AGILEVC My idle thoughts on tech startups

February 26, 2010

Here’s the second installment in my “Cliff Notes” versions of S-1 filings by internet & digital media companies that have recently filed to go public. The first one on Quinstreet is here, they priced their offering at $15.00/sh and started trading publicly on February 12th.
======================================




Founding Date: 2002
Headquarters: Los Angeles, CA

What They Do: Netflix for video games (console – Xbox, PS3, Wii, etc)

How They Do It: Gamefly website used to acquire & manage customer subscriptions, physical distribution centers handle game inventory which is fulfilled by US Mail

How They Make Money: Vast majority of revenue (nearly 90%) from consumers paying a monthly subscription fee ($15.95 – 36.95), small amount of revenue from sale of previously rented games and advertising

Financial Snapshot:
  • 2009 Revenue: $93 million (1)
  • Revenue Growth: 28% YoY (2009 FY), 44% YoY (2008 FY)
  • 2009 Gross Profit: $45 million (1)
  • Gross Margin: 48% –> i.e. for every $1 of revenue of GameFly spends $0.52 on buying game inventory, postage, and other costs of service (customer acquisition spend is not included in COGS)
  • 2009 Net Income: $1.6M (1)
  • 2009 Operating Cash Flow: $4.6M (1)
Subscription Metrics:
  • 334K paid subscribers at 9/30/09
  • 8.2% monthly churn
  • $20.94 monthly avg revenue per user (ARPU) –> this appears to have held remarkably consistent from 2006 to today
  • $25.90 cost per acquired subscriber (CPA) –> appears to have dropped from ~$30 in recent years
  • $169.34 implied lifetime value (LTV) (2) –> i.e. for every $1 of marketing spend GameFly generates ~$6 of revenue or ~$3 of gross profit
Notable Aspects of Their Business:
  • “Back” Catalog Usage: Over 50% of customer rentals are of video games more than 6 months old, which is likely an important driver of GameFly’s profit margins. If most rentals were of newly released titles, their cost of sales would probably be higher for two reasons. First, GameFly’s inventory would be “turned” far fewer times if most rentals were of new games… either stocking fewer games and having more customers churn (cancel their subscription) as they can’t get the game they want, or having to keep a larger inventory of new titles (turned fewer times) but happier customers. Secondly, game publishers typically sell older games for lower prices than brand new ones so GameFly can presumably reduce their inventory acquisition cost (even wholesale).
  • Delivery Beyond Mail Order: It’s not huge surprise, but GameFly recognizes the threat/opportunity to expand into new delivery channels. This includes online gameplay and download which IMO will be difficult for GameFly simply because there are strong incumbents in online casual games (Zynga, EA/Playfish, et al), the challenges posed by online play of “hardcore” games (e.g. OnLive), and the threat of online distribution through console platforms themselves (Xbox Live, etc). But GameFly is piloting kiosk based game rental & sales with two national retailers (unnamed), so if the company began as “Netflix for games” perhaps they’ll soon also be “Redbox for games”.
Pre-IPO Funding History (3):
In total GameFly took approximately $20M in VC funding, which is comparatively modest for a startup with nearly $100M in revenue and poised to go public. Interestingly GameFly’s chief backer, Sequoia Capital, owns a slight majority of the company (51.6%) prior to IPO. While it’s not unusual for a VC syndicate to collectively own a majority of a startup by the time it reaches exit, it’s fairly rare that a single firm has majority ownership. Sequoia was initially a minority investor but participated in every round of funding from Series A thru Series D.

Another interesting wrinkle (well, maybe n
ot that interesting… except to perhaps us fellow VCs) is that Sequoia invested in GameFly out of two separate funds. In and of itself this isn’t that uncommon, but the sequencing of the investments is somewhat unusual in that Sequoia X led the initial investment in GameFly’s Series A and participated in every subsequent round. Sequoia IX, i.e. an older fund, invested first in GameFly’s Series C and then participated in the Series D (final VC round prior to IPO). There’s nothing inherently “wrong” with this and certainly a firm like Sequoia with a long and successful track record has broad latitude with LPs to pursue its investment strategy. But to the extent a VC firm is investing in a company out of two different funds, it’s more common that the two funds initially invest at the same round or perhaps the older fund invests before the newer one.

Series A Preferred
  • $2.0M round, closed 2004
  • Major investor was Sequoia Capital [$1.75M]
  • Pre-money valuation probably no more than $4.7M (4)
  • 8% dividend, 1x non-participating liquidation preference (5)
Series B Preferred
  • $4.6M round, close date unclear (sometime prior to 2008)
  • Major investors were Sequoia Capital [$4.0M] and Peter Thiel [$0.25M]
  • Pre-money valuation probably no more than $8.3M (4)
  • 8% dividend, 1x non-participating liquidation preference (5)
Series C Preferred
  • $6.2M round, close date unclear (sometime prior to 2008)
  • Major investors were Sequoia Capital [$5.5M] and Peter Thiel [$0.25M]
  • Pre-money valuation probably no more than $32.0M (4)
  • 8% dividend, 1x non-participating liquidation preference (5)
Series D Preferred
  • $8.5M round, close date unclear (sometime prior to 2008)
  • Major investors were Sequoia Capital [$4.1M] and Tenaya Capital [$4.0M] (6)
  • Pre-money valuation probably no more than $46.5M (4)
  • 8% dividend, 1x participating liquidation preference w/ 3x cap (5)

Notes:
(1) Actually for 12mo ended 9/30/09, GameFly reports on a fiscal year ending Mar 31
(2) You can calculate an implied LTV for any subscription business where you know both ARPU and churn rate. Email me if you want to understand how to do this, but basically it’s a decay formula to calculate the # of periods an average customer remains a subscriber and then multiplying that by the ARPU.
(3) Gleaned from S-1 itself and the exhibits (investor rights agrmt, certificate of incorporation, etc)
(4) S-1′s typically provide info about number of shares, share price, and sometimes date of issuance for each round of preferred stock financing but not how many common shares were outstanding at the time. By 2008 there were approximately 4.7M common shares outstanding and 5.4M by 2009. It’s conceivable there were fewer common shares at the times of the Ser A-D rounds though there were likely unissued shares for the option pool, so the fully-diluted share count at the time of the round could have been higher (implying a higher post-money valuation).
(5) The preferred stock classes will all be converted to common at the IPO, so liquidation preference and other rights will go away
(6) Tenaya Capital is the former venture capital arm of Lehman Brothers, which now operates as an independent firm

Disclaimer:
This post should in no way be construed as a recommendation to purchase Gamefly stock or any other security, these are simply my own personal observations.