Search Engine Tigers

Just a fun little video with Stephen Colbert discussing the Yellow Pages industry. Enjoy

It’s well known that Google Maps gets data from InfoUSA and, more recently, from business owners going in and editing their information themselves, but are they also getting that data from crawling regular websites? The short answer is ‘yes’ and ‘maybe’.

The long answer is that I’ve seen listings for small businesses in Google Maps where the website displayed for the business is that of a ‘profile page’ on an IYP site. It’s highly unlikely that these small businesses that don’t have their own websites have located their profile pages on these IYP sites (in multiple states), and submitted them as their primary websites. I also know that the IYP site didn’t go in and change the data for these advertisers to point at the IYP pages as these are sites that I have primary SEM responsibility for, and I don’t have the resources to go and add the listing for each advertiser in each phone book to Google Maps. Now as to whether it grabs other data from the website, such as address and phone number, that’s more difficult to detect, as there has to be a known difference before-hand that you can observe for a change. The advertisers that are out there now with these IYP profile pages as the primary website have the same info as in the IYP, which may mean that they always did have, or that they did change unobserved. I don’t have data on this yet, but given that I do have sites out there with metered numbers I may one day be able to say that they do (of course, I would never be able to definitively say that they don’t).

For the majority of sites with the metered numbers I am using this tip from SearchEngineGuide, and having the name, address and phone number of the business listed on one line on the bottom of the pages. It’ll be interesting to follow those that have that, and those that don’t to see if it makes a difference in getting picked up in the local search results.

“Daddy, where do web pages come from?”
“Well, someone has an idea and they find a web developer that they like very much and…”
“No daddy, I already know that from the playground. What I want to know is what is S-E-O all about”
“Oh boy, I didn’t think I’d be having this discussion with you so soon…”

Tonight I asked my wife to describe what I do, and her description was “Um… You try to get traffic to sites, and then someone makes money”. Maybe my 5 year old understands what I do?… ok maybe not: “You work on a computer”. But what about customers? Do they know what it is they’re getting themselves in for? There are three types of customers that I’ve come across, those that don’t know what it is, but they want it and those that know enough to be dangerous (those that actually know enough to do it themselves aren’t necessarily in the customers bucket). For both sets of customers education is the key, understanding what they know, filling in the gaps for them, and setting realistic expectations are the best ways to go. Basically treat them like you’d want to be treated by a vendor. If you don’t it’s only going to come back and bite you in a soft tender spot at some point, and if that happens too often you might find yourself having to explain to your spouse exactly what your new job is… although maybe it’ll be an easier one for them to understand.

(thanks to SEOWife for the idea behind this post)

I’ve been looking at the difference between searching for a local company on the regular Google web search and on the Google Maps search to see if there is any kind of correlation between the two. For this test I decided to search on “Vasi’s Catering Haiku, HI“. Entering that as my first search, I was given a result page that had a local result at the top of the page (which points to their own website), and a nice list of iyp results, starting at city search, going through paradise pages, down into local.yahoo, before actually listing their website in the regular results.

Clicking on the ‘More information‘ link under the local result gives a different result set as the query has changed, it now includes a zip code, radius parameter, and lat and long parameters. The first result is their own website (to be expected since that was shown as the primary local result on the previous page), then it’s a New York cafe website, that lists nothing to do with Hawaii on the site, followed by the previously seen paradise pages result and new allpages and yellowbook listings.

Interestingly enough, clicking back on the ‘web’ results then displays a different set of results, as the new parameters added in in the last step are now passed back through to this search, which returns pretty much the same results as the first time, bar the paradise pages result which was showing in both of the other searches.

Staying in Hawaii (who wouldn’t?), the next search that I did was for “kihei plumbing kahului hi“. This gave me a local search result for Paradise Pages that was also the primary organic result. Following that came Switchboard, BuildingTradesDir, Superpages and CitySearch. Again, clicking the ‘more information‘ button changes the results. Paradise Pages retains pole position, but then Allpages and Kivanetext are the only other results associated with this search.

So what does this show? Basically that there’s not much of a relation between the regular organic web search and the local search on Google. Yes it’s a small sample shown here, but you can do the same searches with most businesses and you’ll see similar results. It’s also not as if the local results are somehow tied to the business in some special way, as in the first example there was a website with no relation to the business, and in the second example there was a listing to a competing plumber on Paradise Pages, as well as to the actual plumber.

One of the easiest places to get links to your site with a decent page rank is from your own profile pages on different sites. Maybe you signed up for MySpace to show your kids that you were hip, then never went back? Maybe you jumped on the recent MyBlogLog fad? Maybe as a professional you’re on LinkedIn? What about when you put those pictures up on Flickr? Maybe you’ve been a member of a forum or two that have profile pages? Maybe you signed up for an authentication service so you could post a comment on another persons blog? There are many different sites where you can have a profile page that allows you to have a link to your site / sites. Quite a few of these sites will generate a decent page rank, given the authority of the sites that they reside on. For instance, I had completely forgotten that I’d signed up for the TypeKey service, when I found my profile page it was at a PR of 4, not a bad little link to throw at myself. Of course, some of these sites either ‘no-follow‘ the profile pages or don’t allow links, making them just plain text, but in the case of the former you can still get traffic from a good site, and in the case of the latter, it can’t hurt.

So what should you do to take advantage of this? For those sites that you can remember signing up for, go out there and add the links to your site(s). For those that you can’t, do a search on your name or your nom de plume and see what you find, there may be some decent profile pages that you forgot all about that you can quickly and easily link to your site(s), and pass the ‘link love’ on to them.

Let’s start off with the disclaimers: I am a SEMPO member, and a member of the SEMPO In-House Committee. Now that that’s out of the way, on to the topic.

A week ago SEMPO launched their Fundamentals of SEO course, covering such topics as
History of Search Engines, How Search Engines Work, Linking & Directories, Keywords, SEO Web Design, Copywriting, SEO Roadblocks, Submissions, Tracking & Reporting, Pay Per Click I, Pay Per Click II, Current Innovations, Intellectual Property, and SEO & PPC Resources. The cost for the course is $499 for anyone, $399 for SEMPO members, and $250 for military or students (12 credits or more). I wanted to see exactly how good the course was to determine whether or not it was something that I could recommend for the rest of my team, so I signed up for it on Thursday.

On Saturday afternoon I started the course. Each topic has a lesson comprising of 35-70 pages of subject matter and mini-quizzes, and concludes with a test that requires a 74% grade to pass. You are not allowed to pass on to the next topic until you have completed the test for the presious topic. By Saturday evening I was done, with all of the tests having been completed. So what was it like? Well, having taken the Sun J2EE and the NSDJA (National Sash and Door Jobbers Association) professional examinations in the past, I would have to say that it was closer to the NSDJA exam. What does that mean? Well, the Sun exam involved studying the course material, then going to a testing center to actually take the exam, under exam conditions. The NSDJA course was open book, with the quizzes being mailed in for grading (this was 1996, it’s hopefully been updated by now). The SEMPO course was basically open book, with a button being pressed at the end of every exam, and if you failed the exam you could retake it. Also, unlike the Sun exam, the questions were the same every time you attempted the exam, so if you happened to mark down your answers, you could run back through and change those that you were not so sure about, and regrade them until you pass. Some of the exams used the same questions that were in the in-lesson quizzes, so if you had completed those, you were set for the test.

There were also some minor typo’s within the lessons themselves (I sent in a list of the ones that I noticed), but since this is new (I believe I was one of the first dozen to sign up for the course), that’s to be expected. The lessons are also slightly out of date already, as this industry moves rather quickly, discussions on Yahoo PPC in some sections talked about Panama, but not in a detailed way in all.

All in all, its not a bad start for the SEMPO institute. I’d like to see them do some randomization in the questions, and maybe implement and external testing / maximum amount of retakes policy. Then the certificate of completion will actually mean more. As for whether I’ll recommend that the rest of my team take it. I will, but probably in a month or so, when some of the kinks have been corrected. I’ll also be signing up for the advanced courses when they become available.

Update: Today (2/12) the certificate showed up at work.

It seems as though more and more companies are starting to move SEM in-house, and as they do so they run into the same kind of situations over and over again.

1) The Perception of Inferiority

This occurs when management falls into either the absolute ‘rocket science SEO’ belief, or just simply the belief that because consultants charge obscene amounts of money for their services they must know a lot more than the people who’d be willing to work for the wages that the in-house staff are drawing. Sometimes the only solution to this is to let it happen (not that you can really stop it if there’s an executive champion), and watch for if there is an opportunity to learn, and if not, the company will realize that they just blew money on a resource that they didn’t need to. Ways to mitigate this perception involve keeping up communication with the executive team, assuring them that the in-house team is up to speed on the latest developments in the industry, and making sure that they are aware of the latest results from the in-house team’s efforts.

2) The Short-term Development Vicious Cycle

Since SEO produces long-term benefits, there is sometimes a tendency within a company to push development resources to work on the items that will have a more immediate effect, thus pushing the SEO tasks out further and further. Of course, this is only an issue in an environment where the development and SEM functions are separated, which is not that uncommon. Again, education is the best way to assure that the SEM items are given a fair hearing and are placed in the development queue.

3) Aiming for the Moon with a Bottle Rocket

The belief that since SEO is ‘free’ and a ‘one time thing’ with maybe some ‘minor adjustments’ over time, there isn’t that much need for staff, and as for linking, well there are automated solutions that do that for you aren’t there? Without an adequate staffing level to support the sky high targets of the SEM effort for the site(s), the chance of success (or the definition of success) is limited. This can then lead into ‘The Perception of Inferiority’, as the full job can’t be done without the full resources.

4) Departmental PPC Wars

If there are multiple departments that are involved with similar products that have competing keywords in PPC campaigns, there’s a good chance that, unless the campaigns are centrally co-ordinated, the departments will be competing against each other, with possibly undesirable results (or rather definitely undesirable results as far as the company as a whole is concerned).

5) The Helpful Developer

‘Ooh, they used H2 tags on this page. Obviously they didn’t realize that you can create custom css tags. Let me go in and fix that for them with a class=”prettyText” tag and put it up in production.”. Changes such as this can occur if the development team isn’t kept notified and informed of the reasoning behind the way the site has been put together for SEO purposes. Of course, a more formal change process would also help to alleviate a problem such as this.

6) Jack of all trades

In small companies the Search Marketer may also be the sole developer, website designer, and chief bottle washer. This invariably leads to a lack of focus in one area or another as a day is still only 24 hours in length no matter how many hats a person wears.

7) Beware the Emperor

The director that was a fuedal chieftan in previous life may set up their department as their own private fiefdom, rejecting or obstructing any inter-departmental co-operation. Different job functions that could be handled by other departments are redundantly replicated in their department, so as to give them control over that function. This can lead to Department PPC wars, and requires finesse or more likely executive management oversight to overcome.

Do you have any more examples that I’ve missed? I’m sure there are quite a few. Also, if you’re a SEMPO member feel free to check out the in-house committee as we’re always looking for input from people with in-house experience.

Now that Panama has given us the ability to manage multiple campaigns within the same account, we’ve taken full advantage of it and gone with 20 campaigns per account wherever we can. When we pull our reports we run the URL performance, as each campaign targets a different URL. Today, however, we ran into a roadblock… about 1/2 of the accounts were displaying their URLs, the rest were displaying the text “DEFAULT URL”.

After contacting Yahoo, we were told that this is due to them not yet having committed the data for this weekend. It is supposed to clear up by tomorrow, and correctly display the URLs.

Update: “Default URL” is still there. After contacting Yahoo support one more time we’ve been told that they have escalated this issue to engineering, and they expect a fix in place by Thursday.

Update2: Fixed as of Wednesday morning, a day early. Way to go Yahoo engineering.

I was all ready to put together a post about the history of the wiki (from Ward Cunningham on down) and then today some news came out about wikis. Wikipedia has implemented ‘no follow’ on all links effective immediately. What this means is that the most popular wiki site is now not passing any ‘link love’ on to any external sites that are linked to. You can still get traffic from Wikipedia, but you’re no longer going to get the fringe benefits.

Why have they gone this way? Well, as SEOmoz also recently announced, there’s just too much link spam out there (SEOmoz are controlling the links on their profile pages, they’ve had the comment links no followed for a long while). People are going out to trusted sites putting links out there in the hope that they’ll get ‘link love’. This really is the fault of the search engines, they make it known that a site is trusted based on it’s links so people go out there to get links for their sites. With wikis giving anyone the ability to create their own links, it was only to be expected that people would take advantage of it, and try to get links everywhere they could get away with it.

So what will this mean for wikipedia? It’s still going to be the ‘go to’ site, based primarily on it’s popularity within Google, but it probably won’t get as much of a workout from people in the SEO community (which their editors will probably be happy with). Now there are other wikis out there (such as this one that I started using in ‘99 long before Wikipedia was on the scene) that haven’t yet implemented no-follow, but whether they’ll follow suit is something that remains to be seen.

Yesterday I listened in on an online webinar given by Microsoft. The two main topics were demographic targeting for Adcenter, and Adlabs tools. However, the webinar got off to a very slow start due to the problems that the presenters had with Microsoft Office Meeting Live… the meeting facilitator in Redmond was doing fine, but the presenters in NYC couldn’t see the slides (not to mention the application crashed on me about 3 or 4 times over the course of the hour).

The demographic targeting was interesting (I didn’t realize the data was based on Microsoft Passport user behavior), but the adlabs tools were more interesting. Here’s a look at 4 of the tools that are available for your use (in beta mode). Naturally all data is based on MSN search data.

The first is the Keyword Group Detection tool. The idea behind this tool is to give you related terms to the keyword that you enter. This seemed fairly buggy, looking at the image below you can see what was returned to me for the term “cookie bouquet” (in case you’re wondering, I stumbled on that term after it was recommended to me as a related term for pizza), not really much related to either cookies or bouquets, unless you get a free cookie bouquet with every new lexus…

The next tool that I played with was the Keyword mutation tool. This tool finds all common misspellings of the keyword that you typed. This is a useful tool, although like the others it does seem to be limited in the keywords that will return results. i.e. “Wolverine” had no significant mutations, however big G did.

The next tool is probably the most interesting of the lot. It’s the funnel tool. With this you can see what searches people either did prior to entering the keyword, or after entering the keyword. It’s handy in that you can see the evolution of a search, and identify high quality tail searches for your particular keywords, so you can see what the user intent tends to be for your particular keywords. It may not be what you think.

The last tool that I’m going to talk about here is the seasonality tool. In theory this sounds like a great tool, giving you estimations and actuals for search terms. Unfortunately it has a couple of big problems. The keywords that can be used to show results must be selected from a limited set of keywords. Then when you select one, it appears that the actual data stops in June 2005… 18 months ago…

The tools have potential, but they’re not there yet, so it was curious that they decided to have a webinar to push them out to the public, although it does have to be said that the problems with the in-house webinar software really set the stage for these tools that aren’t ready for primetime (hence their presence in adlabs rather than actually out in adcenter).