Skype Buys GroupMe
Skype has purchased group-texting service GroupMe for about $85mm, a nice exit for the year-and-a-half-old firm. It’s also an interesting combination: one of Skype’s big strengths is its peer-to-peer structure, which basically coopts users’ computers to keep things running smoothly. GroupMe uses Twilio, which is basically an API for anything a standard phone can do.
Skype is very much a creature of the early stages of the social web: still pretty scrappy, and built on a model that exploited cheap PCs and more bandwidth than we knew what to do with. GroupMe is a more modern kind of company: the model only works because of a good funding environment and someone else’s platform.
Front-Running Twitter Registrations
One Twitter user has discovered that someone is registering Twitter accounts based on new domain registrations. This could be a coincidence, but either way the model breaks down: it’s hard to contact the owner of a dormant Twitter account. If the people doing this scheme serious about it, they should create some kind of general contact page and link to it from each account.
“Why Software is Eating the World”
Marc Andreessen has a fascinating manifesto in the WSJ, on how software is becoming increasingly unavoidable in non-software businesses. But it’s important to be precise, here. Software is changing industries with highly automatable work. Direct marketing is very automation-driven, hence Google and Groupon. Health care and education are very hard to automate, so the biggest successful startups in these fields tend to be very good at navigating whatever payment / reimbursement scheme is in place.
Euro-Panda’s Casualty List
SearchMetrics has released a list of sites hit by the global Panda update rollout. The usual suspects are well-represented: people search, aggregators and scrapers, and a few sites from higher-quality publications that got swept up in the general update. Shopping.com (an eBay property) was one of the more consistently hit sites.
30 Years of Music Industry Market Share
This is fascinating. Digital Music News has a series of images showing which revenue streams dominated the music industry over time. One surprise: CDs hit their peak market share in 2002. The decline from peak wasn’t due to digital music at first—it was due to music videos accounting for more of the industry’s revenue. Digital is getting more significant, but even in 2010 CDs and LPs together had a 51% market share.
Facebook Tests Out Quora
Facebook is using Quora to get feedback on their internships. Clearly they should have started Facebook Questions with nerds, too.
Apple Breaks Mobile Retargeting
Apple has changed their API to make it harder to track individual devices. It’s hard to view this as a privacy PR stunt: most users are probably unaware that they can be tracked at this level of detail.
But what’s interesting is how much of the ecosystem this breaks: not only will it hurt ad-based apps, but it also makes it harder to target non-mobile ads (by matching logged in mobile accounts to logged in desktop accounts, a site could use mobile behavior to target desktop ads—since desktop ads get more clicks but a mobile identity is more persistent, this is a net win).
Apple may plan on reintroducing this feature as part of their iAds network, assuming they don’t shut that down.
(AdExchanger has more.)
StumbleUpon Tests Curated Search
StumbleUpon is testing a new interface that lets people search based on tags. That’s a similar model to Blekko, but with a far wider corpus of data. This looks a little bit like a business (selling CPC ads against head terms) and a lot like a way to pretty up the product before a(nother) sale.
Google is Verifying Google+ Accounts
Finally, Google is doing something a little less scalable: they’re verifying that Google+ accounts belong to real people. One interesting part of this story is that they’re using follower count as a sign of validity. That’s a flawed approach. It’s the Internet: people want to believe. They’re basically outsourcing the hard work to publicists if they assume that the most-followed Lady Gaga is the real deal.
Bing’s Stats on Facebook Fans’ Non-FB Activity
This is worth reviewing: people who “like” Bing on Facebook are more likely to do Bing searches—and so are their friends. As always, it’s hard to tease out correlation and causation. One option would be for Bing to show Facebook ads to a randomly-selected demograpic (for example, showing ads for the Bing fanpage to Facebook users in their 20′s with even-numbered ages, and not with odd-numbered ages) to see if that has a predictable effect on search volume.
The Odd Economics of Paid Online Reviews
Online reviews are currently too cheap and cost-effective to be stopped. The NYT has a piece on automatically identifying them, but the heuristics used are too likely to lead to false positives—especially since people who pay reviewers will now know what to look for. It might be worth thinking of the economics of reviews in a different way: maybe they’re a signalling mechanism, like ads are. A 3-star restaurant is going to pay for different reviews (“Great value!”) than a 5-star restaurant (“Amazing experience!”), and their reviews will only have a normal distribution if the fake reviews match up to the characteristics that the real reviews are going for. Fake reviews may end up being like ambiance: the restaurant is paying for something fake, but it enhances the real experience. Ambiance won’t turn Burger King into Jean Georges, but Jean Georges is what it is thanks to great ambiance and great reviews, not just amazing food.
The most popular solution—the one Yelp is built on, in fact—is ensuring that reviewers are real people by a) having them create persistent accounts, b) matching these accounts to Facebook accounts, and c) finding ways to meet Yelp reviewers in person. Reviewing sites generally don’t have the critical mass necessary for everyone to have at least one friend-of-a-friend’s review of every establishment they’d like to visit. Foursquare might come close, since their interaction is a fifteen-second checkin rather than a fifteen-minute writeup.
Google Gives Site Owners More Control of Search Results
One unspoken result of Google’s change in sitelinks display is that site owners have far more control over how their site is portrayed. A popular site that advertises on its own brand name can now control everything above the fold for about 70% of browsers. Take IBM, for example. Including URLs and title tags, they can now write 141 words of copy that will show up when anyone Googles their brand name. This move makes organic search function a bit more like AdWords, and thus hands even more screen real estate over to existing winners.
Digital Due Diligence Weekly