Understanding Google Maps & Local Search

Yesterday, Google’s Joel H has definitively confirmed that, at least in the whole of the US, Canada & the UK, they are no longer showing the 7 Pack on queries for Web Designers, Graphic Design, SEOs or Advertising Agents when paired with the City & ST.

According to Google, the searches lack sufficient local intent and their being missing is by design not by mistake. What strikes an odd chord for me though is that these very same queries, when paired with cities in Australia, China, Russia, Mexico and Hong Kong, do return the 7-Pack.

Today, we’re intentionally showing less local results for web design / SEO queries. For example, [web design sacramento] doesn’t display local listings today. We believe this is an accurate representation of user intent. In some cases, we do show local listings, however (as NSNA/php-er noted) [web design in bellingham]. I’m sure some of you feel we should be displaying local results for queries like [Web Design Vancouver]. I understand that concern, but based on our understanding of our users, we feel this is the right decision for now.

The issue has been in evidence since early November and had been noted in both the map and webmaster forums (here, here, here & here) as well as in a previous post on this blog.

On November 11, when Google initially responded publicly to the issue, they indicated that the missing 7 Paks was the flip side of the OneBox frequency problem that they had just fixed. Google employee Brian B posted:

This looks like it’s closely related to the issue going on at the thread I’ve linked below. We realize there’s something going on here, and we initially pushed out a fix a while back. There was a little hiccup with the fix, which is probably why the results in Fresno may have gone back up and then back down as addoctane mentioned above.

The team is working on this issue. Stay tuned to the thread below where I will post an update as soon as I hear one.

However, on 12/14 Google Employee Brian noted in a different thread than above:

That said, it looks like a lot of the recent search examples on this thread have to do with web design, SEO, and other services of that sort. It’s possible that instead of being related to the original issue, Google.com doesn’t bring up map results for these types of searches because the search term doesn’t show much local intent. In other words, there’s not enough local information in the query for Google to trigger maps results on Google.com. Searches for the same terms on maps.google.com, however, do bring up results.

The situation brings up more than one question…
(more…)

Google has recently added a “Report A Problem” Link for reporting business Listing spam & errors from within Maps. The link is available via the business listing info bubble where once you select “Edit” you are presented with a “Report A Problem” link that takes you to a short form.

According to Google Guide Cecelia this “is the best way to report spam because it gives us the most information possible. Our team can see which specific listing has an issue, whereas the form only asks for a URL. Sometimes these URLs are broken or people forget to add them”.

The current mapspam reporting form is still available but apparently, over time this new feedback mechanism will replace the mapspam reporting form in an effort to bring “the report closer to the product”.

The feature is mostly available in all countries which have the community edit feature. Although not all countries so noted in the chart as having community edits have the report a problem feature. In Canada is available on main Map listing view near the bottom and is only available on those listings that have already been claimed. Additionally, in the US only, a user report can be filed against an already claimed listing.

Maps Guide Cecelia noted that the “time frame [for spam removal] most likely varies based on the number of reports we receive”. Spam removal has been a point of frustration for many SMBs as Google has been more likely to use the information for algo tweaks than to remove the offending spam. This is a practice that often leaves egregious spam in place for months on end with no indication when if ever action would be taken.

2009 has been an exciting year of growth for local and for my blog. To all of those many folks world wide that have regularly contributed to the conversation and my learning I say THANK YOU!

Here are the 10 Most Read Stories of 2009 (in order of total pageviews):

1. Google replaces TeleAtlas data in the US with Google data
2. Tracking Local Search Traffic with Analytics (Thanks to Martijn Beijk!)
3. Google Maps Proves more Locksmiths in NYC than Cabbies
4. Where are Google Places Pages going? To the Index?
5. Google Maps vs Locksmiths Spammers – Spammers winning?
6. Big boobs bounce back to Top of Google Maps
7. Google Maps now showing Local 10 pack on Broad non Geo phrase searches
8. Google Maps – How to Remove Duplicate Records in the Local Business Center
9. Google Maps – Merging Mania Due to Algo Change
10. Why the Google Local Business Center Fails

The 5 Most Read Stories from the Back Library:

1. Ranking Factors in Google Maps – Cracking the Code @ SMX Local
2. Google Maps still loading slow – try html
3. List of Google Maps Categories
4. Microsoft’s listing in Google Maps Hijacked – (Oops by Me)
5. Google Maps New Local 10 pack Now Live

Greg Sterling posted a Hitwise list of the Top 10 Sites for 2009. It showed ebay.com at the number 10 spot. According to Compete, Google Maps is approaching Ebay and if it can sustain its growth, might even pass them in the coming year even without the Yelp volume.

Barry Schwartz has pointed out that last week, Google Maps added a chart showing which Google Maps services were available in which country. Maps is a global product that is offered in 200 some odd countries and there are a range of features available depending on licensing and data availability. Thus the matrix is incredibly complex.

From this graphic display of the variation in tools one can infer some of the support, licensing and development difficulties in providing a mapping tool world wide. Obviously in some countries there is no digital data of the roads and no lists for the businesses. If there is, the use of the data may be constrained by licensing issues. Here is a brief summary of the features available in a range of developed and 3rd world countries:

The year has been an interesting one. The recent conversation of a possible acquisition of Yelp by Google motivated me to create this chart comparing traffic for the larger sites in the Local arena.

The year started with Craig’s List and Mapquest holding a small lead over Google Maps. Citysearch held a solid fourth and Yelp rapidly bringing up the rear. But the rapid growth of Maps and Yelp, the steady decline of Mapquest and CitySearch has created a significant change in standings:

There are a million ways to look at who has done what in local and this is but one. Here is the URL for the Compete charts. Head on over and add your own favorite local site to see how it has fared and let me know.

When you look at these numbers, it makes the now off Yelp-Google combo look like a market dominating matchup.

There are faux pas and there are FAUX PAS. Google Maps by virtue of its global scope and the occasional tin ear of its algo has managed to periodically insert itself into the culture wars of 21st Century. Google Maps has been accused of Chinese bias by the Indians and Indian bias by the Chinese and of course the real battles of Jerusalem often take place in the virtual world of Maps. However, Google may be trying its luck by taking on the French in both matters of food and language.

Reader Keonda has pointed out on my post about Google Maps newly instituted sentiment analysis feature that Google is now showing the more granular review information in Europe. However, apparently it is showing up on the localized Google language sites in English.

The French take their country, language and food very seriously. By showing sentiment analysis on Google.fr of French food in English, Google may have just stepped into one very huge aggregation of merde. It would seem to me a sure fire way to anger everybody from Sarcozy to the subway conductor. And trust me, you don’t want to anger the subway conductors in France.

Note on this example search for the Lasserre Restuarant in Paris:

Sentiment analysis has to be a very complicated algo and it would seem to be very culturally and linguistically based. There are likely to be even regional differences in expression. For example I can’t imagine how the review for the Medical Marijuana Provider in LA would translate (Hey dude, that was some good shit!). Perhaps for now, Google has just figured it out in English.

If that is the case then perhaps it is premature to be showing it on the regional sites. Unless of course they are prepared to do battle with Le Monde and the powerful French unions.

Last night, reports started circulating (see TechCrunch who broke the story as well as Greg Sterling at SEL & Scoble for analysis) that Google was in late stage conversations to acquire Yelp for $500 million. It is deal that, whether it occurs or not, will have huge impact on the makeup of the local landscape. With Yelp in play, additional consolidation in the space is likely to occur rapidly.

I have no idea whether this deal will go through but I have been an active user of both products and can understand why it is in Google’s and Yelp’s interest that it would. I am more intrigued though by the user experience and what will become of that if a merger does take place.

Since June when I purchased an iPhone, I have travelled without a laptop, using the iPhone exclusively as both computer and to navigate the unfamiliar local environments. Over that time I have experimented with a great number of local apps but in the end have always returned to and continue to use two: Google Maps and Yelp.

I would use Google Maps mostly for the “recovery” process of local navigation; what is the address of Kossars’s? Where is the hotel in relation to the subway exit? What is the quickest way to get 8th Avenue and 14th St? In large urban areas at least, it provides incredibly accurate AND useful results.

Yelp on the other hand, I would use for “discovery”. What tasty ethnic restaurant would meet our budget? Where can we get Ethiopian food? What is cheap and near the hotel? In fact, upon arriving in NY and meeting my “nephew” near my hotel for dinner we had both identified the same restaurant as a likely choice. He by asking his college roommate and me by asking Yelp.

Google clearly had the recovery process nailed and Yelp had a well developed discovery process. It struck me at the time that what was missing in the local space was the ability for the Yelp like discovery process to take place across all local businesses not just restaurants or hotels.

What did Yelp have that Google didn’t?
- More granular data about pricing, dishes and other specifics in the restaurant arena
- A greater quantity of passionate reviews and reviewers
- A more faceted search process

I had always thought that it was the reviews and business details that made Yelp so useful. But in the end, the ability to quickly and easily direct the program to narrow the choices of what met my needs, made Yelp really work for me. This is, at least in a limited way, faceted search. Where the user, by working a little harder at giving the machine more information, can get better results. Some parts of the structured search were provided automatically but all were provided seamlessly.

With Yelp, the process of selecting a restaurant, particularly on the iPhone, allows the user to narrow down the data set to a meaningful number of useful results. What neighborhood, what price range, what quality range are user inputs that quickly allow the service to rank and present a list of restaurants that is satisfying. Yelp really seemed to want to know what I, the user wanted.

Google, on the other hand, has always treated search as a commodity and the user as a simpleton. It has always taken the single field, brute force, we’ll make a great guess at what the user wants approach to providing the answer. With its many algorithms and huge amount of available horse power, this has worked well for web search. It, does however, have severe limitations in local search where ground truth is the ultimate measure of success, not relevance.

But for a structured search like Yelp’s to work and for Yelp to keep growing, they need data and lots of it. That is something that Google has plenty of in local.

So while I think Google would be acquiring a great number of valuable assets if they were to acquire Yelp (a large social network, 8.5 million reviews, a local sales force, a rapidly growing audience), it could also be acquiring a new way of looking at search in the local space. One that clearly works very well.

Since October and Google’s inclusion of the “report an error” link, I have been obsessively reporting errors in the general area of where I work. As I noted yesterday, Google has acknowledged and begun to repair these.

None the less, a few errors remain in Google’s geo data and the business listings so I decided to bring you a custom Map’s (mis)guided tour of downtown Olean. The view covers a roughly 55,000 sq. ft. area and the list is by no means exhaustive. If you are planning a trip in this direction, better give me a call.

Olean, a town of 15,000 where I was born, has lost a number of businesses every recession since the onset of the rust belt phenomena. However, it appears that more than the usual number have been misplaced if not lost during this recession. Thank god Jim’s Vacuum Repair is still around and I know how to get there. As much as I dislike dentists, I am glad that I can still find my dentist as well.

The (mis) Guided Map (Pin=Google’s Location, #=actual location)
(click to view larger):

—

The Key (with the real scoop):

Actual Location – Maps Link	Info Bubble	Notes
A		The centroid of the fair City of Olean, NY at the southerly end of the downtown area
1		My Dentist. Shown on East State Stree. He is actually located 4 blocks west on West State Street. Seems that East State Street is misnamed for some retail locations only.
2		My lawyer. I get a lot of speeding tickets so its a good thing she is in the next office over not 3000 feet away. (Just kidding about the tickets )
3		My office supply store. This is a weird one. The address shows the next town south of but is still located in this one. The pin shows it in north Olean and not on N Union St.
4		Not my bar but a well known watering hole none the less. It seems that Google shows two N Union Streets. One is actually called North Union St Extension. The Pulaski Club is on the real North Union St but is shown on the extension. I suppose there are enough other beer places that this will not cause any undue hardships.
5		It seemed fitting to include a chiropractor in these results. You will need one after going through the many gyrations following the wrong turns on this map.
6		The city skating rink were my son spends his Friday nights during the winter. The pin is actually located where the city office for the rec dept is, so while way off at least I can understand it.
7		My vacumm repair man, Jim. I went to high school with him and having a vacuum repair man is one of life’s treasures.Seems the same West State-East St bug that bit my dentist bit him.
8		Colonial Library eh? Now this one is very weird. Actually (according to Maps) its located in Richburg NY, a town 30 minutes east not near the Hampton Inn as shown.
9		The map shows 2 N. Union Streets. No wonder we can’t find the Pulawski Club
10		W. State, East Sate…what’s the diff? I think I’ll stop at 10.

My test* of Google’s ability to correct and update geospatial data is in. They passed, but just barely. While a passing grade may suffice in French class, I am not sure that it is sufficient in this case.

Google’s long term business plans are predicated on the ability to deliver local ads to mobile users. For that to happen accurately, Google needs to know not only about the existence of a business but where it is located on the globe. To gain control over this critical underlying information, Google, in early October, replaced TeleAtlas’s road and street data with their own. As part of that upgrade Google promised that reports of Map errors would be fixed in 30 days.

On October 30th, I reported the fact to Google that the address 201 N Union St. was located roughly 3000′ to the north of its actual location. This resulted in the 30 some odd offices located in the building (with the exception of mine which had been changed in the LBC) all showing as being located at the other end of town. On November 3, Google acknowledged the accuracy of my claim with a timely email. Corrections appeared to moving along as planned.

The good news? Yesterday, at 1:50 pm, Google reported that the reported issue has been fixed. The bad news? While the street address resolves correctly in Maps, none of the businesses yet do. This is likely an indexing alignment issue and will resolve itself in another week or two. Does it matter to the user getting bad driving directions? Not a bit, they will be angered regardless of the technical explanation.

My grade on their efforts? A C+ or maybe a -B. Much better than TeleAtlas ever did by a long shot and with reasonable user feedback. That being said here is why I am downgrading them to a C.

Firstly, it has been 45 days, not the promised 30 to a fix. My father (a retail animal) always instructed us to under promise and over deliver. That should be a dictum that Google adopts as its own. If it is going to take 45 days then say 60 and folks will be surprised at how quickly it was done. Saying 30 and taking 45 on the other hand, just engenders scorn.

Secondly, finish the job before reporting out. Just because one index is updated, the whole problem needs to be solved before it is really solved.

Thirdly, and this is why the grade might be on the lower side, getting Maps right is the future of Google. Behave like your competitive lives depend on it, because they do! Say what you mean, mean what you say and execute.

Google bought into a huge maintenance and upkeep problem when they decided to replace TeleAtlas as their provider of underlying geodata. It was obviously perceived as a critical technology to bring in house and justly so. That being said, if they are going to do it, do it right. The market is a tough task master and I do not think anything less than an A+ will suffice to keep Google in a market leading position going forward. Any lesser result will mean failure.

*My sample size is one, arguably too small to make a judgement. I have more requests into Google which I will follow as well. Perhaps larger metro areas have been prioritized and fixed in the time offered.

My response: Even if one is this late, that is too many. The user base does not understand sampling and the time to completion stated should be the maximum time in any case.