Tuesday, April 15, 2008

Analysts: You Gotta Love Them Edition


The game analysis business is very competitive. They like to make important or controversial announcements to get some ink. As we see in the graphic I stole from Kotaku's "Analyzing the Analysts" post, historical accuracy does not seem to be an impediment to their continued pronouncements. But sometimes, analysts askew wild statements like "EA will postpone GTA IV" or "iPhone is not a viable gaming platform" in favor of an appearance of real live analytics. In this case, UBS's Ben Schachter established a premise, supported by a bunch of numbers to predict GTA IV sales. Unfortunately, he ignored some facts. Why cloud the model with real world facts, when your predictions can be so much more accurate without them?

Schachter says the performance of GTA IV should be measured by tie ratio to the installed base. This is a great addition to the universe of video game analytics. Really. It is the first time I have seen it incorporated into the analysis of game sales. Too often game companies are modeled on either film or technology companies without regard to the quintennial (I think I made that up) fluctuations in the installed base of consoles. Sadly, this is the last supportable fact in Mr. Schachter's analysis before he takes his step through the looking glass.

I pulled some quotes from the Gamedaily piece:

He explains:

"Investors cannot define success of the game in terms of units of GTA IV sold compared with prior GTA titles. This is because the game can only be sold to consumers who own a console capable of playing the game. To that end, the success of the game is defined by the attach rate (or tie ratio) which equals the number of units sold / the installed base of applicable consoles (in this case Xbox 360 and PS3)," he explained.


Mr. Schachter appears to be looking at the installed base of consoles as a static number. According to vgchartz.com, the current size of the relevant installed based is 30.3 million worldwide. As evidenced by previous console cycles, the installed base grows at an ever increasing rate. Features increase, production costs go down and retail prices go down. The unit sales seem to contemplate Christmas software sales, but not Christmas hardware sales. Moreover, GTA IV is one of those elite titles we refer to as "console movers." Granted the incremental number of people who purchase the console to play the game is not earth shattering, but it is significant. There are a large number of people out there sitting on the fence who will buy a console to play the game. Don't believe me, then tell me why Microsoft was willing to pay for exclusive content, and why is Sony releasing a bundled PS3 with GTA IV?

Moving on. . . .

Circling back to the positive early buzz and the importance of reviews, Schachter stated that there is definitely some upside in his six million units estimate. "Should Metacritic scores come in above 95, attach rates could rise to 30%+, implying potential U.S. sales of 7-8mm units (scores for the past three GTA titles were 95, 95, and 97). Any attach rate above 26% would beat expectations in our view," he said.


Here Mr. Schachter has accurately captured the sentiment of the market. So many of the core gamers who played all the GTA's and purchased consoles are waiting to see what the critics have to say. GTA is definitely one of those titles people never hear about and will not buy unless there are positive reviews. I am not being serious here. Worse, I don't see how the attach rate is extrapolated. His number would put the attach rate on a par with Call of Duty 4 which is currently at 28.8%, but should be a bit higher as the bulk of its sales happened into a slightly smaller base. While this is the top selling title on the consoles to date, GTA has traditionally more than doubled sales of COD and number 2, Halo, in previous releases. But we can't really use prior attach rates because they happened a different point in the cycle. He doesn't mention it, but attach rates for core games are higher on the front half of the cycle than the back. In fact GTA's performance on PS2 has very impressive tie ratios, hovering around 10%. This is a really long way of saying attach rates are slippery. Numbers go up, numbers go down, in the end, they can support anything.

Finally, we get to revenue:

"At 6mm units in the U.S., the game would generate ~$360 million in total retail sales (assuming a $60 retail price), at 8mm units, the game would generate $480 million of sales in the U.S., or about 33% upside from current expectations," he added. "Additionally, we expect the game could do similar numbers in Europe, almost doubling our U.S. totals."


Once again we ignore all those pesky facts which only introduce additional variables. He also uses retail numbers so he doesn't have to worry about all those silly internal numbers like the bottom line. The retail price is assumed to be $60. We all know what happens when we assume. First, the retail price of the base unit is $60 USD in the US. The retail price of the Collector's Edition is $90. As GTA is one of the few titles which can hold retail for over a year, even if the Collector's run is relatively small, it will increase the average retail price. I know what you are saying, the increase is peanuts. Yes it is, but the retail revenue increase in Europe is not. The UK retail price of GTA IV is 39 pounds and the Euro price on the continent is 65 Euros, or at today's rates, USD 76.14 and USD 102.45, respectively. If sales are roughly the same in Europe, or 40% as stated in the article, shouldn't revenue be much, much higher?

I realize headlines are cool, and new angles on analysis are hard to find, but c'mon guys. Let's sprinkle some reality into the work.





No comments: