While various webmaster forums have lots of discussion about the recent Google changes, specifics about particular situations are often lacking. It's much more useful to look at real life examples rather than deal with generic situations. So, as people have been contacting me, I've asked permission to use their stories to illustrate in real life how the changes at Google have made impacts. Here are some stories. You can read them all by scrolling through the article or use the links below to jump directly to sections of interest. Also be sure to see other articles about the recent changes on the Florida Google Dance Resources page.
- My Attorney Is Missing
- We Could Build It, If They Came
- The Honeymoon Is Over
- Deoptimize Or Get Better Links?
- Listing Skate-Away
- A Sinking Feeling
- Directory Flipout
- The Cleveland Zoo Of Search Engine Optimization
Reader Michael Helfand discovered that his FindGreatLawyers.com site can no longer be found in the top results for "illinois lawyers," a position it had held for about three months.
"We follow the rules, try to trade links with relevant sites, have the same PR [PageRank meter score, as shown by the Google Toolbar” as before and have made no recent changes. This is killing other businesses and while it only puts a minor dent in mine, it's almost a pride thing about being rated number one in Google. Any thoughts?," Helfand wrote.
I looked over the site, which provides referrals to lawyers in Illinois. At first glance, nothing leapt out to make you think it was spammy or even overly optimized. Certainly it seems like a better first page choice than an affiliate site for a legal matching service that did survive the shakeout.
Down at the bottom of the page, I discovered a hidden link. Clicking on this led me to a page of reciprocal links. Now, there's nothing wrong with linking to other sites. Your visitors may want to know about them. But hiding links, that's one of the specific things Google warns not to do.
Reason? Some sites are swapping links between sites they own or with sites run by others solely in hopes of getting better search engine rankings. They hide these links to keep human visitors from seeing the links and possibly leaving their sites.
Google tries to assess a web site in the way a human visitor might. If certain links are not being shown to human visitor, then Google may not want to consider them either. In addition, Google sees such link swapping and hiding as an attempt to unduly influence its ranking system, thus the warning not do this.
As it turns out, Helfand says he no idea the hidden link existed.
"In a million years, I wouldn't know that text was there. I emailed my web designer to have her remove this. I think that a lot of small businesses like ours who rely on others for guidance probably get caught in this type of trap," he said.
Indeed, a big worry with search engine optimization has long been that people may hear or learn tips that they believe are fair to do, only to get accidentally caught for spamming by search engines. However, search engines have also historically been forgiving.
A single hidden link isn't the sign of a sophisticated attempt to influence rankings. But a hundred hidden links, along with perhaps some other aggressive tactics? Then you might have spam troubles.
I don't think the FindGreatLawyers home page got dropped in rank because of the hidden link. It was hardly enough, even when seen with other types of search engine optimization techniques on the page, to tip the page into spam territory. In addition, though the link has since been removed (and the PR score risen from 4 to 5), the page hasn't reappeared in the top results for "illinois lawyers."
What else could be keeping the page down? What the hidden link led to, that page of reciprocal links, may provide a clue.
This reciprocal links page was clearly created because people were offering to link to this site if it linked out to them. As a result, this legal site was linking to things such as a diet pills site or a skin care products store. Presumably, these sites were linking back in exchange -- reciprocating the links they received -- hence the "reciprocal link" name for this page and this type of linking in general.
The new system Google is using may no longer be counting these links as much for helping the page rank well. That's particularly the case if Google's doing a form of "LocalRank," which I explore in my Speculation On Google Changes article for Search Engine Watch members. Such links wouldn't be seen as relevant to the search at hand, "illinois lawyers," so having them may help less than in the past.
My feeling is that the site will need to build new, relevant links to make up for the loss of some old ones (and see here for tips). But this is just speculation.
For his part, Helfand buys into the idea many have that the changes were related to advertising on Google, but with a twist. Some believe they were dropped to force them to buy ads, a charge Google denies. Helfand has an opposite view. He was running ads and showing up for this term before the change, something he thinks hurt him now.
"I think that the theory that Google is making changes based on certain words is very valid. We used to be highly ranked for Illinois DUI lawyers. Now we are nowhere to be found. However, we are number two for Illinois DUI laws. I think words like attorneys and lawyers could be pushing Google to not show sites that do AdWords campaigns," he said. "If you look at who is ranking highly for words with attorneys or lawyers in the search, I don't think that link popularity has much to do with it anymore."
Of course, many of the other referral sites that were dropped don't appear to have been doing advertising. So what explains that?
It could be that Google made a change to catch certain types of pages, those deemed to be selling something, and is suppressing them in order to insert more informational pages on commercially-oriented queries. That could have the end result of what's being seen.
Such a move wouldn't contradict the denials Google has made about the recent changes having been done to sell more ads. There's good case that someone searching for "illinois lawyers" might like information in addition to lawyer referral site listings. That's what dominated Google's results before the change, with 9 of the 10 links being for referral sites. A change to decommercialize listings, or at least provide some different classes of information, can be argued as in the searcher's interest.
Certainly "new" Google shows more variety. The state's trial lawyers association is listed, as is the state bar association. Both seem like good additions. There's a link to a FindLaw.com page that itself provides some good links about legal issues for the state. Two referral sites have also survived. One's nothing but affiliate links to lawyer sites, but the other appears to be a good directory of lawyers.
The remaining links are incredibly poor. One is for the Law.com home page. That's a good site but a poor choice for this particular state-specific query. Another is an article from 2002 mentioning that some Illinois lawyers are urging that a death sentence be suspended in a particular case. Another article from 2001 urges Illinois lawyers to support a controversial figure seeking a law license.
Overall, it's nice to see that the changes have brought more variety and some informational sites into the top listings. Not everyone searching for this query will want a referral. To do that, obviously some of the referral sites that were in the top results had to go. However, the change also introduced some irrelevant sites, and one of the referral sites kept was hardly the best.
By the way, this is a good time to note that the observations about results in this article were all done based what I saw last Thursday through Saturday, when I started writing this piece. Results may have changed since then.
What's the situation outside of Google? Helfand's site is in the top results at Inktomi. He's also one of the lucky few referral services shown. Most of Inktomi's results are informational -- and at first glance, seem superior to Google. The state bar is there, as well as a lawyer finding service from the state bar. A "Questionable Illinois Attorneys" database seems a good catch, as do some other lawyer associations. So the Google change -- which knocked out a lot of commercial referral sites other Helfand's -- effectively is already the situation at Inktomi.
It's also interesting that Helfand's site is one of only two paid inclusion links making it into the top results. Inktomi and others that offer paid inclusion have flat out denied that their programs give pages a ranking boost. If Helfand were to stop paying, he might continue to be listed for free and might stay in the top results for free. It would depend on the inherent quality of his web page and links to it, they would say.
However, there's also been plenty of anecdotal evidence I've heard from people who feel paid inclusion has helped them. I've written articles in the past where I saw boosts happening related to paid inclusion. BusinessWeek recently ran its own investigation that found out of 20 advertisers, 10 believed paid inclusion gave a boost.
Even if boosting isn't happening, the point to take away when comparing results on Google competitors to "new" Google is that not all listings at its competitors are free. If you see a Google competitor displaying a lot of commercial sites for free, when new Google doesn't, it may actually be the case that some of those "free" listings involve payment.
At Teoma, there are also some potentially good informational pages shown. But there's also two particular law firms listed, which might not be as good as providing access to a good directory of many local legal firms. A national law firm, using a paid inclusion listing, makes it in to pitch that it can do local work. Overall, if Google hurt a number of "mom & pops," Teoma doesn't appeared to have helped them much to begin with.
At AllTheWeb, it's a mix. There's five individual law firms, one of which is using paid inclusion. There are two referral services. The state bar is present, as is the trial lawyers association. Word-wise, a link for the Lawyers Trust Fund in Illinois may make sense, but content-wise, it's not a good match.
Alistair Lees of OnlineBizBuilder operates a site that until recently was in the top ten for the incredibly competitive query of build website. After the change, he was out -- despite being an AdWords advertiser.
"We have made absolutely no changes at all to our website that might have caused this, so far as I can see - the text is the same, the URL is the same and there are no practices that are in any way underhand," Lees said.
Looking at the home page, nothing immediate appeared as a problem. But given how competitive that query is, you'd expect links to come into play far more than page content.
Unfortunately, Google's reverse link lookup tool is undependable. It won't show all the links it actually knows that point at a page. Google did find 13 internal and external links when I checked (it reported 22, but only 13 were displayed, and the ability to subtract internal links wouldn't work). At AllTheWeb, I got a more accurate count of 70 external links.
Most of these links seem to be in credits that are placed at the bottom of pages created by the OnlineBizBuilder tool. For instance, see the "Powered by onlinebizbuilder.com" link at the bottom of this page. It may be that in the new Google system, small, repetitive links like this, easy to spot because they follow the same pattern, are not being counted. That could have had an impact on the site.
What is surprising is that just having links pointing at you isn't enough. For a competitive term, you'd really want those links to be using the terms you want to be found for in or near the link. In fact, the current number one site for build website at Google shows something exactly like this. Many of the credit links pointing back at the number one site make use of its domain name, which have the search terms build website within it.
So the number one site has what I expected to find for Lees site -- and his site doesn't appear to have ever had links going for it. Why does Google not like it now? It's a mystery to me, as much as it is to Lees.
How about the results now at Google? Unfortunately, I didn't save a "before" look, and the filter test is no longer working on this query. While I can't compare to what's dropped, I can evaluate what's there now.
Six of the links are for sites that appear to help you learn to build web sites or offer free tools, rather than commercial web site building services. That also suggests that perhaps links to these pages from educational or non-profit web sites might count for more in Google's new system. This would be a good technique to use, if you wanted to decommercialize results to make them more informational.
It's also honestly hard to say that listing sites with website building services for this particular query, which I'm guessing was the case, is better than listing informational sites. The user might be better served by this change.
Of course, we do get some commercial sites listed. iVillage offers free web site hosting, while major hosting Homestead is present.
The seemingly inevitable link to an Amazon page is also present. Why it made the top results initially wasn't clear. A backlink lookup failed when I tried it at Google, as did my attempt to get backlinks via the Google Toolbar.
Very odd. It was suspicious for this page to do so well without backlinks. But the fault turned out to be mine. When I visited the page, the URL changed from this:
See the part in bold. Glenn Fleishman -- whose ISBN.nu site lets you compare prices from major online bookstores -- tells me that's an individual session code assigned to me and automatically added to the original URL. If someone else goes to the page, they'll get a different code.
Doh! It was a stupid mistake to make, but it illustrates how careful you need to be when analyzing things. A backlink lookup on the actual URL came up with 174 matches.
The Amazon page also points out something else that's important. Assuming all of its links are counted, the words in these links would often be the title of the book, "Building Accessible Websites." Yet the book is ranking well for a search that involves the root verb of building -- build -- and the word website in singular form, not plural.
Enter the change that Google quietly launched along with the new algorithm -- stemming. With stemming, Google automatically searches for variations of words such as plural and -ing forms.
Lees and others might think they searched for pages containing only the words "build website," as would have happened in the past at Google. Instead, now they are actually searching for words such as:
The expansion of terms is another reason why Lees's page might not do so well now. Suddenly, it faces much more competition. However, overriding stemming by using the +symbol in front of each word, +build +website, still doesn't bring Lees back. Stemming can't be blamed at the only thing that's knocked him out.
The result also have a perfect example for anyone who wants to razz Google about irrelevant listings -- a link to Habitat For Humanity. It's a great charity that builds homes for people. I've no doubt links to Habitat often contain words related to "build" and "website," since the organization is obviously involved with building and has a web site. But Habitat is clearly not the right result for this query.
FYI -- only a few hours after I've wrote this, the results for this query changed -- and the Amazon page disappeared. However, Habitat For Humanity was still there, last I checked.
How about Google competitors? Inktomi has a mix of informational and commercial web sites relevant to the query build website, Glancing quickly, it seems on par with Google. A link to "Businesses United in Investing, Lending and Development" seems odd at first, but the acronym of that organization is BUILD -- and so a search for "build website" makes sense for it to show up. It is the Build web site! It's more of a stretch for the Build Freedom web site to appear -- and the Build-A-Bear Workshop shouldn't be there at all.
AllTheWeb shows a leaning toward informational sites. I liked the AllTheWeb listings a little better simply because I recognize some of the sites shown, such as Dev Shed, Suite101 and Search Engine Watch's own sister-site, WebDeveloper. However, I couldn't say conclusively that the sites are better than Google's without visiting each one. I'd also need to know what type of user we're talking about. Someone brand new to building web sites? Then the Dev Shed site about using open source tools for web development might not be the best starting place to offer.
At Teoma, we see again a leaning toward information, though some commercial sites such as Homestead are there. We also get a glitch similar to the ones seen at Google and Inktomi, this time with the Open Directory coming up first.
Renee Duane runs Unforgettable Honeymoons, which she says was one of the web's first honeymoon-only travel sites and which has been featured in Newsweek, USA Today and the New York Daily News, among other publications. The company is both brick-and-mortar and on the web.
"We are probably one of the largest, if not the most well-known, honeymoon specific retail travel agencies in America," Duane said. "I am definitely feeling the pain and so are my loyal, hard working employees. Right before Christmas could not have been a worse time."
Google liked the site too, until the change. Then it disappeared from the first page position she says it has maintained at Google since 1998 for honeymoon packages.
Once again, nothing leaps out that seems wrong with the site itself. How about links? Could be a problem there. Google comes back with a count of 105. Most of the external ones seem to come from a site that lists wedding resources on a state-by-state basis.
Duane's site is listed on all these individual state pages, from what I can tell. That's appropriate, since she can serve any state. But in Google's new system, it might be that it recognizes that these links all come from the same site, in the same style, so perhaps it effectively collapses them down into being worth a single link. That would be helpful in preventing a site from receiving too much link credit, in a situation like this.
In contrast to Duane's site, the top listed site on Google has links from many different web sites. I spot checked a few and found the words "honeymoon packages" in close proximity to some of the links.
Ah ha! Well, maybe not. The site listed second, featuring Bali honeymoons, is like Duane. It seems to have many links from only a few web sites. I could also see that some of these links also weren't descriptive. Instead, they were links from the word "Reserved" in a copyright statement saying, "All Rights Reserved." Given this, it's hard to know why that site is getting top ranked. Links don't seem that helpful.
Indeed, why that site -- which offers honeymoon packages in only one location -- is better for searchers than Duane's site that offers them worldwide, is also unclear.
The argument that Google changed things to oust either small businesses or commercial sites overall doesn't hold up, in this particular case. Duane's small business disappeared for this query, true. However, all the other sites listed except Club Med (one of the new entrants) seem to be small businesses.
How about the idea of Google shifting over to show "informational" content. Not proved, in this case. The "before" listings were all sites selling honeymoon packages. The "after" listings may have some different sites, but they all sell packages, as well. Indeed, you'd be hard pressed to know why some sites in the "free" listings shouldn't be in the ad area and vice versa.
Competitor tour time. Inktomi's results might as well be Google, in that they simply list a bunch of sites selling honeymoons. But Duane's site does get listed in the top results, good news for her for when Yahoo begins to use Inktomi results in the near future and when MSN shifts over to them fully in mid-January. Hers is also the only paid inclusion link on that page.
As at Inktomi, so too at AllTheWeb, except that her site shows up and but isn't a paid inclusion listing there. Teoma is nearly the same, except Duane's site doesn't show there and a paid inclusion "article" from eLibrary about a particular beach resort stands out. I say article, because it's really just a press announcement for a particular resort offering a holiday package.
"I don't use spam at all. The only thing I can think of are the links that my boss exchanges with other sites," said Jones.
What happened? One thing that stands out are the site domain names:
If any search engine is looking for a sign that a site may be highly optimized, multi-hyphenated URLs are a sure sign.
Google has expressly said that it doesn't penalize for hyphenated URLs in the past and once again recently in public forums, so this may not make a difference. However, it could be that while hyphenated domain names alone aren't bad, using them and other aggressive SEO techniques could tip a site into being seen as spam.
At the air purifiers site, there are other signs that it was optimized for search engines. The page has a meta revisit tag on it, which no major search engine uses. But some people optimizing for search engines see this tag on other pages, assume it must be helpful and add it to their own. Thus, a search engine seeing the tag could interpret it as a sign of optimization, in case it wants to filter optimized pages.
The words "air purifiers" appear in the body copy twice in bold text, for no apparent reason. In the past, Google has said that it may consider words in bold to be a bit more relevant to a page. My guess is that this is why the words were placed into bold on this page -- to try and please Google.
The HTML title tag, a comment tag and the meta description and keywords tags all have the words "air purifiers" in them. That's a big sign that the page owner wants to rank well for that word.
It's possible that Google might now be scanning for signs like this. You search for "air purifiers," and Google gets back results, then runs down a checklist like this:
- Are terms in HTML title tag?
- Are terms in meta description tag?
- Are terms in meta keywords tag?
- Are terms in bold in body copy?
- Are terms in comment tags?
- Are terms in ALT tags?
- Are terms in a hyphenated URL?
- Is meta revisit tag used?
Match several of these items and others not listed, and you might find your page now dropped from the top results -- at least that's the thinking of some. Those believing this now think they need to "deoptimize" pages to do well in Google. Certainly to me, the last four items wouldn't hurt to reconsider, if they've only been done for search engine reasons. But the rest, if done appropriately, really shouldn't be necessary to change.
In fact, the air purifiers site uses the words "air purifiers" fairly appropriately. They aren't repeated billions of times in any of these places. OK, it does sound awkward to see that phrase used three times in the opening paragraph of the page. However, it only appears seven times overall on the entire page, in the visible body copy. This isn't a major spam problem. If it was, Google should have nixed the site long before now.
Rather than too much on-the-page optimization, a more likely culprit could be that some reciprocal links are no longer counted in Google's new system. This page had a number of links placed at the very bottom under the heading, "These links are for reference only. Many products are here for you to enjoy." The links, almost 50 of them, lead to unrelated sites that link back in return.
As explained earlier, these may now be useless reciprocal links, no longer carrying as much credit or any credit as before. If that's the case, the site will need to seek out some quality reciprocal links from sites related to the topic of air purification to get back on top.
I also got a Google point of view about the site. They affirmed that having the reciprocal links at the bottom of each page in the site could serve as a "hint" that the site is geared too much toward links and placement.
How are the Google results now for air purifiers? Before, the listings were entirely for sites selling these products. That's great if you want to buy but perhaps not if you want general information to help in your purchase decision. After, Google still has some of the sites that were there before. However, it also has picked up an advisory page about purifiers from the US Environmental Protection Agency, a nice addition. A page from a site reviewing purifiers, rather than just selling them, is also present.
Overall, the change at Google is a slight improvement, I'd say. It's also an improvement that could only come at the cost of at least one of the 10 product sites previously listed getting dropped.
As for heart rate monitors, "old" Google showed all commercial sites pitching these products. New Google is virtually the same. It did gain a comparison page from Epinions. It also gained a book listing from Amazon about heart rate training: somewhat informational, yet off topic.
And elsewhere? Inktomi is almost entirely showing sites that sell purifiers. Only one listing is more informational, a comparison chart -- though products are for sale via this page, as well. Overall, Inktomi looks a lot like "old" Google and certainly isn't better than "new" Google. It's also notable that 5 out of the 10 listings come from paid inclusion.
As for heart rate monitors, Inktomi looks virtually the same as Google, mostly commercial listings. Inktomi also has an Epinions review page and gets a slight bonus for variety by showing a page on selecting such tools from About.com. Two of its 10 listings, including the Epinions review, come from paid inclusion.
Reader Jim Ingrum of SkateMall.com doesn't buy into the idea that Google has made its change to hurt those who don't advertise. In fact, it's the opposite.
"If you do a search under roller skates, Skatemall.com was listed third last month. However, we are also a big advertiser with Google and are now nowhere to be found. What is also interesting is that Newskates.com and Skate-Buys.com also are not found and they also advertise through Google. Netskate.com, however, does not advertise through Google and they are the only skate shop still listed on the first page of the results. It seems to me that the people that do advertise with Google are the websites being punished," Ingrum emailed.
Ingrum's site isn't backed by a giant company, so his ranking loss is an example of how some feel the Google change has dropped "mom and pop" small sites. Yet other small sites remain in Google, such as the top listed site that sells Aircoast-brand skates, the Roller Warehouse site or RolleryHockey.com.
Other dropped pages are listings no one will miss. One was an affiliate page for an online shoe retailer, promising best prices on a particular brand of roller skates but not delivering. Three other affiliate-style pages were dropped from the top results, including one sending traffic to eBay and one sending traffic to Amazon.
Are the new results any better? My mother, who was a professional roller skater and had me skating at three, would be happy to see the informational page from the US Olympic Committee about roller skating now in the results. But the two pages from colleges on roller skating ordinances? Completely off-target -- keeping Ingrum's site would be better than adding these.
Let's do the tour. Inktomi gets marks for having two informational pages in the top results and adding variety, rather than the single decent one Google gained. Four of the listings shown are paid inclusion -- and SkateMall isn't present until the second page of results. It's much better than at Google, but the page still won't be seen by an estimated 85 percent of searchers.
At Teoma, it's all commercial -- and at least two of the listings come from paid inclusion. To be fair, there is some informational content in the Resources area of Teoma's results -- but these are easily overlooked by a typical searcher. SkateMall doesn't make it until the second page of results -- behind two more paid inclusion links for DealTime and Walmart.com, which already had links shown in the first page of results. DealTime and Walmart then each get paid inclusion listings again at the bottom of the second page of the results.
At AllTheWeb, it's all commercial except for the movie web site listing for the remake of Roller Ball. SkateMall does make it into the top results here.
Brian Lane of Above Ground Pools writes, "Over the past several years we were coming up in the top five of Google's search results when you typed in above ground pools. But, now when you type it in, you get results that look more like a listing of information sites and partnerships with sites who offer stores."
As with many of the other case studies, nothing jumps out to explain why Lane's site and several other commercial sites would have lost ranking. Some sites selling above ground pools do remain. However, the new informational content isn't very helpful -- above ground pool regulations for the city of Philadelphia or for St. Louis, for example.
Both of those listings make this meteoric rise without the benefit of any external backlinks pointing at them, that I can see. Given this, it's arguable that the results are worse. They don't help those looking for information, and they do hurt those who may want product info.
Competitor tour. Inktomi's all commercial, with two paid inclusion links in the top results. Lane's site doesn't make the top here. AllTheWeb's all commercial, but Lane does make it here. Teoma's all commercial, but a listing from Service Magic is unusual in pitching contractors, rather than pools themselves. Walmart's also unusual in offering a guide to above ground pools that can be purchased. Notable also is that both of these exceptions are paid inclusion listings. Lane's site doesn't make the top results here.
Tim Cohn of Advanced Marketing Consultants isn't upset about any particular ranking loss. Instead, he's disturbed by the fact that since the change, the way he appears in the Google Directory hasn't stayed consistent.
The Google Directory gets its listings from the volunteer-run Open Directory Project. Google takes the Open Directory's listings and shows them in PageRank order, which means in order of popularity based on how Google counts links to the sites from across the web.
Cohn's site is listed in the Smart Business > Marketing and Advertising category. He noticed that since November 13, if he changes sites from being listed in PageRank order to alphabetical order, the name of his company would change back to the old "Ideal Marketing Associates" title. He doesn't know why it is happening, but it's left him feeling that all is not right with Google.
"My contention is Google has been broken since November 13th, at least as it pertains to their ability to consistently deliver this one piece of data," Cohn said.
When I checked, the problem seemed to have now gone away.
Try a Google search for cleveland search engine optimization. This is a great search, because it shows weaknesses in both Google and Yahoo.
One of the top listings at Google is a Yahoo Directory category for Cleveland Zoos & Aquariums. What's that got to do with search engine optimization? Visit the page, and you'll see how Yahoo has put a final stake into the heart of its once useful directory.
Overture contextual paid links have now been added to the Yahoo directory. I first noticed these about two weeks ago, when I was teaching a class on web searching. I always demonstrate the Yahoo Directory as a useful way to get a professionally edited list of web sites.
Whoops! Now when you get into some categories, the new paid listings from Overture can make it hard to see what's the editorial content. The Cleveland Zoos case is an extreme example. One editorial link versus up to four paid listings that I've counted.
Worse, the paid listings aren't even being automatically targeted right by Overture. SEO ads on a page about zoos? The final kicker is that Google has indexed the content of these links, so now it thinks the page is about search engine optimization! So much for those who think Google gives too much weight to links and not enough to page content. In this case, the page content was completely misleading.