Search quality highlights: 52 changes for April

[Donna]

Search quality highlights: 52 changes for April
5/4/12
[insidesearch.blogspot.ca ]

Some of the more interesting highlights :

Increase base index size by 15%. [project codename "Indexing"] The base search index is our main index for serving search results and every query that comes into Google is matched against this index. This change increases the number of documents served by that index by 15%. *Note: We’re constantly tuning the size of our different indexes and changes may not always appear in these blog posts.

New index tier. [launch codename "cantina", project codename "Indexing"] We keep our index in “tiers” where different documents are indexed at different rates depending on how relevant they are likely to be to users. This month we introduced an additional indexing tier to support continued comprehensiveness in search results.

Keyword stuffing classifier improvement. [project codename "Spam"] We have classifiers designed to detect when a website is keyword stuffing. This change made the keyword stuffing classifier better.

No freshness boost for low quality sites. [launch codename “NoRot”, project codename “Freshness”] We’ve modified a classifier we use to promote fresh content to exclude sites identified as particularly low-quality.

More authoritative results. We’ve tweaked a signal we use to surface more authoritative content.

Fewer autocomplete predictions leading to low-quality results. [launch codename "Queens5", project codename "Autocomplete"] We’ve rolled out a change designed to show fewer autocomplete predictions leading to low-quality results.

More efficient generation of alternative titles. [launch codename "HalfMarathon"] We use a variety of signals to generate titles in search results. This change makes the process more efficient, saving tremendous CPU resources without degrading quality.

More concise and/or informative titles. [launch codename "kebmo"] We look at a number of factors when deciding what to show for the title of a search result. This change means you’ll find more informative titles and/or more concise titles with the same information.

Better query interpretation. This launch helps us better interpret the likely intention of your search query as suggested by your last few searches.

Anchors bug fix. [launch codename "Organochloride", project codename "Anchors"] This change fixed a bug related to our handling of anchors.

More domain diversity. [launch codename "Horde", project codename "Domain Crowding"] Sometimes search returns too many results from the same domain. This change helps surface content from a more diverse set of domains.

And more .....

Probably its not only Panda and Penguin what's causing all the commotion last month.

[edited by: Robert_Charlton at 12:19 am (utc) on May 8, 2012]
[edit reason] added quote box and fixed title [/edit]

[deadsea]

The new index tier is the most intriguing to me. Its kind of like they took the supplemental index and expanded the concept at some point.

I'm sure only the best sites are in the top tier index. If some algorithm decides your sites is lesser quality, or you get a manual penalty, it sounds like you site might get re-indexed in a lower tier. To recover, you'd have to have the penalty lifted AND wait long enough to be re-indexed in a higher tier.

[Sgt_Kickaxe]

Better query interpretation. This launch helps us better interpret the likely intention of your search query as suggested by your last few searches.

How do I stop my last few searches from being used in this manner? Is being logged out enough? Google needs to clarify this as it has privacy concerns attached.

MLB search feature. [launch codename "BallFour", project codename "Live Results"] As the MLB season began, we rolled out a new MLB search feature. Try searching for [sf giants score] or [mlb scores].

The #1 result is ESPN which shows scores right in the serps and Google's new display of scores is apparently coming from ESPN. I think ESPN might have something to say about their data being scraped like that...

[Robert Charlton]

MLB search feature... data being scraped like that...

Sgt_Kickaxe - Your question goes over the same ground as this discussion you started here in Oct 2011, on much the same topic....

How do I stop Google from poaching data for snippets?
http://www.webmasterworld.com/google/4370321.htm [webmasterworld.com]

The discussion is worth rereading, as it was thorough and should have covered your concerns about sports data licensing. The major league teams do not go unrewarded because of such results in the serps.

If you try one of the suggested baseball searches, you'll see that it's the data display that has been enhanced considerably, perhaps even helping the teams by making it easier for fans to follow them. Fans can then visit official sites for more in-depth information.

[matrix_jan]

More domain diversity - BS...

[Sgt_Kickaxe]

Sgt_Kickaxe wrote...
I had started another thread in the Google policy section of WW <about copyright issues>. Either this post or that thread should probably be removed, they ask the same thing. Cheers.

See copyright discussion here...
[webmasterworld.com...]
.

[edited by: Robert_Charlton at 7:32 pm (utc) on May 7, 2012]

[BillyS]

No freshness boost for low quality sites. [launch codename “NoRot”, project codename “Freshness”] We’ve modified a classifier we use to promote fresh content to exclude sites identified as particularly low-quality.

I suspect this means IF you are hit by Panda, then keeping data fresh does not help.

If that's true, it means Google's telling us to take action on Panda first and foremost. Anything else you try to boost rankings will not work.

[creative craig]

More authoritative results. We’ve tweaked a signal we use to surface more authoritative content.

I now only use Bing because of this crap - I got sick and tired of Wikipedia being #1 for 99.99% of all search terms. I know Bing shows Wikipedia as well, but nowhere near as much as Google!

[jonathanleger]

I find it laughable that Google calls these search quality highlights. Apparently they haven't looked at their SERPs lately. It's either big brands and domains with massive numbers of backlinks (including sites that allow users to post their own content with little or no quality verification, like Blogger and YouTube), or it's thin content or content from ancient pages that haven't been updated in years.

Many of the same sites that were ranking based on massive amounts of SPAM links are still there, too.

Quality? Where's the quality? They axed most of the quality sites with Penguin.

[blend27]

In the past 4 month our main ecom site went from 4 to 9 to 11 to 21 to 51 to 76 for the main money making term.

So did most of the SMBs from US/EU and replaced by Amazon, Wiki, Ovestock and HSN on the first page. Along with 3 dead beat sites, that have not been updated for 5 years or so. These dead beat sites do not participate in any Social Media what so ever. No link building, nara...

Completely unique niche. Not one Item is the same, ever due to part of the widget being made from a several million year old tree sap. Never needed to compete with BigBox junk sellers. And no this not is one of the things you cant buy in CN(wholesale), people who buy it most likely do a research before making a purchase.

We still sale, not a bad month was April, but G's Traffic, conversion from SERP is ZILCH.

Several competitors dropped from top 20, well worth the positions, down to 80-90-ish.

Quality - my-my G. I had 7 visitors from Goog 3 days ago, down from 500-550 a day couple of month ago.

And Yes, all White Hat, no tricks.

[CainIV]

Sounds like an all out shutdown on utilization of exact match. Reducing "low quality" autocomplete might include exact match domains trying to cash in on popular terms + how to, or even warnings, dangers, etc. You see this exemplified lots in affiliate owned SERPs

[WebPixie]

More authoritative results. We’ve tweaked a signal we use to surface more authoritative content.

A large percentage of our sites are smallish niche sites. In almost every case the niches are being dominated by this authoritative content, if by that they mean a 300 word article about sky blue widgets on a mega-site out-ranking sites dedicated to sky blue widgets with at least 50x the content.

We have done okay because we thankfully decided to over-build these site years ago, meaning if we thought we could cover the topic well with 10,000 words we went ahead and did 20,000 words just to be safe. But we have seen many very good competing sites buried by this worthless content on the mega-sites.

Is beating spam worth killing most of the sites not run by MegaSuperOmniCorp? So far it appears that G- thinks so.

[Martin Ice Web]

The realy big question is, will this silly quality authoritative content stick? Can we go now to work, or will it change tomorrow?
I must be blind or my arithmetic unit must have an error, cause I can´t see it. If Cutts could even give an example, but he can´t, that´s a pitty.

[atlrus]

More authoritative results. We’ve tweaked a signal we use to surface more authoritative content.

I now only use Bing because of this crap - I got sick and tired of Wikipedia being #1 for 99.99% of all search terms. I know Bing shows Wikipedia as well, but nowhere near as much as Google!

^^^ This x100

It should've been written as "More Wikipedia results" instead.

Talking about throwing in the towel - wiki is now showing atop everything I search...Why don't they just substitute the "I'm feeling lucky" button with "Wikipedia" button and be done with it? It's essentially the same thing.

How much of a failure for a search engine is the fact that Wikipedia is the #1 result for pretty much everything?

[Scurramunga]

How much of a failure for a search engine is the fact that Wikipedia is the #1 result for pretty much everything?

Indeed, why have a search engine at all? All that's needed is a static html page with a Wikipedia link, an Amazon link, a Zazzle link, a couple of Youtube links, a Pinterest link, a Walmart link, a Facebook link and some other big brands thrown in for good measure. Google can still run ads on that page, after all is that what counts most?

Forget p2 results because no one looks at those. A static html page with just a Google logo.gif would save them squillions.

[Shaddows]

How much of a failure for a search engine is the fact that Wikipedia is the #1 result for pretty much everything?

It depends. Say, if Wiki was in the top 5 results, it achieved an average 50% CTR. That would strongly suggest most searchers preferred Wiki to RandomSite.com.

Now I'm not privy to actual CTR, but I know from casual shoulder-surfing that most people go straight to Wiki, regardless of its position. Ergo, Google is returning the most popular destination in the top spot. That would not constitute "broken"- just annoying for the other 99.999999% of site owners. Unfortunately for those owners, Google is not trying to please them.

[atlrus]

It depends.

Of course it does :)

We can look at it in a different way, though:

Say, if Wiki was in the top 5 results, it achieved an average 50% CTR. That would strongly suggest most searchers preferred Wiki to RandomSite.com.

Or it strongly suggests that Google failed to return any other authority site, which is likely the more plausible scenario. Wikipedia is great and all, but for Wiki articles, where I am quite a knowledgeable expert - I see so many mistakes and misinformation that Wikipedia is hardly an authority site over some very good real expert websites (some show in the search results, others don't). Wikipedia's generalization and fragmentation of information, as well as actual facts, are skewed enough for me not to consider them at all, when I need real information. In other words, Wiki is an authority website only because Google says it is.

but I know from casual shoulder-surfing that most people go straight to Wiki, regardless of its position. Ergo, Google is returning the most popular destination in the top spot.

But is this really what a search engine should do? The most popular destination does not need to be returned for every result. People generally go to a search engine to find alternatives. Showing the same website over and over for pretty much every search possible is saying one thing : "Our algo is so messed up - we can't return anything good out of the billions of websites available, so here is a link to Wikipedia instead."

[rlange]

atlrus wrote:
How much of a failure for a search engine is the fact that Wikipedia is the #1 result for pretty much everything?

Bing, Not Google, Favors Wikipedia More Often in Search Results [Study] [searchenginewatch.com]

--
Ryan

[netmeg]

People generally go to a search engine to find alternatives.

I don't think this is true. In my experience with hundreds of clients since long before the days of Google, they go to a search engine to find an answer. Not alternatives. An answer. And generally they're happy with the first one or two they find. WE ARE NOT NORMAL USERS. Normal people don't search like we do.

[Robert Charlton]

Normal users also don't resent Wikipedia because it ranks above them. ;)

[jonathanleger]

My loathing of current Google results has nothing to do with losing my own rankings. I didn't lose them.

My loathing of the results has to do with the trash that has replaced the good sites. For the first time since Bing went live I have had to go to Bing to find the answers I'm looking for. Google's relevance is way off right now.

I tried searching G for "used green widgets". The words are there, but the obvious meaning of the query is clearly ignored. Bing does not suffer from such lack of relevance for the same query.

Moderator's Note: please do not post the literal keywords for searches (good or bad).
There is no end to such examples, and other people may use the opportunity for nefarious
purposes.

Report your specific search complaints directly to Google rather than here. At the bottom
of every results page there is a link for "Give us feedback."

[edited by: tedster at 7:22 am (utc) on May 9, 2012]

[deadsea]

jonathanleger: bookmark google verbatim search, you will like the results much better.

All their synonym stuff ruins my search results because I do precise searches and often it finds not quite what I was looking for or ignores words from my query.

But looking through my log files for a site I have that doesn't have much SEO work done on it, I find that that site is ranking really well because of the synonyms. I hadn't done keyword research on some of the pages and used the synonyms and alternative phrases on the pages. Now those pages are ranking anyway, unless I have verbatim search on.

[atlrus]

I don't think this is true. In my experience with hundreds of clients since long before the days of Google, they go to a search engine to find an answer. Not alternatives. An answer. And generally they're happy with the first one or two they find. WE ARE NOT NORMAL USERS. Normal people don't search like we do.

I think I worded it wrong. What I meant is that people would often search to find an alternative to already known websites. For example, if you read the news at CNN, when searching for news you are likely looking for an alternative to CNN. Or if Wikipedia is your primary source of information (God, I hope not), you are searching for an alternative answer, otherwise you'd just go to your Wiki bookmark.

And yes, I agree, people do often search for answers and Google fails to deliver there just as often. Do a search for "flat tire" (I hope the mods will let this one stay). Do you honestly believe that Wikipedia being the #1 result returned is good "searchengining"? :)

Normal users also don't resent Wikipedia because it ranks above them. ;)

I don't think anyone is really loathing Wikipedia for ranking high. I know I don't, even though it ranks atop of almost everything I try to rank for. But Wiki is really not a competitor of mine in any way, so whether it ranks above me or not - it's irrelevant to me. I am lucky to be in a field where people are searching for the actual product and service, not its general description Wikipedia provides.

My point about Wikipedia is what's wrong with Google:

1. Arbitrary decision that a website is an “authority”. Wikipedia is not an authority website, rather than a huge blog made up of anonymous posts. There is no real experts standing behind what has been posted – a post could be true but it could be false, who knows? Where is this “authority” coming from?!?
2. Spray and pray – Google takes the said authority website and returns it for everything. Why? Because it's easier than trying to fix their algo which is really broken beyond repair from the years for tinkering. Like I said, in my field Wikipedia is absolutely useless as a returned result. I know it and all the people who search for those keywords know it. Only someone who has never heard of Wikipedia would click on it when searching for those keywords and chances are they would come right back to the results, because Wiki is not what they are looking for. Yet Wikipedia continues to be the #1 returned result. Just mind boggling...


I can't believe all Yahoo is doing is bickering over the CEO's resume instead of taking advantage of this opportunity to run some advertisements and get back a large chunk of the SE business back. Idiots.

[jonathanleger]

jonathanleger: bookmark google verbatim search, you will like the results much better.

My point is that I shouldn't have to do that for such a basic query with an obvious meaning.

Rather than getting back pages selling used green widgets, Google is returning results about how to use green widgets, or just ignoring "used" entirely and returning results on green widgets.

I may know how to use the verbatim option, but the typical searcher has no idea what that is - nor should they have to.

[netmeg]

When I search for "flat tire" the Wikipedia entry tells me what it is, what could happen if I drive on it, and differences between car tires and bicycle tires. It also includes prevention information, assorted methods of repair, and an odd paragraph about the dangers of changing a flat tire. #2 result is an authority site's page on how to change a flat tire. That looks pretty darn good to me.

Assuming many people search for just "flat tire" in the first place, which is a big assumption

[atlrus]

When I search for "flat tire" the Wikipedia entry tells me what it is, what could happen if I drive on it, and differences between car tires and bicycle tires. It also includes prevention information, assorted methods of repair, and an odd paragraph about the dangers of changing a flat tire. #2 result is an authority site's page on how to change a flat tire. That looks pretty darn good to me.

Assuming many people search for just "flat tire" in the first place, which is a big assumption

Are you serious?!?

The only actual information Wikipedia gives is scraped from the second result which is the Edmunds "how to change" article! Yet Wiki is ranked above it, just because...why?

But that's not all, just look at the rest of the top 10 results - only two websites related - edmunds and ehow, the rest are not even close:

* flatire burgers (not even flat tire, just "fla" tire)
* Fat Tire (Belgium beer)
* facebook page for a bar

WTF!

My point exactly - Google cannot return good top 10 results, period, so they just throw in Wikipedia and hope for the best...

P.S. Look at the Adwords, though, all of them about tires, no burgers, beer or bars somehow...

[netmeg]

And that's exactly the point. It's a reasonable result to me, and it isn't to you. It's all subjective, and it goes to how people search and what they expect. That's why you don't (and probably will never) see a mass exodus to Bing or Blekko or DDG.

In the absence of any clue about user intent, Google served up a variety of possibilities - "flat tire" doesn't mean anything by itself.

In this very forum we have people screaming for fewer choices, and people screaming for more choices.

But everyone knows Wikipedia, and in the absence of a clear choice, that's what Google serves for that query.

[JohnRoy]

When you use bing.com/search?q=flat+tire you would basicly get the same results, just some images added.

[jonathanleger]

When you use bing.com/search?q=flat+tire you would basicly get the same results, just some images added.

Um, no. In Google 6 of the 10 results aren't related to changing a flat tire or talking about flat tires:

1 Burger Joint result
1 Brewing House
1 Bar (Facebook page)
1 Decorating Company
1 Bike Shop
1 PARKED DOMAIN

Bing:

1 Burger Join
1 Band Page

So while Bing's results aren't perfect, they are far better than Google's (especially that parked domain!)

[atlrus]

In the absence of any clue about user intent, Google served up a variety of possibilities - "flat tire" doesn't mean anything by itself.

Oh, come on, now... You and I and everyone else imagines the same exact thing when they hear "flat tire", i.e. a deflated car tire. Google with all the talk about being more intuitive and what not returns beer which is NOT flat tire beer but FAT TIRE beer. Even you have to admit that this is just unbelievable.

Flat tire is pretty much a "blue widget". You can pick anything simple and do a search and sit back and laugh. I just did a search for "blue agave" - first result, naturally, Wikipedia, the following three are restaurants in California, while I am searching from the East Coast with location enabled...