Wednesday, June 15, 2011

No more doubt: Oracle wants billion-dollar amount from Google

[Update] A new filing by Oracle confirms that Oracle's damages claims in this case are in the billions of dollars. [/Update]

Yesterday's announcement of the Apple-Nokia settlement led to much speculation over how much money the Finnish handset maker receives. All analysts agreed that it must be a lot. But there are also other smartphone patent disputes of similar economic magnitude. In particular, Oracle's dispute with Google may be settled or decided within a matter of months.

I said on other occasions that the amount of money Oracle is hoping to be awarded in the patent and copyright infringement case against Google is probably comparable to what SAP had to pay (among the verdict, interest and legal fees, SAP's payments to Oracle amounted to approximately $1.5 billion). Having analyzed Google's latest filings in the case, I have no more doubt that Oracle's damages expert calculated damages that would, at least if tripled due to willful infringement, amount to a billion-dollar figure. And that's just damages: if Oracle wins, Google will also face an injunction and need to negotiate a license deal (in the alternative, all existing Dalvik-based apps would have to be ported to a new app platform).

I previously blogged about Google's criticism of various positions taken by Oracle's damages expert. That filing, even though redacted and referring to a sealed report, was a request for permission to file a so-called Daubert motion. It revealed some interesting and previously unknown facts about Oracle's stance.

Yesterday Google filed the Daubert motion, which amounts to 48 pages including supporting declarations. In addition, there are more than 500 pages of non-sealed supporting documents and an unknown number of pages of sealed exhibits. Even some of the non-sealed material is redacted, and it makes reference to the report written by Oracle's expert, which is not publicly available. That's why it's very difficult to analyze all of what Google writes. Oracle will reply soon, and then Google may file a response to Oracle's answer, and thereafter the judge will decide. As this unfolds, we'll learn more. For now I just wanted to focus on gleaning some useful numbers and other indications of the damages Oracle claims to have suffered.

Ads, lost profits, Java fragmentation

Google claims in one of its new filings that Oracle's expert proposes the following basis for his theory:

[...] adds together as the basis for calculating a royalty rate: (1) Google's incremental profit from advertising on all Android devices world-wide, and (2) the purported harm to Sun's and Oracle's business from lost profits and alleged "fragmentation" of the Java platform.

That actually means that three components are added. Google just uses the numbers 1 (for what relates to Google's business) and 2 (for what relates to Oracle's business), but the second part has two parts: (i) Oracle's lost profits and (ii) Oracle's harm from "fragmentation" of Java.

Google's filings provide clues as to Oracle's position on each of the three pillars: ads, lost profits, fragmentation of Java. I'll discuss them in the following sections.

Ad revenues: "even split" demanded

In one of its new filings, Google says this:

That shows what Oracle is really after--any revenues that could be associated with advertising on Android Devices.

As I discussed in my previous post on this case, Oracle wants to include Google's Android-related advertising revenues while Google wants them excluded. On this question I concur with Oracle.

It's not known what total amount of Android-related advertising revenue Oracle's damages expert calculated. On a per-device basis, this passage contains some interesting estimates:

[...] Google would earn net revenues of $0.58 per device for Symbian phones in 2011 while it would earn approximately $3.35 per Android and $3.07 per Apple iOS device in that year. Thus, while Dr. Cockburn estimates that Google loses approximately $2.77 in net revenue for every Symbian device sold, he estimates that Google makes almost the same net revenue on Apple iOS phones as it does with its own Android phones (a difference of $0.28).

I guess the underlying assumption is that iOS and Android users have similar mobile Internet usage patterns while Symbian users make much less use of the Internet from their phones. Before I started using Android I had various Nokia phones, and some of the more recent ones such as the N95 and N97 were basically smartphones but my data volume indeed went up hugely with Android.

Last month Google announced the activation of 100 million Android devices. That number is growing rapidly, currently at a rate of 400,000 activations per day. Presumably, Oracle's damages expert takes a high number of Android devices in use at a hypothetical point in time and multiplies it with $3.35 to calculate annual revenue.

In my previous post on damages I already quoted a passage in which Google complained about a 50% royalty rate on Google's mobile advertising income. That claim is also stated in one of the new filings:

Cockburn assumes that all of Google’s advertising revenues associated with Android Devices would have been fair game to Sun and that Google would have given Sun "an even split" [meaning 50%] of its incremental profits from advertisements, plus compensated Sun for half of what Sun lost.

We'll get to that "plus compensated" part in the next two sections (on lost profits and fragmentation of Java). I quoted this in context because this shows again the additive approach: on the one hand, Oracle wants to be paid for the benefits of Android to Google's business; on the other hand, Oracle wants to be compensated for the harm it claims to have suffered.

Lost profits: over $200 million

Google denies that Oracle is entitled to recovery of lost profits in this case. According to Google, this would be subject to criteria that Oracle's damages expert hasn't even addressed in his report.

This is the only one of the three pillars for which I've been able to find a reasonably specific amount: over $200 million (and that's prior to a possible tripling for willful infringement, of course). Here's the quote:

Dr. Cockburn attempts to recover lost profits as part of his reasonable royalty analysis, increasing his estimated damages award by over $200 million (see Cockburn Report, Exhibit 26).

The word "increasing" again shows the additive approach. This amount of more than $200 million is just a limited part of Oracle's overall damages claim, which in turn would be tripled should Google be found to have acted willfully in its alleged infringement of Oracle's intellectual property.

Fragmentation of Java: possibly billions of dollars all by itself

I already discussed in my previous post on this subject the concept of "fragmentation of Java" as an important component of Oracle's damages claim.

In its latest filing, Google states:

A substantial portion of Cockburn's damages calculation is based on so-called "fragmentation" of the Java platform into numerous incompatible sub-standards.

The word "substantial" is very important here. This means it's a fairly large part of Oracle's damages claim.

With reference to Sun's Java-related settlements with Microsoft, Google writes the following:

Cockburn focuses on the $900 million settlement of the case in 2004 even though the case involved far more than fragmentation and the agreement was a broad business arrangement that involved more than settlement of litigation (i.e., included covenants not to sue under the parties’ entire patent portfolios). At the same time, Cockburn completely ignores a 2001 settlement in which Microsoft paid Sun just $20 million to settle litigation involving claims related to fragmentation. (Exs. N & O.) In any event, the Sun–Microsoft settlements relate to all of Java, not JavaME. Cockburn gives no reason why the recovery here, related only to JavaME, should be several times that amount.

The court and the parties know the amount that Oracle's expert calculated. I don't. But I can deduce that the expression "several times that amount" is far more likely to mean "several times $900 million" than "several times $20 million". There are three good reasons for that assumption:

  1. According to Google, Oracle's expert "focuses on the $900 million settlement". If he focuses on that huge number, why would he then arrive at a number in the tens of millions of dollars? That would only make sense if he assumed that Java became much less valuable than it was in 2004. That's not the case -- and even if it were, Oracle wouldn't say so.

  2. The sentence that contains "several times that amount" begins with "[i]n any event". That's a concession by Google that even if one accepted (as Google does not) the more negative scenario for Google ($900 million), there would be no justification in Google's view to attach more value to JavaME (Micro Edition, particularly for mobile devices) than to Java as a whole.

  3. Finally, let's consider the word "substantial" again (from the passage I quoted prior to the one I'm discussing now). An amount in the tens of millions of dollars would certainly be "significant" but hardly "substantial" in this context, in which just one of the three pillars -- lost profits -- already amounts to more than $200 million.

So it's a reasonable assumption that Oracle's damages expert believes "fragmentation of Java" has inflicted damages in the billions of dollars. It's another question whether Oracle wants to receive all of that or believes that as part of an overall damages computation only a part of the fragmentation damage would count. Just to show why this is unclear, the filing also contains the following passage:

After inflating his royalty calculation as described above [adding up mobile advertising, lost profits, and fragmentation damages], Cockburn opines that Oracle should be awarded more than 50% of the net benefit he claims Google would receive from licensing the patents and copyrights at issue from Oracle, resulting in a greater than 20% royalty rate.

Even if I repeat myself a lot on this important item, anything that the expert believes to be reasonable compensatory damages could still be tripled based on the theory of willful infringement.

So there are some great unknowns here. Yesterday's filings added enough pieces to the puzzle to know for sure that this is -- in a worst-case scenario for Google -- about billions, not millions. But I don't know yet whether it's billions even without a tripling based on willful infringement or only after that. Prior to tripling, the amount could be "only" in the hundreds of millions. At the same time, just "fragmentation of Java" could amount to billions even prior to a possible tripling...

If you'd like to be updated on the smartphone patent disputes and other intellectual property matters I cover, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents.

Share with other professionals via LinkedIn: