- SEOmoz analyzes how Google handles local results. In most cases, Google puts Places results ahead of local organic results, but generally behind at least a few national results. This is something worth tracking over time, though; Google has shifted their local-versus-national and Google-owned-versus-web leanings several times in the past.
- Google is truncating results significantly for at least a few very targeted queries (so far in the form “topic” + “publisher”). There are probably subsets of queries where nearly every click after the first few is an accidental click, so this makes sense from a UX perspective. And not too many sites survive by ranking for their competitors’ brand names and news beats. So the net effect of this should be minimal. As a precedent, it’s interesting.
- A recent study claimed that Wikipedia ranks for 99% of searches. The methodology was terrible, though, so disregard it.
- Google has launched expanded sitelink-style AdWords. This pushes the top organic result down by several more pixels, and turns ads into more of a navigational tool. It’s not clear how often Google is using this, but it should raise their monetization per search incrementally, especially for some branded terms.
- Google has also optimized symptom-specific health searches in a way that sends more clicks to end condition pages, rather than intermediate symptom-discovery pages. The impact on WebMD is unclear, but it should at least cost them some navigational pageviews.
- In an interview, Google Search’s lead designer promises: “Don’t break search.” Good advice.
- Indie search engines DuckDuckGo and Blekko are both using Google ads, among other means, for monetization.
Owning the Dot-Com
Via Rand Fishkin, here’s a smart explanation of why and how to buy a .com domain name. Some of the advice makes sense, and some of it is pretty ridiculous (who starts a cold email by insisting that he’s ready to pay immediately?).
RJ Metrics Gets Pinterest’s Numbers
RJ Metrics has somewhat cleverly gotten some numbers on Pinterest’s user activity and its changes over time. They’ve found some ways to do simple cohort analysis, which provides much more robust data than what the company would release. Notably, Pinterest users are very active, and tend to stay that way—although comparisons to Twitter at the same age rather than at the same time obscure the fact that online culture has moved much more towards retweets/repins/tumbls as a medium of expression.
Using UPS Data for Merchant Loan Underwriting
Kabbage, whose objective underwriting standards are apparently substandard but whose PR and technology are top-notch, is now using UPS shipment data as a way to estimate loan quality. This demonstrate the shear volume of “data exhaust” being generated and tracked by the economy. As this kind of thing becomes more common, companies will have to optimize their implicit data output, both as a way to keep counterparties happy and as a way to thwart competitors.
And on the demand side of that equation, the SEC is talking about expanding the definition of insider trading to cover one-on-one talks between management and analysts.
Digiday Touts Performance Pricing
One of our bugaboos has long been the fetish for performance pricing in online ads, which tends to select for agent-principal conflicts: agencies know better than their clients how to manipulate the end result so it hits the right numbers at minimal cost, and the most effective way to do so is by doing things that hurt the brand’s long-term value (whether through manipulative SEO tactics, display ad buys that are indifferent to what content is displayed next to the ads, or some as-yet unexplored techniques). Digiday disagrees.
In other ad agency news, at least one agency is trying to diversify into investing in mobile games. “Branded content” has generally seemed more like a way to mark up content than to market better, but this time may be different.
Kayak Partners with TripAdvisor
TripAdvisor is a pure-play SEO company. Kayak is a search engine in need of a) content worth analyzing, and b) content users will care to click through. So their new partnership is entirely sensible.
Zynga Reports Earnings, Aims for Mobile
Zynga has over 15 million mobile users, a massive gain. That’s particularly good news for them given Facebook’s ownership of their user acquisition and a substantial fraction of their revenue. Meanwhile, earnings for the quarter were down even ignoring special charges; they haven’t optimized their hit generation so it coincides with quarters (although between their analytics-driven culture and the demands of, well, analysists, they may get better at this over time; channel stuffing can happen at the consumer level, too).
Zynga has also launched a licensed, gambling-oriented game, and is apparently just daring investors to remember that IGT paid 50X sales for a Facebook gambling company a few weeks ago. Zynga alsowants to create a publishing platform, making it a platform-on-a-platform given current circumstances. (That, or a non-Facebook platform that is cometing with Facebook as a platform and offering an inferior alternative, in history’s most acrobatic case of channel conflict ever. Don’t miss Chris Dixon on platform risk.)
The eBay Strategy: Decentralized Commerce
Chris Dixon (now working at eBay, though probably not for long) explains the difference between eBay and Amazon. eBay decentralizes commerce, and provides simple tools for increasing the number of stores/storefronts in the world. Amazon centralizes it, and provides tools for cutting costs as long as Amazon keeps the data and any change in the margins. At one point, he cites this theory: “Who can beat Amazon on price? The companies whose products are sold on Amazon.” Not necessarily. Factor in costs of capital, and Amazon can often easily compete with its suppliers’ direct channels. And add in the game theory side (Amazon can maintain a high enough order volume to make all-or-nothing demands), and their competitive position is far stronger than it looks. It’s easy for online entrepreneurs to round frictional costs down to zero, but Amazon is a monument to nonzero frictional costs.
Yes, Twitter Broke the Whitney Houston Story
Although the first few tweets were unconfirmed rumors, Twitter did, indeed, manage to break the story that Whitney Houston died. The next step is to plug Twitter Firehose access into a set of templated potential stories (e.g. “[Politician's name] concedes,”[Other celebrity] dies,”) and add a layer of Narrative Science, and we’ll be able to have automated breaking news without any human involvement.
Demand Reports Earnings
Demand Media reports marginally higher media and registrar revenue, plus writedowns due to junking low-quality content. They’ve finally seen a drop in content monetization on a quarterly basis, but they’re still seeing pageview growth—an indicator that they’re successfully replacing high-monetization search traffic with low-monetization social media clicks.
Taking the Upfront Online
A few video publishers have decided to borrow the practices—and some of the language—of TV “upfronts,” to create the “Digital Content New Fronts,” per the WSJ. Part of the point of the upfront practice is that TV is extremely predictable; selling video ads this way sells them short. (It is, however, a good way to distinguish between expensive premium inventory and the rest.)
Valuing the App Ecosystem
Asymco calculates the net present value of apps per iDevice, without going into too much detail on marginal data from new quarters (which, admittedly, will muddle things singificantly). The growth rate and margins in apps imply that a) they’ll be a big chunk of total iThing margins in the future, and thus b) Apple may be able to release phones with lower price points, and monetize through high-margin apps instead.
Anonymity and Writing Style
33bits, the privacy-busting academic blog, links to a new paper that can identify anonymous writers in part based on writing style. As we’ve argued before, this may end up being used for ad targeting: even if it’s a false positive, two people who write very similarly will likely share other traits (location, education, profession), which can be enough to enhance ad clicks.
AOL’s Patch: “12 of the 863″
Fortune has a brutal article on the weakenesses of Patch, which is growing fast from a base of basically zero. This business depends on a peculiar kind of economics, where some parts (technology, ad sales) are scalable, and others (labor markets, which are tapped by Patch and local newspapers) are not. So far, it’s not working well, though what’s really not working is Patch’s status as a small startup embedded in a publicly-traded company whose other businesses are more amenable to traditional analysis.
Digital Due Diligence Weekly