New Web Metric Likely To Hurt Google

StonyandCher write(s) with news that one of the largest Net measurement companies, Nielsen/NetRatings, is about to abandon page views as its primary metric for comparing sites. Instead the company will use total time spent on a site. The article notes, "This is likely to affect Google's ranking because while users visit the site often, they don't usually spend much time there. 'It is not that page views are irrelevant now, but they are a less accurate gauge of total site traffic and engagement,' said Scott Ross, director of product marketing at Nielsen/NetRatings. 'Total minutes is the most accurate gauge to compare between two sites. If [Web] 1.0 is full page refreshes for content, Web 2.0 is, "How do I minimize page views and deliver content more seamlessly?"'"

But

[Verte]

you can't measure that...

Idiotic

[ucblockhead]

In my experience, most people don't bother to close their browser when they are done browsing. It's even worse for people used to tabbed browsing. How many times do you shut down the computer at night with tabs containing something you looked at with your morning coffee? I know I do as often as not.

So?

[Mr.Progressive]

Google is its own Web metric

Tabbed browsers

[gmuslera]

Sometimes i visit a site that links a lot of places (an common one is a google search) and open every site in a different browser tab, and then i read. Now, the last tabs are likely to be there for long time, either till i close it, read it, or even click on links there. How that kind of behaviour gives more weight to the sites i opened at the end?

And to make it worse, most browsers now support tabs.

Re:But - well, what about sessions?

[Verte]

Sure, sessions will work for sites like forums. However, is there going to be anything shown by session length that won't be shown by page views in that case? What about pages that you can really spend days to weeks at a time staring at, such as the glibc [gnu.org] manual or the Coyotos microkernel specification [coyotos.org]? If the user never refreshes the page before the end of the session, information-packed sites aren't going to be measured at all.

Re:Idiotic

[needacoolnickname]

That doesn't matter. Assuming you don't have some kind of page refresh every n seconds, most analytics software have timeout values between page loads. If you don't close your browser and then come back the next morning and continue where you left off, the analytics software should see that it's been more than 30 minutes between page loads and consider it a new visit.

That might be true, but what about when I open a link in a new tab from something I am reading but don't get to it for another 20 minutes. After I get to it I notice that the link is crap and close it right away. Total time spent = 4 seconds. Total time they think spent is 20 minutes 4 seconds.

only one "metric" matters

[timmarhy]

the bottom line, is how often does someone click on your ad and have it result in a purchase? anything else is meaningless.

No good can come of this (a new slower web)

[G4from128k]

I predict this change will lead to more sites where all interaction and pacing is under the control of a designer, not the user. I can see it now:

PHB: "How can we get people to stay longer?"
Eager-Beaver Designer: "Let's put everything in Flash, put fewer words per screen and longer pauses between new screens."
PHB: "Great!"

My point is that I am a browser and I use a web browser. That means I want to browse. That means I want to be able to glance at something, make a quick decision, and control the movement to the next chunk of content.

This emphasis on viewing time will cause designers (and their bosses) to try anything they can think of to slow down the user.

All well and good.

[xigxag]

But surely advertisers don't care how long you stay on a site except insofar as it increases your exposure to their ad. E.g., on Slashdot, you might spend ten minutes reading comments but quickly scroll past the ads in the first 30 seconds and the rest is all content. However, if you choose to post a comment, an ad is visible on the comment pages and stays visible during the duration of your composition. I'd say the second ad, continuously viewed during the three mintues it takes to write a comment, is more valuable than the first ad, which goes off screen almost immediately.

Seriously retarded method

[goombah99]

What matters is not page views or page durations but redirects to pay-sites. That's the value of any site from an advertisers point of view. When I read the NY times I spend a long time there but I'm not likely to be shopping and even if I was in a mood to shop the probability they happen to show me an ad for something I'm interested in is close to nil. Plus the adds they tend to show there are delux flash moving ads or big long columns.

Now when I go to google and type in "blow up dolls" or Airline miles or 629 investments or some purchase worthy topic, I read the ads. Not only that but the ads are short. so I don't spend much time. But I click through a dozen of them into tabs in a few seconds.

When I watch you tube, how long to I look at the ads? probably not at all--I scroll then off screen. But I do see the adds on the leader of the video. But that's only a few second on a 5 minute video. A good and focused 5 seconds yes. Even subject worthy 5 seconds. But not 5 minutes.

How does the "Web 2.0" metaphor work with Porn?

[Mr. Underbridge]

If [Web] 1.0 is full page refreshes for content, Web 2.0 is, "How do I minimize page views and deliver content more seamlessly?"'"

If 1.0 is surfing one handed while jerking off, Web 2.0 is having a USB pocket pussy connected to an interactive AJAX interface.

In all seriousness, can we dispense with the fucking Web 2.0 crap? It is today's "information superhighway". If you use the phrase without a hint of sarcasm, you are an idiot.

Hurt Google?

[teebob21]

Can someone please explain the rationale for declaring that a metric change will "hurt Google"? When is the last time someone decided to use a particular site based on a commercial web-rating? I certainly don't use Alexa to decide which news sites interest me, at which banks to do online banking, etc.

Certainly there are a few closet Google employees around here... So tell me, are the higher-ups even remotely concerned with a traffic ranking? I mean, if suddenly MSN Search spikes up over Google in the ratings because its so goddamned user-hating that it takes 3 minutes to search a single topic...is anyone going to blow a gasket, provided traffic and revenue remain at present expected levels?

I sincerely doubt it.

I'm sorry...

[Rachel Lucid]

Isn't the sign of a good site that people are able to get what they want QUICKLY?

This can only 'help' pages that take forever to load...

Re:Of course you can

[grcumb]

Use some programmability/flash/whatever to keep pinging back to the host.

Right, so the users behind my NAT are going to be measured as one person spending all day on somepopularsite.com, in 8 different places simultaneously? What about the four other open tabs currently open in my browser? Am I still visiting those sites? The answer could be 'yes', but I don't see how that adds value for advertisers.

HTTP is a stateless protocol, which means that it's inherently difficult (i.e. impossible) to consistently get accurate data about the duration of a given visit. It can be argued that you can derive data that's statistically significant. You can argue further that if everyone uses the same metric then they'll be valid for comparison purposes, which is enough for the MBAs in Marketing, I suppose.

I personally think time spent on a website is a silly metric, and will continue to hold that opinion until someone can make the case that staring at an advertisement for longer period of time actually encourages a person to finally click on it, rather than tune it out completely. (This works well for branding, but for little else.)

There's a lot of nuance that can be brought into this discussion, and this is where the good advertisers and marketers earn their keep. Assuming that either page views or time spent on a site are sufficient to make a solid judgement of the value of a given website is, uh, a little short on nuance.

Re:My God!

[vic-traill]

Now that's good fer a god-honest knee-slappin' guffaw!

Thanks - I needed that.

Just so I don't get karma-slapped upside the OT head ... I've always thought of Nielson as a mechanism for pricing ads; like all representations of average behaviour, it doesn't say shinola about a particular individual's viewing habits. So, as long as the advertisers think they're getting value out of the metric, that's fine. But I've never talked to anyone who used a Nielson rating as a TV viewing guide.

Similarly, I've never talked to anyone who uses Nielson/NetRatings as a measure of the usefulness of quality/level of interest/etc. of a web site. And NetRatings doesn't even have the mindshare of Nielson the TV dudes. Anyway - in the context of a mechanism for ad pricing, google is the web equivalent of a TV ad about TV ads, which doesn't make any sense for a NetRatings rating. For that matter, what's the NetRatings measure of http://www.nielsen-netratings.com/ [nielsen-netratings.com] ?

Methinks that this announcement of a change in metric is just an attempt to get some profile on NetRatings' existence, and the notion of affecting google.com's measure for ads is plain absurd, because google *is* the advertiser. Drawing an equivalency between an indexing and search discovery mechanism like google and a less meta-focused content site is just boneheaded.

A bit of a lame submission IMHO.

Re:Hurt Google?

[Anonymous Coward]

In all its years as the number one search engine, Google has never run an ad on TV. I don't think that they are worried that they are going to lose business because of a change in a random metric. Google provides advertisers with customized information on how many well their own ads are doing, and I doubt that most of the advertisers even know that there are other companies ranking these websites.

Re:Love the headline

[MillionthMonkey]

This is the most hilariously worded Slashdot headline I've seen in like a week. The guy was basically describing an algorithm, nothing more. For technical reasons page view statistics are becoming irrelevant- so now they calculate a new metric that supposedly gives weight to longer user session lifetimes. Maybe they just pay more attention to overall HTTP query traffic or something. The effect of this would be, say, to boost a site such as AOL chat (an extreme example of a site with a low page view count and long session lifetime), and de-emphasize a site such as Google (an extreme example of a site with a high page view count and short session lifetime). For purposes of illustration, he just picked two examples that would make sense to people.

The article submission takes the angle that this is a kick in the nuts for Google! As if Google depends on Nielsen's reporting high metrics to advertisers so that they can charge more for banner ads! So Nielsen would report a low metric for Google! Oooh, what intrigue! Nielsen has their balls in a sling now! How will Google retaliate?

But that wasn't the point the guy was making at all; for him Google was just a good example of an extreme example. I would guess that nobody in either company is really concerned about Nielsen's calculated metric for Google. Google acts as its own Nielsen and competes with Nielsen using a not-quite-equivalent business model. It's a sort of integrated content provider/ratings company all on its own. They don't need to have their metrics reported to advertisers. Advertisers are showing up with money already for that AdSense program, and the cost is associated with a metric calculated for a search term, not Google as a content provider itself. The advertiser has already chosen Google (as the content provider) so implicitly of course they also have to agree to the terms of Google's ratings service since it's part of the package. Nielsen's rating of the Google home page doesn't enter into it! Just ask anyone using AdSense if they gave a crap about Google's Nielsen rating.

So what?

[Greg Lindahl]

Google is a company that doesn't care about ratings -- they get paid for clicks, not time spent ignoring the banner ads. What a dumb story.

-- greg

Re:Seriously retarded method

[Torsoboy]

You bring up some good points for a majority of ads, but I'd like to add something. You are basically saying that ads are only useful if they result in a direct purchase. The thing about ads... It's not whether or not it leads to a direct purchase or not... It's the fact that you know the product exists, and is mainstream. For example, you can't buy food from McDonald's online (yet), so using your logic, an internet ad for their product would be useless. It's more of a long term customer base they are building. You are familiar with their product because you see it everywhere, whether or not you like it or not. After a commercial, you may not go, "Oh, I should go buy a burger right now!" but maybe next week, when you have to choose between "Lil' Tony's Big Burger" or "McDonalds", you might (and probably will) choose McDonalds.

More likely Google doesn't give a shit

[Moraelin]

More likely Google doesn't give a shit whether or not some Web 2.0 metric gives it a pat on the shoulder or not. Google:

1. Makes its money out of serving ads, not out of being the site where you spend an hour on the same page. If you came, searched and looked at their ads, that's it.

2. Google's secret sauce is the brand name and search algorithm, not its Nielsen rating. People go to Google because they have something to search, and gets new users by word of mouth and by the deals it has with the likes of Mozilla to make their site the default home page. It's not like users start with Nielsen's Top X page and find out about Google there.

In other words, it seems... surrealistic to read the title and summary that Nielsen's ratings will hurt Google. Google doesn't get any income or users out of what rating it has, so the amount of "hurt" will be anywhere between insignificant and none whatsoever.

3. It seems to me like a flawed rating anyway, _especially_ coming from a usability expert. Google's search is a tool. Being able to just do what you needed done, quickly and with a minimum of useless fluff, is what a lot of us would call a good tool.

And the need for such tools won't go away just because some other sites work in a different way. Just because Ebay existed (as an example of a site where users spend a lot of time in a row), didn't make Google obsolete before, so why would it now?

4. Why the heck does it even matter, other than techno-fetishism, in Google's case, whether it's page refreshes or some AJAX kind of thing that fetches the results in the same page? No, seriously. Each search produces a different list, so essentially it _is_ a different "document". The browser is already perfectly able to display a new document. Why would anyone sane want to try to, basically, reinvent the page refresh in Javascript instead of using the browser's existing mechanism? No, seriously.

AJAX and the like make sense when you can actually have most of the data and the processing client-side, and you can actually offer some purely client-side functionality. In Google's case that's not even possible. You can't transfer the whole search database to the browser as XML and let the user tinker with the search expression locally, in the same document. So it's going to involve a round trip to the server and displaying a new result list anyway. So why not just let the browser display the new page?

Nielsen is generally a smart guy, but maybe there is no One True Metric to bring them all and in the darkness bind them. For some things it is a usability advantage to do more client side and not refresh the page, while for other things it makes no sense whatsoever. The focus should be on how well and intuitively is the user served by the site, not on promoting one arbitrary metric like time spent, taken out of context, for everything.

Exactly

[Moraelin]

Exactly. Maybe the fact that a perfectly usable and popular tool scores badly on their new metrics, _and_ there's no imaginable gain whatsoever if it changed itself needlessly to fit the new metric, should only tell them that their new metric needs some more work.

Re:More likely Google doesn't give a shit

[poot_rootbeer]

Why would anyone sane want to try to, basically, reinvent the page refresh in Javascript instead of using the browser's existing mechanism? No, seriously.

Why?

Because on a typical website, half of the page content is exactly the same on every single page: logo, header, footer, navigation rail, etc. The content well is the only part that's different from page to page.

Why should the client and server request and return those page elements over and over again if they never change? AJAX allows only those elements that are different from page to page -- the real content -- to have to be transferred.

Imagine how horrible Google Maps would be if you had to wait two seconds for the whole page to refresh every time you dragged the map a quarter of an inch.