Two days ago, I published the chart below with monthly estimates of Groupon’s U.S. revenues. The chart shows a startling 30 percent falloff in February from the month before. As I noted in the post:

Again, these are just estimates based on the equivalent of scraping Groupon’s site, and thus could be missing something.

Well, at least for February, it looks like those numbers are way off. The post obviously caused some ripple effects to the extent that Groupon had to start addressing the issue with potential hires. As a result, it knocked loose the real revenue numbers for February and January. Groupon wouldn’t comment on the revenue numbers when I asked them about it, but according to a source, Groupon is now privately countering the numbers in my post: instead of $62 million in U.S. revenues, the company did $103 million in February. And that is up from $92 million in January (compared to the $89 million in the original data below).

I did some checking around, and I’ve been able to confirm that these two numbers (the $103 million and the $92 million) are right. I was also able to confirm that the 60/40 mix between U.S. and international revenues is about right.

Google is moving in many directions—mobile, browsers, productivity apps, operating systems, social. At first glance, it may seem like it is trying ever so hard to move beyond its giant one-trick pony: search. What people keep forgetting is that it is a pretty good trick. Benchmark Capital VC Bill Gurley reminds us how good this trick is in an excellent post that looks at Google’s market expansion strategy not as one of a series of aggressive offensives, but rather a highly defensive strategy.

Warren Buffet famously describes the best businesses as “economic castles protected by unbreachable ‘moats.’” Search is Google’s economic castle (perhaps with other forms of online advertising such as display thrown in there), and everything else is a moat trying to protect that castle. Android is a moat. The Chrome browser is a moat. The Chrome OS is a moat. Google Apps is a moat. These are all free products, subsidized by search profits, that are intended to protect the economic castle that is search.

It appears there’s an arms race taking place in the mobile gaming world, and as is often the case in Silicon Valley, it starts with engineering and development talent. Google, for one, took some serious measures at the end of last year to stem the flow of its engineering talent to Facebook. Many companies go out of the way to court and retain top developers, but it’s traditionally been the deep pockets and reputations of big players like Apple, Google, and Facebook that win out.

Apple mobile iOS devices (iPads, iPhones, and iPod Touches) are used by 130 million people, but they present a huge blindspot to advertisers. All Apple mobile devices use the Safari browser, as do millions of Apple laptop and desktop computers. Safari blocks third-party cookies by default, which is good for privacy and good for consumers. But it is bad for advertisers who rely on browser cookie tracking to measure the effectiveness of their ads.

Marin Software, which offers a way to manage paid search advertising, conducted a study it provided to TechCrunch which shows that 80 percent of the time iOS devices don’t count paid-search conversions (i.e., clicks) because cookie-tracking is turned off. On the Mac, the undercounting occurs 50 percent of the time. All told, when you count all browsers, 38 percent of all paid-search clicks are not being counted.