How do I optimize an e-commerce site without rich content?

A difficult question from Paddy Moogan in the UK. “Hi Matt what are you opinions on optimizing an Ecommerce website where the main pages/products may not necessarily be rich in content?

That’s a tough question, alright. Essentially you are saying here’s a place where you can buy products and there’s not a lot of content or may be the content is duplicated from a bunch of places. So my short answer is put on your user-hat. If you type in a product and you get a ton of places to buy things and there’s no real information on that page and there’s no real value add and all you get is places to buy then you get pretty annoyed. In fact whenever we ask your users, what’s the top issue for you these days it’s less about web spam like cloaking or hidden text or other stuff like that and more about search quality issue, Oh I don’t like how many commercial results I see or I get too many products or too many comparison shopping sites or things like that. So that’s the sentiment that we’ve heard a lot and what you should be asking yourself is do I want to take that step do I want to make and ecommerce site if I don’t have a lot of original content or value add or if I have a lot of duplicate content or if it’s in a fillet feed and if there’s really not that much that I’m adding to it and so I’d ask you, if you really want to jump into that and start optimizing that or do you want to look into something more original something more compelling some other hook that you could get a lot of visitors for and so my advice is, if possible think about how you can move towards that high value add, unique sort of site not just a site that somebody might view it as cookie cutter or they get annoyed about if they land and find a page just looks like 500 other pages that they have just seen on the web. So those are the sort of things to think about.

Two questions about the link: operator

Sven Heb from Wiesbaden, Germany asks “how accurate is Google’s back link check (link:..)? Are all nofollow back links filtered out or why does Yahoo/MSN show quite more back link results?”

The short answer is that, shortly we only have room for a small percentage of back links because web search was the main part and we didn’t have a ton of servers for link: queries and so we have doubled or increased the amount of back links that we show over time for link: but it’s still a sub-sample, its relatively small percentage. I think that’s a pretty good balance, because if you just automatically show a ton of back links for any website, then spammers and competitors can use that to try to reverse engineer someone’s ranking. And you don’t necessarily want someone spying on your rankings and try to figure out how they can compete with you by getting every single link that you’ve got. What we do instead is a nice compromise; if you register your site in google.com/webmasters or Webmaster council then you can see all of the back links that we know about you. So a vast vast vast the majority of the back links that we know about are there in Google’s Webmaster council. So you can look at the sub sample for any website or any page on the web, but if you want to see pretty much the full dump of what we know about you can see it for your own site but not necessarily for your competitors. We think that that’s a pretty good compromise and that’s probably the policy that we’ll have going forward.

San Diego Tim from San Diego says “If you have inbound links from reputable sites but those sites do not show up in a link:webname.com search, does this mean you are not getting any ‘credit’ n Google’s eyes for having inbound links?

No, it doesn’t. link: only show a sample, a sub sample of the back links that we know about and it’s a random sample and it’s not like we only show the high page rank back links that’s what we used to do and anyone who had a page rank of four or below wasn’t able to see their back links because they weren’t in the high page ranks they weren’t getting high page rank links. So we made it more fair by randomizing which back links we would show and we also doubled the back links we would show at that time. Now what’s interesting if you only show links that flow page ranks or that we trust or that do not have nofollow then people could kind of reverse engineer that and say, oh I’ll try to get the links that are really valuable. So we show the links that carry a lot credit in our system and we also show the links that we don’t really trust or don’t carry a lot of credits in our system. So it is truly a random sample, stuff that’s nofollow, stuff that is followed, stuff that we believe a lot, stuff that we don’t trust that much. Just because you don’t see one particular link in link: it doesn’t mean it doesn’t or does flow reputation or page rank or whatever you want to refer to it as. If it’s your own site you can use Google’s Webmaster council sign up and get a very complete basically the vast majority of links that we know about as a dump that you even download as a CSV file. So if you do want to get a really good idea of your back links that’s the place to get a pretty exhaustive list of your links according to Google.

Why does Google index blogs faster than other sites?

Lots of questions from the UK! Lee Willis from Cumbria, UK asks “Why does Google crawl/index blogs (specifically sites notified by “WordPress XMLRPC pings”) so much faster than a “normal” site submitting a revised Sitemap. What is the impact of that on the overall “quality” of the index?

Well, we always try to maximize the quality, the relevance and the accuracy of your index, and you want to make a distinction between crawling and indexing because site map submission does not guarantee that we will crawl the URLs on that list, it is very helpful to help us discover new URLs or to make canonicalization decisions. But we don’t guarantee that if you submit a site map we will go ahead and crawl it. There have been some people who did some experiments where they saw that happen but I’m not going to confirm it or deny that for the policies can always change on how exactly we do use site map submissions. But crawling and indexing is different, so if you do a ping a lot of time Google will come and crawl you but often its Google blog search because if you are doing those WordPress or web blogs or FeedBurner pings those pings are often those sort of things that are blogs. So the blog search can come and crawl you five minutes later. But then if you show up, you might show up in the blog search corpus not in our main web index corpus. So just because you get crawled it doesn’t mean that you are getting any sort of index boost or anything like that. We do sort of try to rationally decide what is the best quality of data, how do we get that, sometimes its crawling stuff immediately like with blog search you have very fast, very real time sort of results. And sometimes it’s taking site maps, it might result in crawling at a different pace or you may not get any boost at all but we do use that information not to waste and help us try to improve canonicalization and help us try to improve the quality of the index. So I wouldn’t say Ping that’s the way it automatically get crawled or anything like that. We make great content you get to be well known we will probably crawl you relatively frequently and see updated content any time you make a good change.

Are Google SERPs moving to Ajax?

Here’s a question from Owen in London. Owen in London asks “Can you confirm if the Google SERPs are moving to AJAX, http://tinyurl.com/be5shp, if so do you think it’ll affect analytics which rely on the keyword information being in the URL?”

So, Google did roll out a change few weeks ago, which is for a very smaller percentage of users, very small like under 1% right now are doing almost what you might call java script in hand search results. So you show up on Google’s page and as you are typing you can do neat things on java script and so you can try to make things faster and you can try to make things smoother for users, there’s a lot of really smart stuff you can do. The team really didn’t think about refers and how that might break analytics packages and stuff downstream. So it’s a very small percentage of people this is being trialed on and people are thinking about, are there ways to have refers anything that you do is so useful to have refers. So ten years from now, if refers are not the conventional browser sense then may be browsers can return everything after the poun sign. For example even though after the # mark or after the poun sign isn’t officially part of the URL or URI and if browsers were to pass that along then that would help all sorts of refers and analytical packages. So the way that I think about it right now is we have try experiments as how to make search results better and faster and cleaner and it’s not the intent to break refers but we have to keep trying out new things and we do want to have the ability where analytics packages can still continue to work.

More than one H1 on a page: good or bad?

A very short to the point question from Erin, south of Boston. Erin asks “More than one H1 on a page: good or bad?”

Well, if there’s a logical reason to have multiple sections then it’s not bad to have multiple H1’s. I would pay attention to overdoing it, if your entire page is H1 that looks pretty cruddy. So don’t do all H1s and then you see assessed to look like regular tags. Because we see people who are competitors complain about that if users ever turn off CSS or if CSS doesn’t load it looks really bad. It’s alright to have a little bit of H1 here and then maybe there’s two sections on a page so may be have a little bit of H1 here. But you should really use it for the headers or the headings which is what the intent is and not to just throw H1 everywhere you can on the page. Because I can tell if you just throw H1 everywhere on the page people have tried to abuse that and so our algorithms have tried to take that into account so it doesn’t really do you that much good. So I would use it where it makes sense and more sparingly but you can have it multiple times.

Will Google add guest accounts to Webmaster Tools?

Here is a question from Ian M from United Kingdom. “Is Google planning to create read-only “Guest accounts” for Webmaster Tools? Many clients (particularly in heavily regulated industries e.g. banks) are very reluctant to provide access to a third party.”

Great feature suggestion! I have no idea! Because the Webmaster Tools team they have to plan their resources and what they work on just like any other team and I can see a valid use for this, at the same time there are other things that the tools folks are working on that are really useful. Some people want infrastructure updates so that back link reports always rock solid or new data is really really fresh and it’s hard to play that off. So it’s a valid suggestion, I appreciate the suggestion I don’t know what level priority they give that, because there’s probably a relatively limited impact compared to making reports rock solid or overhauling our UI’s and things like that, that’s going to be useful for every single person not just for a smaller faction. But that’s something I can imagine as doing in the future so we’ll definitely take that into account and we appreciate the suggestion.

Does the position of keywords in the URL affect ranking?

Interesting question from Adeel from Manchester, UK “Does the position of keywords in the URL have a significant impact: example.com/keyword/London is better than example.com/London/keyword?

Truthfully, I wouldn’t really obsess about it at that level detail, it does help a little bit to have keywords in the URL but it doesn’t help so much that you should go stuffing a ton of keywords into your URL. If there is a convenient way that is good for users, where you have four – five keywords, that might be worthwhile. But I wouldn’t obsess about it to the level of how deep is the URL in the path or how am I combining it. For example, on my blog when I do a post I’ll take the first 4-5 words or 2-3 words related to that post and I’ll use that as the URL. But you don’t need to make 7,8,10 or 20 words because that just looks like spamming users and people will probably not click through as much in the first place. So position is going to be very very second order kind of thing of keywords and URLs. I would not worry about that so much as having great content that people want to link to and people want to find out about.

Is redirecting a large number of domains suspicious?

Cweave from Dallas asks a really interesting question “When permanently redirecting (301) a large number of domains (read: more than 10) to 1 domain does Google flag this as suspicious? What considerations does Google look at? For the purposes of this question let’s assume this as a consolidation move.

I think there is plenty of valid reasons why somebody might do this. For example, if you Google, there are ton of people who have registered Google typos and we try to get those as we don’t want people to get confused to get malware. So we end up with a portfolio with lots of Google related domains even things like Google sex and Google porn. And so I think it’s perfectly logical to have misspelling of Google and all that stuff just to a 301 to Google’s home page. So that’s what I think Cweave was talking about when they said consolidation move. At the same time if we see a ton of 301’s all going to one domain, then we might take a look at that, you could certainly imagine someone trying to abuse that or do spam so we could take a second look or scrutinize that. But if all you are doing is trying to consolidate misspellings or a bunch of brands, and by brands I mean a bunch of domains that you have registered that are very in-severe domain and you really only have that one domain, I don’t foresee that being a problem because people would check it out or someone reported it as a spam thing, if we took a look we’ll just see oh yeah they are just consolidating their brand. So Google might take a look but I don’t consider that to be a large problem.

Will Google use non-link references as a signal?

This question comes from Boston, EA; Eric Enge asks “Do you think web search will ever make use of references (web site mentions that are not links) as a ranking signal?

So there are two answers. The first one is, I never want to take a ranking signal off the table like, I’ve joked, that if the face of the moon can help us rank the search results better, I’m willing to use the face of the moon. At the same time think about how people would attack the use of references. Right now a lot of people rely on getting links, if all they have to do is example.com in text and then you can leave as comments all over the web and all over blogs and all over forums it will almost be anywhere you can stamp any user generated content people would be leaving those references. So that’s the sort of reason why you might be skeptical about why we’d use this sort of signal, because people could abuse that sort of thing they could just leave mentions of the URLs even if they can’t generate links. But I’ll say we are willing to look at it, we would run the analysis we would say is there a way to pull out that signal from that noisy data where we could find a way to improve it. But that would definitely be a sort of thing that people would try to abuse it.

Should large corporations use rel=canonical?

Terry Cox from Orlando, Florida asks “In regards to the new canonicalization tag, does it make sense for large corporations to consider placing that tag on every page due to marketing tracking codes and large levels of duplicate URLs like faceted pages and load balancing servers?

So this is a great question, should you put the canonical tag on every single page, well there is a short term answer and a long term answer. Short term answer is I would probably say not right now, take a little bit of time, study your site architecture, think about URL normalization, beautification, whatever you want to call it think about the structure of URLs you want to have and take a few weeks or few months or couple of months you sort of assess where you want to go. I don’t think you should throw the canonical tag on every single page on your site immediately and move it around, because it is a powerful tool and people do have the ability to shoot themselves on their foot. So on the plus side we’ve seen a quarter of a million pages show up within just a few days, where people are using this canonicalization tag, which is fantastic. It is good to see the traction and the adoption move very quickly. On the down side we have seen one company, very large company, computer company, I won’t call them out by name, where they had a home page and the home page was doing a redirect and they also had a canonical tag and the canonical tag pointed to a page that we hadn’t crawled at all and so those sort of cases can be very difficult to try to do the right thing, and we do the right thing. But it can take us a couple of days to sort it out or go and find that URL and crawl it. So I wouldn’t just jump in deep in the pool without doing some planning. The longer term answer is that it doesn’t hurt to have this on every single page of your site. Ideally you’d find other ways to solve the canonicalization but it doesn’t hurt to say on every single page this page maps to this canonicalized, very pretty, very preferred version of this URL. But what you want to do is, you want to make sure that, it is absolute URLs, ideally goes in one hub, it’s a logical system that you designed you haven’t just jumped out and started playing around with. I don’t see any harm in having that sort of thing because we’ll just follow those, what we almost think of is many 301 redirects within that site and we’ll try to canonicalize according to those suggestion. We don’t guarantee that we’ll do it but it should work just fine with no problems. So feel free to do that but take some time and plan it out a little bit.