(Our suggested “moonshot” for MZ: ending urban traffic congestion via a real-time pricing platform + “connected car”)
(Our suggested new tagline for MZ: “put a price on it.” Shoutout to Portlandia for its “put a bird on it” tagline for a hand-crafted gift store capturing its big picture strategy)
In a year when valuations of so-called Unicorns — startups valued at $1+ Billion — are being marked down by investors, we will present the case that Machine Zone, recently rebranded as MZ, is a $10 Billion Unicorn in the making.
This is audacious claim. A January 2016 Unicorn list compiled by Fortune Magazine assigned a $3 Billion valuation to Machine Zone based on a WSJ report in June 2014 of a funding round of $250 Million led by JPMorgan Chase. There was second hand confirmation of this in Pitchbook.
Machine Zone was not even listed on any Unicorn list a year ago simply because such lists required that valuations be based on reported equity financing with implied valuations of $1+ Billion. Machine Zone’s last reported funding round listed in Crunchbase was a Series B done a full four years ago when Machine Zone was just beginning.
We found Machine Zone’s absence from 2014 Unicorn lists both perverse and ironic. It was perverse in that Machine Zone didn’t need financing so it was excluded from successful startup lists. It was ironic because, unlike most other Unicorns, Machine Zone’s revenue levels and revenue trends are observable daily via app store data reported by analytics companies such as App Annie or Thinkgaming.
On July 15, 2015, Bloomberg reported that the company was in discussions with investors for an additional $200 Million in funding at an implied valuation of $6 Billion. Dean Takahashi of VentureBeat also reported rumours of this new funding round. But, he reported that Machine Zone was seeking $500 Million at an unstated valuation — not the Bloomberg figures. Takahashi’s source also said that “the pitch has met with skepticism.”
Machine Zone has refused to comment on any venture capital interest or funding. As we blogged at the time, compared to most startups who would relish disclosing funding rounds that would confer Unicorn status, Machine Zone “walks the walk, not talks the talk”.
The Evolution of Machine Zone’s Identity
Until this year, Machine Zone’s CEO Gabe Leydon averaged about two interviews a year and never talked about revenue, valuation or IPO plans. He never talked about the state of the mobile game industry. In fact, he rarely talked about Machine Zone’s two hit successes Game of War: Fire Age or Mobile Strike.
Instead, he used rare interviews to advance the theme that Machine Zone was a technology company with software platforms whose applicability and marketability extended beyond games.
In a 2013 interview, Leydon said that Machine Zone had developed a “game engine” that could be “re-skinned” to create other genres of games with the same underlying play and communications innovations. This comment was designed to counter the perception that Machine Zone was a one-hit wonder deserving less of a valuation than mobile game rival Kabam with multiple Top 10 hits at the time.
In 2016, Machine Zone has done just what Leydon predicted in 2013. It had “re-skinned” their top revenue rank Game of War: Fire Age to release another Top 5 revenue rank game Mobile Strike, published by their downtown Palo Alto studio Epic War LLC. What is remarkable to us is that there does not seem to be much cannibalization going on between the two games.
In 2014, Leydon talked about Machine Zone’s real time, crowd-sourced chat translation engine. We wrote several papers speculating that this chat translator would be a valuable addition to Slack as it would open doors to large multi-national corporations.
In a March 2015 interview with Bloomberg’s Robert Kolker, Leydon identified what he thought was the “Wow” factor of its hit game Game of War: Fire Age — the low latency of the game play.
“…Game of War accommodates about 3 million users in simultaneous play, with what the company clocked as a 0.2-second response time…. This is the largest real-time concurrent interactive application ever built. There’s nothing even close to it.”
He also hinted at the marketability of this technology outside of gaming.
Shortly after than interview, we wrote a paper speculating that Machine Zone’s game engine must be a NewSQL database. This was based on job requirements posted on its website and a Michael Stonebraker YouTube video in which Stonebraker said that database requirements for today’s massively mobile multi-user online (MMO) games are the same as modern, cloud-based online transaction processing (OLTP) databases required by banks, airline reservations, order entry systems, real-time ad auctions.
It turns out that Machine Zone’s MMO game “purpose” shaved years of the development of a modern OLTP database compared to development path taken by Google and Facebook. The original “purpose built” databases of the likes of Facebook, Google, and Yahoo were designed to be massively scalable and globally distributed. They did not have to handle transactions.
Database design involves tradeoffs. As the online world’s need for monetization increased, especially real-time ad auction exchanges, a reversal in trade-offs has occurred. In 2012, Google made the often quoted declaration that if it had to choose between a NoSQL and a “NewSQL” database to handle OLTP, it would choose the latter:
“We believe it is better to have application programmers deal with performance problems due to overuse of transactions as bottlenecks arise, rather than always coding around the lack of transactions”
So, Google has morphed its “NoSQL” BigTable into Spanner, a “NewSQL”, which it now uses for its mission-critical sell-side ad platform.
An additional signal of Machine Zone’’s intent on being a fundamental technology company was a report in Las Vegas Sun that the data center builder Switch would be expanding its Southern Nevada facility to house an additional 4,000 servers paid for and managed by Machine Zone . This announcement came in the very month in which Zynga announced that it would begin selling off its own dedicated data centers and return to Amazon AWS as a cost savings move.
On February 18, 2016, Machine Zone and CEO Leydon had a “coming out party”. He broke out of his pattern of infrequent print interviews to give a full blown 39 minute video interview at the important Code/Media 2016 Conference.
In our opinion, it was here that Leydon first demonstrated his charm and ease at speaking as he mixed in fond memories of 90s skateboard videos with big picture views of the state of ad-tech. The interview was convincing evidence to us that Leydon was capable of leading an IPO and being the spokesperson for a publicly-held company.
Within the first minute of the interview, Leydon articulated a more focused view of Machine Zone as “real time” technology company. However, because the audience were media and ad-tech people, Leydon did not mention its database technology at all.
Instead Leydon startled the crowd with sharp criticism of 3rd party buy-side ad-tech platforms and the state of ad-tech in general. He casually revealed that Machine Zone had developed it own ad-buy platform specifically tailored to the acquisition and retention of freemium game players aka “whale targeting and retention”.
This platform was an alternative to relying on outside platforms like Chartboost and Tapjoy, used by Machine Zone’s rival Supercell and other top mobile game companies. According to Crunchbase, Tapjoy has received a total of $2.47 Billion in VC funding over the years.
Once again, by building its own buy-side ad-tech platform, Machine Zone has set itself apart from other mobile app Unicorns. Supercell, its chief rival in the mobile game industry, uses Amazon AWS for infrastructure and Tapjoy for ad-tech. Supercell does not have any internal chat function for players to communicate.
On April 4, 2016, Machine Zone issued a press release stating that it had changed its name from Machine Zone to MZ to underscore its new identity as a “real time” technology company. It also announced that it would begin licensing its real time transactional database, branded as RTplatform™. Leyton suggested in a follow-up conversation with Venturebeat that RTplatform ™ had wide-spread applicability ”from financial service companies to connected car companies to government institutions”
Valuing MZ’s Mobile Game Business
What follows is an estimate of MZ’s current valuation based solely on its mobile game business. There are three pieces of data required: (1) App Annie revenue ranks for MZ’s games; (2) an estimate of a power function relation between annualized revenue run rate (ARR) and app store revenue rank; and (3) “market-derived” valuations of pure play mobile game companies as a multiple their ARR.
For example, Activision Blizzard recently bought King Digital for 3.08 times ARR. Using that as a comparable and an estimate of MZ’s mobile game ARR of $2.0 Billion, we would arrive at a valuation for MZ of 3.08 * $2.0 Billion = $6 Billion.
We have used this methodology to value MZ over the past 2 years: Machine Zone: The $4 Billion Unicorn that Walks the Walk ; Machine Zone and the Perversity of Unicorn Lists and Machine Zone: IPO or What? (for Seeking Alpha).
We used the same methodology in articles to value other publicly-held mobile game companies — King Digital, Zynga and GLU Mobile — and the start-up Kabam. Finally, we have used the methodology to make prescient buy recommendations for two undervalued Japanese mobile game companies — Mixi and KLAB.
Below are two “market-derived” valuations of pure play mobile game companies as a multiple of ARR. The first is a valuation of 3.08 * ARR that Activision Blizzard paid to acquire publicly-held King Digital in late 2015. Using King’s ARR, as reported in 10-Qs to the SEC as a checksum, we present below an estimate of the distribution of King’s ARR by individual game revenue and associated revenue rank as reported by App Appie.
The second is a market-derived valuation for Finland-based Supercell. While the company is not listed on a stock exchange, it is required by Finnish law to report financials once a year. In 2015, Supercell reported revenue of $2.326 Billion. We coupled that with a reported $5.5 Billion valuation that Softbank placed on Supercell when it bought an additional 22 percent stake in Supercell (bringing its ownership to 73 percent) in mid-2015.
As with the King valuation, we use Supercell’s reported 2015 revenue as a checksum when estimating the distribution of Supercell’s ARR by individual game revenue and related revenue rank.
For our valuation of MZ here, we chose the lower, more conservative, Supercell valuation of 2.36 * ARR. The higher 3.08 * ARR that Activision-Blizzard paid for King Digital was 20% higher that the market value of King at the time. Plus, most financial pundits felt that Activision-Blizzard paid too much for King.
In past valuation of MZ, we chose 2.5 * ARR based on market-derived valuations of publicly-held Japanese gaming companies. Given, the general downward drift in Unicorn valuations, the use of the lowest multiple of 2.36 * ARR seems appropriate today.
Based on the estimates above of individual game revenue associated with various iOS Apple USA revenue rank as reported by App Annie, we derive an estimate below of a 2016 power function of global ARR vs iOS USA revenue rank.
We now present a current valuation of MZ based on its two hit games alone which rank #2 and #3 on the App Annie iOS USA revenue charts.
Note: during the writing of this paper in April 2016, Supercell’s Clash Royale and MZ’s Game of War have traded #1 and #2 positions multiple times. We are being conservative in our valuation here by using the lower #2 ranking for Game of War. Had we chosen #1 for Game of War with an associated ARR of $2.1 Billion, our valuation for MZ’s game business would have come in at $7.3 Billion instead of $5.7 Billion
Using “top-line” metrics like sales or monthly active users to value Unicorns has become suspect today. Observers of the startup scene have come to the realization of the futility of growing the top line if unit margins are negative and not likely to turn positive with scale. A freemium mobile game company has zero value if the advertising costs of acquiring a new user are greater than a user’s long term value (LTV), as measured by the discounted present value of money spent. Valuations based on solid top-line data have a greater validity if they are supplemented with some rough estimates of what a Unicorn’s full P&L looks like.
So, to add weight to our $5.7 Billion dollar valuation, we present below a rough estimate of MZ’s full operating P&L. First, our estimates shows that MZ has been responsible in growing headcount consistent with revenue.
Our estimates for MZ’s contribution margin (sales — advertising cost) is a healthy positive number. It is likely that MZ has THE highest contribution margin in the mobile game industry given an estimated average annual in-app game spend of $550 per MZ game player.
It is likely that MZ currently is showing a small operating loss as measured by GAAP, but it would be positive if non-cash, stock-based compensation were backed out. The company is likely cash flow positive from operations. Because MZ as a mobile game company has no inventory or material accounts receivable, it does not need cash for working capital.
As CEO Leydon has observed, mobile games are the most efficient cash conversion operation in the history of modern business. At the Code/Media 2016 Conference, he observed that there can be a 120 second turn-around from cash out for an “call to download” ad to a new user download of a game to the first payment for in-app boosts posted to MZs cash account at the app stores.
MZ does not need cash for working capital or to cover operating losses. It has been reported that their new data center in Nevada is costing them $50 Million, and we could see them needing $100 Million per year for the next 5 years to expand data centers globally.
Use Cases for RTplatform™
We place the MZ’s valuation today at $9.1 Billion as a fundamental technology company. We think the valuation for its ad-tech platform is fair at $1.0 Billion, give that VC’s have poured over $2.7 Billion so far into Tapjoy, a comparable platform. We think the licensing value of its chat translator is fair at $400 Million, given the doors it might open for Slack. There might even be interest in the chat translator from Facebook or Microsoft, given the current interest in text messaging as a replacement for apps and mobile OS.
Admittedly, our $2.0 Billion valuation for RTplatform™ is the most speculative component as no comparable market-derived valuations are offered. One factor that caused us to value it so highly was the very fact that MZ hyped it. Here was a Unicorn company and CEO who had “walked the walk” for years and never made comparisons. Suddenly, it started “talking the talk.” as in “our specs crush your specs” and “ our new specs crush our old specs.” We believe the company can make good on the hype, given their amazing string of accomplishments.
According to the Venturebeat interview on the day of the launch, the company said its platform was “much more scalable than what is currently available in the market from rivals like Amazon or Google…” Leydon said PTplatform™ was “100 times bigger” than its current platform running Game of War.
The other factor underlying our high valuation was the use cases and market potential we were envisioning for a platform described by the company as a
- “massive platform for doing high-fanout data processing,”
- “many-to-many applications”
- “an infrastructure that allows you to do some extremely large things in real time at scale.”
- “unique ability to interconnect ‘billions’ of endpoints worldwide and transmit data at low latency”
In the Venturebeat interview, the company hinted at use cases “from financial service companies to connected car companies to government institutions”. In earlier interviews, Leydon hinted that its game engine was transactional with ultra low latency, which we speculated as meaning NewSQL. He compared it to platforms required for high frequency trading.
We present the following broad use cases for a real-time pricing or auction platform coupled with the “connected self” or the “connected car”:
- eliminate information asymmetry and “moral hazard” between insurers and customers;
- eliminate the “tragedy of the commons” like urban traffic congestion or overfishing;
- eliminate transactions costs causing “sticky prices” for services whose performance over time is uncertain;
A specific use case for (1) would be real time auto insurance pricing. In 2014, consumer auto insurance had been estimated to be a $190 Billion market. MZ should be targeting one of the top 4 auto insurers — State Farm, Geico, Allstate, and Progressive — as an exclusive licensee. They should aim for an announcement within the next three months, with a roll-out and initial monetization within a year.
This “early win” will shock the auto insurance industry, impress the VC investment community, and finally clue tech writers that MZ should listed along with handful of unicorns — Uber, Airbnb, Palantir and Slack — as having the greatest upside potential.
A specific use case for (2) would be a real-time auction for peak commute time on urban freeways. There is an article in Forbes citing a report which estimated the direct and indirect costs of traffic congestion at $124 Billion in 2013.
There was also something called the Millennium Project out of UC-Berkeley in the mid-2000 which used (then novel) mobile phones to gather data on drive times and traffic congestion in the Bay Area. In 2011, there was a report which presented in detail the problems in ”scaling up the Mobile Millennium traffic information system using cloud computing and the Spark cluster computing framework”.
Surely, the 2016 RTplatform™ would be a prime candidate to underpin any solution to urban traffic congestion. Needless to say, solving this problem would require government sponsorship so monetization by MZ for this use case might be a 5+ years off. But, announcing that it would be involved in a project to end urban traffic congestion would place MZ alongside only a handful of companies undertaking a “moonshot” and “make a difference in the world” type of project.
A specific use case for (3) would be dynamic pricing for sporting and entertainment events. Many Major League Baseball team are setting aside bleacher sections with individual game day tickets that vary by day of week, opponent, and weather. The National Football League is also starting to set aside individual game day tickets that vary over the course of the season by attractiveness of the matchup.
We could envision MZ’s platform taking this dynamic pricing of sporting events to a “real time” level by allowing both baseball and football fans to bid on game day seats inning by inning or quarter by quarter. Obviously, this use case seems ludicrous, but it does emphasize widespread instances of “sticky prices” due to transaction costs for a steam of services with uncertain, highly variable quality.