Has the 10-pack become a 7-pack?
On Thursday, October 8th, something new started showing up on Google’s local blended SERPs:
What previously showed ten local listings now shows seven! Reports are that the 7-pack is showing for local queries in Canada and the United Kingdom.
It’s unclear whether this is a permanent change or simply something Google is playing with, but the fact that I am having trouble finding any 10-packs now suggests the former.
Some say that Google may be making space for its coming Local Listing Ads, while others believe Google is just cleaning up its interface, but the impact on those of us in the Local SEO game is substantial. A reduction in top page listings of 30% means we all need to work a bit harder to get the results we want.
Rankings in Google Maps Dancing All Over
The last few days have been exciting for business listed in Google Maps. And, by “exciting”, I mean irritating has hell as I’ve watched some longtime 10-pack rankings for clients drop away.
A visit to Google Maps Help shows that my clients are not alone. The wailing and gnashing of teeth was constant. Clearly there were (are?) problems across the US and Canada.
And so the conspiracy theories began. Many thought it was connected to edits made in the Local Business Center. Others thought it was algorithm changes. Having multiple businesses at the same address was thrown out as a possibility.
None of these theories matched what I was seeing. I saw drops in competitive markets and in non-competitive markets. I saw some clients stay exactly where they were.
What I did notice is that the “Pages” tab in the detail for a company’s LBL seemed to have changed. In some cases, the number of pages had dropped; in others, the tab had disappeared completely.
There were reports of Reviews disappearing as well, but I did not see this happen to any of my own clients.
My instincts were telling me to wait it out, but I’m not a patient man. So, I emailed Mike Blumenthal, who has probably forgotten more about Google Maps than I know.
Mike had two theories, both speculative and unproven:
1) Switch to (Google) Places has caused disruption with associating businesses with underlying cluster data thus leading to loss of reviews, and web pages and attendant loss of rank.
2) They have implemented some sort of more draconian penalty system for keyword or business title stuffing.
Of those, the first seems more likely, and sounds like the sort of mistake Google might make (especially within the Google Maps system).
If it was an anti-spam smack, it was a rather ineffectual one. Mike says he hasn’t eliminated it as a possibility, even if the odds are slight.
Surprisingly, after the conversation with Mike, the listings for my clients began to come back. As a result, I really believe it was some sort of data error within Google Maps, and that they are in the process of fixing it.
I would not be surprised, however, to see continued volatility in the rankings for the next few days.
Google Maps Still Spammed for Locksmiths
Two months ago, we discussed the abundance of spam listings in Google Maps for “locksmith” keywords. Indications at that time where that Google was attempting to crack down on these spammers, so I decided to take a look and see how they were coming along.
I looked at the listings for the six largest cities in Texas: Houston, Dallas, San Antonio, Austin, El Paso and Fort Worth. All have populations north of half a million people.
For each of these, I searched on the city name plus “locksmith”. So, for Houston, I searched on [houston locksmith]. For each Google Maps listing returned on the standard Google results page, I recorded:
- Display Name
- Display URL
- Phone Number
- Actual Name from Site Returned when the Listing is Clicked
- Actual URL from Site Returned when the Listing is Clicked
In addition, I recorded several subjective judgments for each listing and whether the site had duplicate listings within the same query.
All told, I looked at 53 listings. All the queries returned a 10-pack of map listings, except for Houston, which returned a 3-pack. Given Houston’s status as the largest city in Texas and the fourth largest in the United States, it was surprising that only a 3-pack was returned. Perhaps this is a clumsy attempt at spam reduction by Google.
Subjective Factor One
Is the listing for a local company?
Many (most?) of the companies advertising in the locksmith industry do not employ any actual locksmiths. They are simply lead aggregators that collect your info and sell it to a locksmith in your area. Hopefully that locksmith sent out is licensed and trustworthy, but there is certainly no guarantee.
Unfortunately, only 35.85% of the listings surveyed actually reflected a local company. Nearly two out of three were clearly just pretending to be located in that geography.
Subjective Factor Two
Does the display name match the website name?
I was relatively liberal on declaring a match. If the company had added a keyword, but the actual company name was recognizable, it counted as a match. In addition, a keyword-only name that used the same keyword name on the website counted as a match.
Only 60.47% of the time did the names match. This is something that Google clearly needs to address, but it will be difficult to do so in an automated manner. And Google hates to do things by hand.
If the listing did not link to a website, it was not counted in this metric.
Subjective Factor Three
Does the display URL match the website URL?
One of the standard Google Maps spam tricks is to include a keyword URL in the display and have it redirect to your actual branded site. The hope is that the keyword URL would count as another instance of usage of the keyword.
Google does seem to be cracking down on this, as they matched 90.70% of the time. On the other hand, this should be very easy for Google to check. Simply have Googlebot crawl the link, and if there’s a redirect, shut them down. I’m not sure why Google is not taking a more aggressive stance on this obvious spam.
If the listing did not link to a website, it was not counted in this metric.
Subjective Factor Four
Is the listing name keyword spam?
This was an assessment of how much the name of the listing on Google Maps appears to be just keyword usage. 47.17% of the listings had names that were nothing but keyword spam.
Again, this is hard to police in an automated manner, but the accuracy of a listing’s name is critical to the integrity of the system.
Subjective Factor Five
Is the listing spam?
This was an assessment of whether the listing itself is spam. Over-usage of keywords and misrepresentation of location where the most common cause of a listing being judged as spam.
Unfortunately, 64.71% of the listings for “locksmith” in the six largest Texas cities are spam. In other words, barely a third of these listings can be trusted. This is an unacceptable quality level.
One company came up for both Austin and San Antonio. They appear to be a company licensed as a locksmith in Texas, although they are servicing multiple geographies. It’s not clear where they are actually located, however. I did not count them in the spam numbers as they are an unusual case.
Six of the listings examined were duplicates within the same geography. In other words, the same company was taking up multiple listings within the same search. There was one match pair in Dallas and two matched pairs in Houston. This is another sort of spam that should be very easy for Google to detect.
By city, here are the number of spam listings:
- Austin: 7 spam, 2 non-spam, 1 unclear
- Dallas: 7 spam, 3 non-spam
- El Paso: 1 spam, 9 non-spam
- Fort Worth: 9 spam, 1 non-spam
- Houston: 2 spam, 1 non-spam
- San Antonio: 7 spam, 2 non-spam, 1 unclear
For the keyword “locksmith”, at least, cattletown Fort Worth has the most can pork by-product in the state of Texas. And El Paso has the cleanest listings.
ionadas local at Interactive Strategies 2009
Interactive Strategies 2009 is a full day conference sponsored by the Houston Interactive Marketing Association and includes tracks on Social Media, SEO, Creative and Strategy. The keynote is by Brian Solis, the principal of FutureWorks, a PR and New Media agency in Silicon Valley.
Attendees can register online and receive a 15% discount by using the promotional code “bcombs”.
Anatomy of the Implied Local SERP
While 40% of searches contain explicit local intent, geographic intent can be inferred for countless more queries. For many of these queries, Google includes its Google Maps listings for the locality of the searcher automatically.
I was curious about the structure of Search Engine Results Pages (SERP) for queries that imply local intent, so I spent part of yesterday studying them.
To build a profile of implied local SERPs, I looked at over one hundred such queries and recorded pertinent information such as where the map was listed, what sort of map was listed (10-page versus 3-pack), and whether any other vertical search results were ranked above the map.
Only two of the queries were three words long. The rest were almost evenly split between one word and two words queries.
All queries were performed on August 18, 2009 and were made on a Time Warner Cable internet connection in Cedar Park, Texas. Geolocation services consistently resolve this connection as being in Austin, Texas. Screenshots of the SERP were recorded for each query. They were also taken for the query plus “austin”, so that comparisons could be made.
For an example of an implied local SERP, see this screenshot taken for [grocery store].
The natural location of the Google Maps listing on an implied local SERP is the fourth spot, but this can be pushed down by an additive vertical listing above it. In one case, the third organic listing had an indented listing after it, which pushed the Google Maps listing down to fifth. Otherwise, if there were no vertical listings above the Google Maps listing, it was always fourth.
There were four types of vertical listings that could be ranked above the Google Maps listing.
Type Rate of Occurrence News 42.86% Video 11.43% Image 3.81% Scholar 0.95%
The distribution by rank of the Google Maps listing was as follows:
Placement Percentage 4th 50.48% 5th 40.95% 6th 7.62% 7th 0.95%
Placement in the 6th or 7th spot happened when two or three vertical search listing types out-ranked the Google Maps listing for a particular query.
The 10-pack was shown 92.38% of the time, while the 3-pack was shown 7.62%. Ie did not see the 1-pack on an implied local SERP.
For 12.38% of the time, Google placed the designation “Customized for Austin metro area, US” at the bottom of the SERP. This indicates that Google is blending pages that score for Austin within the organic listings. This is equivalent to adding “Austin” to the query for those organic listings. It did not change the ranking for the Google Maps listing.
In nearly all cases, the Google Maps listing on the implied local SERP was identical to the Google Maps listing on its companion explicitly local SERP. The only exceptions were the occasional authoritative 1-pack on the explicitly local SERP, which occurred when the company name matched the query exactly.
If a particular query returns an implied local SERP, that increases the importance of optimizing for its local variant within Google Maps. Google has not released and data on the ratio of traffic from queries with implied locality versus those with explicit locality, but it’s reasonable to assume it is non-trivial.
Implied locality greatly complicates things for national brands. Many of these companies are used to dominating the SERPs for their queries, and the addition of local listings across all localities (assumably) results in their having to fight hundreds of little battles. Most large companies are not structured to compete in this manner.
Concordantly, it’s a great opportunity for the local business!
One unanswered question is how implied local SERPs differ by the location of the searcher. I’m especially interested in comparing whether a 10-pack or a 3-pack is returned. Is the selection driven by the nature of the query itself, or the nature of the local listings for a particular geography?
If anyone is interested in helping me by performing some queries (and taking screen shots) in other cites, please let me know.