NetRatings: Tiny Google Decline, But Not The First Time & Yahoo Growth

I looked at the latest figures out of comScore that showed Google seeing its
first drop in search share for nearly a year. My review of rating service
figures continues today with NetRatings. They also show a Google drop, but far
less than the comScore figures.

We’ll get to the NetRatings figures in a moment, but first a look back at comScore. I
said I had some follow-up questions for them. comScore kindly responded, and
you’ll find answers
postscripted here.

I poked a bit at comScore for spinning its figures in press releases to gain
attention. But exactly as Frank Barnako

, all the ratings services do it. Heck, newspapers do it, and we do it
ourselves on occasion, look for an angle that will attract attention to a story.

The main thing I’m trying to encourage with any of the figures you review
— or really anything you read — is to use that old "critical thinking" stuff
that my high school teachers used to pound into my head. Question everything and
wonder about what’s not being said or examined. Amr Awadallah

an example of this from earlier in the year.

As another example, we have the situation now, where Google saw a July 2006 decline in marketshare as reported by comScore.
Beginning of the end, or perhaps something you’d expect? As I noted,
month-to-month or month-to-year-ago figures might not be enough to tell the
story. As it turns out from the long-term comScore chart I showed
earlier, Google’s had declines before. That’s not in the
press release that went out, but comScore did highlight it in the
follow-up they
sent me:

We agree with your assessment that a single-month decline does not constitute
a trend. In fact, comScore also observed a similar seasonal decline for Google
during the same period last year. Fewer work days, more vacations, and reliance
on academia could all contribute to Google’s core user group showing lower
online activity and conducting fewer searches during this time period. Bottom
line – Google could be more impacted by seasonality than other engines

Now let’s head into NetRatings. I was going to do Hitwise today, but ever so
coincidentally, NetRatings put out its latest search share figures a few hours
after comScore released theirs. Funny how that happens, eh?

The release isn’t online yet, but you can watch for it via the NetRatings press page
here. There was no particular spin
to it, just a simple headline of:

Nielsen//NetRatings Announces July U.S. Search Share Rankings

FYI, NetRatings didn’t play down an angle in the wake of my comScore article. Go
through the archives, and you’ll see that this is how releases of search share
have been put out by NetRatings since May, all low key. Of course, go back a bit further and then you get into
releases that also pitch angles.

How about those figures? Here they are, all searches within the United

July 2006


Searches (Millions)

Searches Per Day (Millions)





















My Way




























Now let’s look at some trends. Pulling the chart from my now updated
NetRatings Search Engine Ratings
page, we get this

You can see the tiniest of drops for Google from June to July. For Yahoo, a
marked rise. MSN continues that slow, slow drop. Now for some drill downs.

One of the key things I said yesterday about dealing with search share figures is to watch
patterns over a long period of time and also look at what’s happening with
different services. If you only watched comScore, Google had nearly a year of
rises until July’s fall. Seeing a drop and only knowing about comScore figures might be alarming.

In contrast, the chart above shows that Google’s gone through several
declines more recently, at least according to NetRatings. In April 2006, it had
a peak of 49.8 percent of the search share in the US. The next month in May it
dropped nearly a full point to 49.1 percent. The following month in June, a
slight rise to 49.4 percent. Then in July, a slight drop to 49.2 percent.

Go back further to May 2005. Starting that month at 48.0 percent, Google
drops each month through September 2005 to 45.1 percent, nearly three full
points! Alarm bell time!

Perhaps at the time, some might have felt that way. But then you see the
traffic come back up. As comScore noted, it might be that Google gets hit more
by seasonal swings. If they have a heavier than normal school and university
audience, perhaps that group is searching less during the summertime in the
Northern Hemisphere.

When I’m looking at the figures, I’m watching to see if the line moves
through important bands. Again, see the NetRatings chart above. Every fifth
percent mark has a solid line. I’m less worried if Google goes up or down
between the 45 and 50 percent marks. I’m more interested if it breaks out of
that band in either direction for a long period of time.

That’s one reason why when comScore figures were ringing some alarms with
some analysts earlier this year, I felt pretty mellow about Yahoo. In the
comScore stats, it was staying within the 25 to 30 percent band it had been in
for several months. Moreover, the NetRatings figures had Yahoo pretty solid in
the 20 to 25 percent range.

In contrast, it’s really hard to be positive about MSN. comScore showed it
marching out of the 15 to 20 percent range back in October and into the 10 to 15
percent band. OK, Yahoo stepped
out of the 30 to 35 percent range around the same time. But NetRatings also
showed MSN doing a long dead-man-walking decline, unlike the case with Yahoo. The latest
NetRating figures have just taken it into the 5 to 10 percent range, from the 10
to 15 percent range it previously was in. Even more alarming is the fact that
this decline all comes after significant spending MSN has done to try and build
its search brand. That spending clearly hasn’t helped.

I’ll be watching over the next several months to see if the MSN decline
reverses. Of course, domain "roll-ups" might be a factor.

Every service "rolls-up" a
number of domains to create traffic figures. For example, the "Ask" figures from comScore will include plus some Ask-powered web sites. NetRatings, in
contrast, will count traffic separately than traffic to Ask-powered
sites such as iWon.

The MSN figures might not be counting some of the Windows Live
traffic that Microsoft is generating. Windows Live is the new leading brand, and
more and more traffic is being pushed there. I’m checking with NetRatings about

As a side note, I wish all the ratings services spelled-out exactly all the
domains that are used to measure the traffic to respective services. It would
make it much easier to know if they are neglecting anything important.

How about the other major players on the chart? You’ll see AOL has a big
spike in August 2005. I’m not sure what happened there, but it was probably a
change to counting methodology. I’m checking with NetRatings on this. But that
type of spike is another reason that taking two particular months and comparing
them out of context with surrounding months can be dangerous.

Finally, there’s, which keeps plugging away. The key thing to keep in
mind here, as I noted already, is that NetRatings measures for while
comScore measures for the entire Ask network.

Tomorrow, I’ll look at some Hitwise figures. Hitwise was actually the first
out this month with ratings from July. If you’re itching to see them, jump on

. Unlike comScore and NetRatings, they show Google making a gain from
June 2006 to July 2006. Like the other two services, they do agree on a Yahoo
gain between those months. Bill Tancer over there has also

just done
a weekly breakdown going into August that shows slight Yahoo and
MSN gains and slight Google losses.

Also, just a last note of encouragement to all the rating services. I like
them. I know they have problem and that the stats aren’t perfect. But they give
us all a starting place. The purpose of this series isn’t to knock the services
but rather help educate those who look at this figures on how to better assess

Related reading

A primer to forecasting the value of SEO
How to get started with Data-Driven Attribution in Google Analytics
Conversation Mapping