Monday, June 29, 2009
Back in 2007, we briefly reviewed the major M&A activity in the digital ad serving technology and digital agency spaces. Many of our panel observers felt that Google and Microsoft should immediately divest themselves of the agency parts of the DoubleClick/Performics and Aquantive/Razorfish acquisitions, to clear out some of the conflict of interest inherent in major agencies owned by the sellers of supposedly performance-based, impartial media platforms.
Google moved relatively quickly to divest itself of Performics, whereas Microsoft held onto the Razorfish business -- until now.
Amazingly, the $6 billion acquisition of Aquantive counted as Microsoft's biggest ever acquisition. Considering its size and clout, Razorfish (formerly Avenue A | Razorfish) has had a relatively quiet two years since then. Perhaps this can be chalked up to this mega-agency's DNA; its first go-round in Bubble 1.0 was in the fast-hiring, website-overbuilding, overvaluation heyday of 1995-2000. And the company still seems to favor slow-loading, expensive-to-build, semi-indexable pages. Like ALPO, a recent client win for Razorfish, could it be that Razorfish represents a previous era of overpackaged, overstrategized goods with the same old ALPO inside the can?
While some may ask why Microsoft is selling this business, Daily Finance reporter Douglas McIntyre proffers: "What it is doing with the company in the first place is anybody's guess."
Certainly, if the long delay in selling Razorfish was helpful in demonstrating that Microsoft's decision-making process was totally independent of industry opinion that they should divest sooner, the delay was effective. Unfortunately, the value of the asset may now be sharply reduced. But so are many assets... such as the parts of Yahoo Microsoft is still considering strategically partnering with or buying.
Labels: msft, razorfish
Wednesday, June 17, 2009
Microsoft has gotten us one step closer to the hard-line Traffick stance of jail time for click fraud. Making money by directly yanking money out of advertiser pockets in industries like auto insurance must have felt very good to these alleged fraudsters: until the day they were slapped with a $750,000 lawsuit. Ouch.
Labels: click fraud
A lot of ink gets spilled about everything digital media. And some wonder why those of us in the search side of the industry are, well, a little bit... noisy.
You can see why by visiting Alexa (or if you have a subscription, a service like Hitwise). For your favorite site, check out the tab called "Clickstream". The website that is most visited prior to visiting your favorite site is, of course, Google.
It might look like:
Or it might look like:
8% Some other site.
Or in some other cases Yahoo actually is still key (though not necessarily Yahoo Search):
And the "post visit" clicks are also (although, hopefully, less so) to these search engines as well. Users either go back to the search engine right away due to search dissatisfaction, or they use a search engine for something else, after they're finished with one task.
Other common upstream and downstream sites are site closely related to your favorite site, like
Or functionally related ones like:
as people tweet what they find.
For some popular sites like cuteoverload.com, Twitter is hitting 2 and 3 per cent of post-visit visits. For this blog, Traffick.com, Twitter clocks in at an amazing 10% of downstream visits!
Still though, Google is at the top of pretty much every list of upstream and downstream clicks for pretty much any website.
Bing.com? The bad news is, the site is too new and too little-used to show up in many upstream lists at all. The good news? Downstream, it shows up in the list for blogs like this, and tech industry publications. Microsoft will need much more than mentions from the likes of us, though, to gain market share.
Saturday, June 13, 2009
Remember how Yahoo helped Google rise to prominence, then dominance, ultimately obliterating the search and portal brand that gave it that timely helping hand? If you've been in the industry less than five years, you don't. In 2001, Google's future was still far from assured. But the power its brand gained during its two-year partnership to display Google results on Yahoo did serious damage to Yahoo's brand, and bought Google the credibility it needed to move to top-of-mind as the world's leading search engine.
Is Google about to do the same favor to Twitter? On the surface, it seems that adding microblogging search capability that features mostly Twitter results would only benefit Google, by giving it equal or superior capability to the as-yet-to-be-honed Twitter Search. Another feature to solidify the world's leading search brand. So as for whether this helps Twitter do to Google what Google did to Yahoo: probably not. Google today is far more than Yahoo was in 2002-2004. Still, it's interesting to contemplate the possibility that a deal that looks only so-so for Twitter might have a salutary effect on Twitter's already strong brand, and turn out to be the mainstream exposure they needed to go from hot to white-hot.
Labels: google, twitter
Friday, June 12, 2009
Even in a down economy, or perhaps because of it, companies are investing more than ever in performance-based digital marketing. So it's perhaps no surprise that the SES Toronto conference turned in another strong performance this week. I was delighted that so many of the top speakers (too many to name here, really... ok... Mark Evans, Keith Boswell...Miriam Warren... ok I'm stopping...) we invited were able to make it. And if I do say so myself, the tracks (Nuts & Bolts, Corporateville, and Geek) seemed to be working well in getting folks connected with the type of knowledge they need in this fast-moving field.
The demands on marketers have begun to change more rapidly as reputation management, social media, and universal, personalized, blended, local, and mobile search have created a much wider array of relevancy signals and visibility channels. While introducing newcomers to the basics, the theme of the show was intended to expose us to data and debates about what's new in the field. I feel that companies that opt not to show up to these events - especially those constructing digital marketing plans from scratch - may be failing to ask the right questions in planning. It's too easy to fall into the trap of micromanaging tactics based on 2002 assumptions, on one hand, or to concoct hip-sounding but amateurish attempts to bust into social media (the Meatball Sundae syndrome).
For a quick flavor of how the conference went for some, we offer for your consideration a list of tweets about SES Toronto 2009; some of the SES Toronto interviews posted on YouTube; and a selection of the pictorial evidence on Flickr.
Our keynote speakers, Tara Hunt and Emanuel Rosen (who just tweeted that Toronto is now his favorite city in North America!), deserve special thanks for helping us "anchor" digital marketing in real-world reality. Search engines measure relationships and relevance, but in the past have done a poor job of it. Marketers who only focus on the tactics suited to imperfect search technology from the past fail to see the need to create a variety of connections and authentic social capital upon which strong referrals are built.
The opening night party at The Drake Hotel (Acquisio sponsoring, with co-sponsors Page Zero and NVI Solutions) must have been good, too. MJ Lepage of Acquisio (pictured here) convinced me to speak French. I was shy at first, but those Acquisio t-shirts have amazing powers of persuasion.
Labels: ses, ses toronto, ses toronto 2009
Saturday, June 06, 2009
Interesting piece at ReadWriteWeb about how the Twitter insiders actually use Twitter. Called out on their idiosyncratic use which doesn't resemble that of some power users, Ev Williams responded directly via email. A highly relevant excerpt, to me, is:
"Many people fall into the trap that you should follow all or most people back out of a sense of politeness or so-called engagement with the community. But the fact is, having more followers does not give you more time in the day (as much as I'd like to sell that). At a certain point, you're not actually reading any more tweets by following more people -- you're just dipping into the stream somewhat randomly and missing a whole lot of what people say.
"That's fine, but I believe people will generally get more value out of Twitter by dropping the symmetrical relationship expectation and simply curating their following list based on the information and people they want to tune in to."The argument that the developers' own use of a tool affects the company's approach to building the tool, and its sensitivity to different user constituencies, is pretty much specious. I don't expect the developers of Word to be sending memos inside a law firm, I don't expect AdWords product developers to open an online apparel store, and come to think of it, I don't want to know what's on Steve Jobs' iPod, or how often he listens to it.
Thursday, June 04, 2009
Miriam Warren, Director of Marketing with Yelp, is speaking on the cool mobile apps panel next week at SES Toronto, moderated by none other than Mitch Joel.
She was generous enough to answer my many questions about all things local and mobile. We've posted the results over at the Search Engine Watch blog.
If your iPhone literally "tells you where to go" (with the help of an uber-reviewer like "Andre D."), rest assured Miriam will be pleased.
Labels: mobile search, ses toronto, yelp
Wednesday, June 03, 2009
"Go to Bing. Try this?!," I said.
Carolyn and I were in the middle of watching Blue Jays highlights; me, fresh back from the Microsoft Canada demo of the Bing search engine. Her, an intelligent Ph.D. who somewhat knows her way around the Apple OS, fruitlessly trying to use Google to get the answer to a baseball trivia question we were pondering. If Roy Halladay had 14 strikeouts in last night's game in his masterful, complete-game win over the Angels, what's the rundown on the most strikeouts he's ever had in a game?
"Bing! dot com," I burbled, helpfully.
"What?," she "replied." No response yet. Apparently "Go to Bing" does not yet resonate as a domestic communique.
I never do this.
"Bing! Really -- try it at Bing.com this time. Type 'Roy Halladay shutouts single game' or whatnot." (We search experts try not to give more precise directions. It makes normal people feel uncomfortable.)
I've seen many demos (by Microsoft, even). But rarely do I do just like the people in Microsoft focus groups did: notice a difference with the engine's usability, and recommend it to someone.
We didn't find the stat quickly.
Then I remembered that the Bing engine has been tuned to offer more orchestrated consumer-friendly results pages when you type in "Roy Halladay" without any qualifiers.
Sure enough, Carolyn saw a pretty useful Roy Halladay search result full of photos, stats, and search refinements... just one notch short of a Roy Halladay shrine.
We try a similar search for "Scott Rolen". A similar result ensued. Carolyn remarked that the page came complete with a "nice BIG picture" of Scott Rolen. It's at moments like these that it hits me out of the blue: women do like men, and will occasionally admit it even in your presence. Then again, it was just a bad head shot. Perhaps she meant he had a fat head in that shot. That's pretty much my take on it, anyway. Kidding, Scott! Love the Gold Glove play out there! Keep it up! Whew, didn't want to alienate that guy.
The top "related query" in the left navigation was Scott Rolen wife. Being a curious sort, and a man, I thought I'd love to know why. I still don't know. The query results lead to some pretty generic looking Ten Blue Links. Undaunted, I tried the image search for the Scott Rolen wife query. Things break down a bit from here: Bing's image search is as unhelpful as Google's often is. The top photos are all of the ballplayer, other ballplayers, managers, shots of dirt, etc. The highest ranking woman pictured is of an Asian bible college student who is not Scott Rolen's wife. I don't ask. In any case, the page is now down.
Wow, search is distracting.
I look over at Carolyn, who seems to be talking. She's talking about how she likes the layout of the search engine, remarking on how useful it would probably be - I coach her a bit to reinforce the idea that it will pull up better pages for these popular queries. She reacts pretty well, contrary to the usual reaction you see to a demo. Basically - she doesn't hate it. And had a few good comments.
This home focus group would have probably produced positive feedback clips on a par with those Microsoft showed us in the demo, of the average cafe or office users giving their take on whether they found it useful: the dainty twenty-something with tattoo-covered arms; the office manager who doesn't like to "re-work" when he's searching for a hard-to-find stat like "World's Best Rapper,"; the affable round-faced chap who "never gives up" when he does his goal-directed searches for things like "nightlife when he's in Italy," but would rather get to what he's searching for faster.
Microsoft's research is impeccable. According to Stacey Jarvis, search lead for Microsoft Canada, a high number of search queries simply do not deliver successful results. She noted, metaphorically if nothing else (since it probably isn't competing with the Print button), "the back button is the most frequently used button in search." So search still sucks? Well, it's definitely unsatisfying a lot of the time. But luckily not quite to the point where users won't keep trying.
Also germane to all of the above users, and you, and me: about 50% of queries are about returning to previous tasks. So we need to get to information more quickly.
One evocative persona was a woman who took 27 minutes and had to retry her search five times to find a retailer for the Merrell shoes she was looking for. You think that sounds painful? That's a US example. In Canada, stock is much spottier and domestic e-commerce players are often harder to find and harder to shop from.
The "managed" results on broad, popular queries I alluded to above are called Best Match. If you type Toronto Blue Jays, you get a schedule, a Wikipedia entry, news, and other "most useful" items... every time.
For search optimizers, this should be a continuation of the trend set in motion years ago by all the search engines. If the door was slowly closing on the opportunity for "just any website with the right SEO strategy" to rank well on broad queries, does this slam it shut? Good for consumers, bad for SEO's? Forget ranking on head terms, folks.
Freshness is a big part of the rolling mandate here. The Toronto Blue Jays result page also shows a current box score with the annoying news that it's 3-0 Angels in the third inning. Whoops, 4-0. Let's hope Casey Janssen settles down: he's got Halladay caliber composure and stuff... minus the extra heat since his injury.
Other useful real-time features are being dutifully built into Bing.
Type "Flight Status" and you get a handy app to look at your flight's status. A bit of inside baseball came up at the demo: Mark Evans asked "Is that part of Farecast you're integrating?" and Travel Ninja Stuart MacDonald interrupted from the crowd to say: "Flightstats isn't Farecast." Apparently, Farecast integration is down the road, if that means anything to you.
Yes! Matthews jr. struck out. Oh no! Janssen hit a guy with a pitch.
Back to this.
Microsoft - and others - are good at researching pain points and they're good at building features. Information retrieval is complex, though. And users are using one main technique to find answers: typing short or long queries into search engines. Repeating queries, trying different results. That's the case on Google as much as it is on Bing.
Bing is following, using more sophisticated technology and database integration, the spirit of the old Ask Jeeves answer engine or the recent thoughts behind Mahalo. Figure out how to be reasonably close to the right intent on the most popular 80% of queries.
For hard-to-find information, it may well be the quality of the underlying results that blocks us from getting there, so is it really up to a search engine to fix that? Well, certainly Wolfram Alpha is trying to get us answers to very specific queries like "what are the highest single-game strikeout totals for Roy Halladay?"
We got to the answer easily in this case only because the headlines told us that Halladay's 14 K's was a personal best. Search engines don't always make it easy to get specialized information.
That just underscores the point of Microsoft's research: many search journeys today still end in frustration. The fact that Microsoft's response will leave some users pleased, and others most certainly frustrated, still seems like progress. The new buzz in the industry being generated by the likes of Bing and Wolfram Alpha is good for consumers, even if they all still offer only very partial solutions to the problem of finding exactly what we "want."
So if Bing's flavor is to do well on popular queries with an unsophisticated "mass" audience, does it follow that it will gain in popularity, backed by an $80mm advertising campaign? When specifically asked, Stacey Jarvis was candid that Microsoft's current search share in Canada is 4.9%. But no one at the company is willing to stand behind any boasts about concepts like "switching search engines," market share, and (with some exceptions, like in the focus group videos) "Gooogle." They're united behind what is evidently a very good product that can improve more with refinement. And quite rightly, daring to predict widespread consumer adoption of Bing isn't in the Microsoft script at the moment. Do they have a plan here, or just a solid product?
It's now 5-0 Angels, with the Jays still hitless. But I probably won't watch the rest of the game on a search engine. It's off to the gym, where I hope Dr. Phil isn't showing on the TV's.
Note: Stacey Jarvis, Search Lead for Microsoft Canada, will speak on the Orion Panel on the Future of Search at SES Toronto, next Monday, June 8.
Labels: bing, microsoft search
Microsoft's new search engine, Bing, is now live in Canada, one of three countries (the US and the UK are the others) to get full access to the engine out of beta. I attended an impressive launch demo last night in downtown Toronto.
Before getting to the serious nitty gritty review -- you knew this was coming right guys? -- I have to mess with the Microsofters a little bit. Bad news first, full review (mostly good) later this evening.
The bad news is, you can't search Bing to find any relevant results on the query "microsoft canada bing launch." Looking at the screen shot below, that isn't the worst of the problems. The worst problem is that the familiar old eBay dynamic keyword insertion ad made it to the top of the right rail on that long tail query!
Fantastic prices on Canada. Buy &
Love it! Always loved it!
Ha, and here I thought our dollar was making a comeback. :(
To their credit, on my favorite "embarrass the ad-engine" query, [schnauzers], the Bing result is really nice - with no ads, and helpful refinements in the left nav.
This does highlight a key issue with Bing - both a strength, and proof that there is much left to be done. There's a gap between the satisfaction you'll likely receive from popular queries -- the Microsoft team is optimizing results to provide quick answers to people's questions on the most popular generic queries -- and the not-much-improved functionality for longer tail queries.
Danny Sullivan reports on an update in Google philosophy and algorithmic emphasis. As I interpret the story, at one point Matt Cutts didn't rule out the idea that you could slap nofollow tags on some internal links on your pages to conserve "link juice" for your important pages. Hence, the practice of "PageRank sculpting" was born. SEO's had another cool story to tell their clients, and each other.
Danny's interpretation of this is that making this kind of technique available, and then taking it away, is a violation of a broad principle of "backwards compatibility". Shame on Google, he implies, for making the advanced SEO's scramble to undo what they already did now that Google's algo has supposedly undergone this massive shift and a page with ten links passes only 10% link juice to each link on the page, rather than, say, doubling the juice on the remaining links if you nofollow half of them.
I don't think I agree. Here, Danny is standing up for the constituency of advanced SEO's, many of whom are currently attending SMX Advanced. My take is that SEO's taking actions on speculations about the algorithm are themselves building the new "features" that lack "backwards compatibility." This is especially the case when the "features" (tactics) address no known principle of third-party trust or relevancy of sites or pages.
But for those of us who don't believe all of Matt Cutts' stories and non-stories, and take a holistic view of business strategy, information architecture, audience development, and traffic growth, we had a lot of lower-hanging fruit to work on than using a short-term fad method of "telling" Google which pages are important.
Long term, search fails when site owners try to "tell" search engines which pages are important, short of burying the unimportant ones in their architecture so they're literally invisible. Importance shouldn't be arbitrarily determined by site owners, though certainly users and engines appreciate it if they provide indications.
Search Engine Land itself has undergone a surge of traffic in recent months, all no doubt a product of holistic audience development. I'd love to hear Danny's take on how much of that improvement in fact resulted from deliberate PageRank sculpting. None? A lot of it?
No matter: holistic brand building and audience development and overall quality content, combined with sound organization/architecture of the content, are what gave Search Engine Land its mojo - not short term tactics. And that's how most companies should look at the SEO exercise.
What was supposedly "given" to advanced SEO's in the short term has now been taken away. Nofollow and for that matter XML sitemaps are just supplements in a much more important larger "grand scheme of things." PageRank sculpting turns out to be just another time-waster that contributed mainly to "Advanced SEO" bragadoccio on the barstool. It's gone now? Boo hoo.
If you spend your life hanging on Matt Cutts' words about SEO, well then mark these words: you will, in turn, find yourself annoyed with Matt Cutts.
In conclusion, I propose a new tag that should only be used by Advanced SEO's - square brackets used so as not to screw up the HTML on this page: [this page is really frickin' important] [/as you were]
Rand Fishkin, in a spiffy flow chart, seems to approximately (and diplomatically) agree with my take, highlighting the key low hanging fruit that comes before frivolity like PR Sculpting: content development, information architecture, link acquisition, internal link structures, and conversion rate optimization. But as for Rand's example of a site that is a large one with many deep URL's, as an example of one that might benefit from the sculpting, this might depend on the query we hope to rank for. Overall, I've seen no major problem ranking very deep pages on relevant long tail queries (for example, at HomeStars.com). Those pages rank or don't rank for a variety of reasons, as many as the tail is long. And from what Cutts is saying today, the point is moot anyway, as the provisional tactic/loophole has now been closed. Back to working on the important parts of the business.
Hmm, and I'd love to think my take on this is just common sense and uncontroversial, but the statistics out there seem to indicate (even outside of major sites like Wikipedia) that there was a massive rush to play with the nofollow tag among the "SEO community". Like Rip van Winkle, I slept peacefully through the stampede.
Labels: information architecture, pagerank, seo
View Posts by Category