spending all his time writing and being a self-described wimp) as he needed a break from “failing to save the world.”  He chronicled his year of training competitively as a cross-country skier, transforming himself from a wimp to an accomplished athlete, and coping with the unexpected illness, decline, and death of his father.  If you are an athlete, aspire to be an athlete, or love beautiful, emotional, and intense writing from a master of words, this is for you.

Jack Fish was next.  This was pure mental floss – wacky, silly, fun, colorful – all the typical mental floss adjectives.  Brain candy.

In response to my On Bullshit post, one of the readers of my blog (sorry – I forgot who – I know, bad form) recommended Why Business People Speak Like Idiots.  Terry Gold had a good series of posts on Jarbarish and I don’t think I ever need to hear the words “space”, “traction”, “leverage”, “mindshare”, or “value-added” again in my life (please hit me if I say any of these stupid, meaningless words).  This book was written by some ex-Deloitte Consulting folks.  While it was predictable, it went down fast and had some entertaining moments.  While not nearly as philosophical as On Bullshit (nor as short), it helped identify best practices for more fully leveraging your value-added communication to your constituents.

The best part of The Number was Mark Cuban’s introduction – just because it was so right.  While this was yet another “post bubble – what happened – why it always happens – and why it will happen again” book, it had some useful history, completely trashed (appropriately) the accounting industry, and had some good anecdotes that I hadn’t heard before.  Oh – and it trashed the accounting industry (did I say it trashed the accounting industry – it did – and it was fun to read (and correct)).

Sapphire Sea was more mental floss – I believe it was the author John Robinson’s first novel and – although it needed a much better editor – it was fun.  I’d like to believe I learned something about Madagascar, but I think the half-life of the book has already passed.

Normally I’d fight the urge to read any book with the word “marketing” in the title (ok – I wouldn’t have to fight the urge very hard), but I looked forward to The Marketing Playbook since it was by John Zagula and Richard Tong – both of Ignition Partners (and ex-Microsoft).  This one belongs on every entrepreneur’s bookshelf – John and Rich did a great job of distilling their battle-tested marketing approaches into five different “plays”, explained them well, gave plenty of relevant examples, and only got hung up in marketing blather a couple of times.  The second edition needs a better editor (notice a pattern – it’s remarkable the different between a well edited, not so well edited, and poorly edited book) – 25% of the words could go.  But – still well worth the time. 

Touching The Void kept me up until 3am last night (which is part of the reason I’m up again tonight at 1:30am Paris time).  This is Joe Simpson’s classic book about the disastrous climb he and his climbing partner Simon Yates had in 1985 on Siula Grande in the Peruvian Andes. Joe survived a near death experience to write an amazing book about it (and subsequently live a great life – this man did not waste his second chance).  Wow.  I never need to climb a mountain like this.

So – I’m back on line, refreshed, a little goofed up on time, and very happy to be part of the human race.  It only took a week to get back to something that resembles normal.


Warren Buffett is one of my business heros – I think he is amazing on many different levels.  His Berkshire Hathaway annual shareholder letter is legendary (and a model that everyone should follow) – the 2004 annual report is out and the letter is available online. It’s required reading for anyone running a company – for substance, organization, style, tone, and wit.

If you feel ambitious, you can (and should) read the Berkshire Hathaway shareholder letters going back to 1977.


Allen Morgan at Mayfield has a nice series up on his blog about the ten commandments for entrepreneurs.  His post today is Commandment #6: Explain Your New Ideas by Analogy To, or Contrast With, Old Ideas.

He’s right.  Mostly.  At the end of his post, he asks for ways to “categorize the new ‘It’” (if they are VCs).  My constructive addition to his post is the notion of the analog analog (also known as the “analog analogue”, but I like my version better).

In the mid 1990’s, I met Jerry Colonna, we invested in a few companies together, and became very close friends.  I love Jerry and – while I rarely see him since he’s in NY and I’m in Boulder – I feel connected to him in a way that’s unique.  Maybe it was our joint experiences together, maybe it was something we drank one night, or maybe it was merely a cosmic connection – in any case, I smile whenever I think of the things I’ve learned from him and the experiences we have had together.

One day, when we were talking about a deal, Jerry knocked me on my ass by saying “what’s the analog analog?”  In true Feld fashion, I responded with a “huh?”  Jerry went on to explain his theory of the analog analog (which I’ve written about before) – specifically that every great technology innovation (or technology business) has a real world, non-digital analogy.  It’s not the “nothing new is ever invented” paradigm – rather it’s the “learn from the past” paradigm.

I’ve found this to be a much more powerful lens to look through when evaluating a new business than the “technology analog” lens (which is the one Allen is describing in his post).  While “Tivo for the Web” or “eBay meets CNN” are useful analogies, I recommend entrepreneurs take a giant step back – out of the technology domain (or at least our current technology domain) – and get to the core analogy – optimally a non-digital one.  Then – walk forward from the analog analog through other analogies to the current idea.

Throughout my life, I’ve heard the cliche “history repeats itself” over and over again.  This is never more true then in the computer industry.  Earlier this morning, I wrote about Ryan’s post on Mr. Moore in the Datacenter and alluded to the migration from mainframe to web to ASP to SaaS (aren’t they all different versions of the same thing?).

All hail the analog analog – the more things change, the more they stay the same (ok – that’s a cliche also).


Everyone knows that exponential curves don’t last forever.  However, it’s remarkable to look at how they apply to specific cost curves – especially when they enable a radical transformation of how technology is used and deployed.  My partner Ryan McIntyre – one of the co-founders of Excite – talks extensively about this in his post Mr. Moore in the Datacenter.  It’s a great read with some fascinating real data (yes – I remember paying $1k / megabit / month in 1996 back when SaaS was called ASP – but I was too young for it back when it was called timesharing – although I bumped into Multics a couple of times at MIT).

In the mid-1990s, Dan Bricklin told me that computing had gone full circle – in the early 1980’s it used to be that you typed some stuff on your screen, hit enter, and waited a few seconds for something to happen.  By the mid 1990’s we were doing the same, only using a PC over dial-up lines using HTTP.  Those mainframe dudes had it right, but thanks to Mr. Moore and crew, we no longer have to wait a couple of seconds (except every now and then, or when I’m accessing the web from my Sidekick.)


I’ve never completely understood the whole cell phone industry subsidize handset thing, but I’ll roll with it (ok – I understand it – market share, total value of the customer, service plan lock-in, uplift pricing for additional functionality over time).  I love my T-Mobile Sidekick II (as you’d expect since I’m an investor – but I really do love it even though Amy forbids me from taking it to bed with me).  You can get it for NEGATIVE $50 from Amazon right now.

Paris Hilton loves her Sidekick II also (note to self – don’t use your famous dog’s name as your password).  If you want a demo of how she uses it, click here.