SEO Analysis for Good!

Just a week ago I saw a post come through the Seattle Digital Eve mailing list from someone asking for help on behalf of a friend.

The post:

I have a friend who is looking to improve the SEO on their ecommerce site.  They are a small business, and the SEO people they've talked to want to sell them a big, complex package, when what they really need is some coding improvements on the site, consolidation of 2 sites into one while preserving link juice, better keyword/metadata, etc.  

Anyone out there with some decent SEO expertise?

Offering to Help

I emailed the person that posted to the group:

If it's a small project I can probably spend some time on it.
Even though I have a ton of experience, I can keep the fees down since I have a day job.

My email was quickly forwarded on and the business owner (Martin) emailed me right away with a simple "Hey Jenn, I am interested in your SEO skills. Please call me...". I called him on my way home as I was stuck in traffic that evening. Martin described the two websites for me, how they were getting rankings and traffic, and now aren't. He said he had done some work to one of them, and traffic kicked up. He asked me to take a look at them and see if it is something to do with the website, if either had been hit by Panda or Penguin, or if people just aren't searching for rugs online as much as they used to. I quickly thought - how odd that this is almost exactly what we have been going through with the websites at ADP. Not to mention that they have gone through this up and down, all the time me trying to figure out if it was just general loss of interest, change in searcher behavior, or if the site was going through a penalty or hit from just general Panda and Penguin updates. While this man's sites are a small version of what I deal with on a daily basis, this could not only be fun, but should be fairly easy to figure out.

Martin asked what my hourly rate is - but I immediately responded with a giggle (at $250/hr I'm sure it's probably more than he was expecting to pay) and told him that for the initial conversation and if it's something quick, I won't charge him. I then asked him if he has a Google Analytics account attached to the websites, of which he responded with a "Yes". I told him how if he can look at the Webmaster Tools data in Google Analytics and look at impressions vs. Avg position, and then impressions vs. Click Through Rate. If the Avg position, and CTR stay fairly steady, but impressions drop, then that is a quick way to see if interest has gone down. Another way to back that up, is look at how many terms get impressions one week, to the next week. If those numbers drop, then there is a good chance that rankings are dropping and there is a penalty. He was a bit overwhelmed, and asked if he gives me the login if I would take a look for him. I excitingly said "Sure" (since most people don't like to give out their login, it's easier to tell them how to do it - but this was the next best thing). I told him the next steps - I would spend some time that evening and see what I can find, then come up with a diagnosis and a plan of attack going forward of which he can either do himself, or if he has the budget, hire me to help him with it.

The Evaluation

That evening I settled in, opened up the computer while watching Hulu on the TV (we don't have cable), looked through both websites, and then logged into GA to dig through the analytics.

Checking 

Behavior vs. Penalty/Panda/Penguin

I started by going back as far as I could in the analytics account looking at organic traffic only. The month that both sites did the best was just a few months back this year. I also checked the timeline against Penguin and Panda updates via the Moz.com Google Algorithm Change timeline and noted if there were any clear drops that sync up with an update. There was a slight drop with one of the updates that hit one of the sites I work on for my day job that is built similar to his.

I first pulled the Impressions against the Avg position to see if the position stays the same while impressions go down. Though given that there is a slight drop after an algorithm update, I figure that there will be a drop in position, impressions, and number of terms getting impressions.

My First Chart:

Site #1

Impressions vs. Avg Pos.
Note: I took out numbers to protect the client even though he gave me permission.

Site #2

Impressions vs. Avg Pos

As you can see - there is a drop in impressions but the average position stays fairly consistent, and even more-so for Site #2. Both improved the last few weeks in positions, and impressions.

My Second Chart:
Site #1

The number of keywords showing impressions week over week for site #1

Site #2

The number of keywords showing impressions week over week for site #1

I noticed that the number of keywords getting impressions (meaning how many terms showing up when a person searches regardless of position) drops when the impressions drop in the earlier charts, and then goes up when impressions go up.

This clearly shows a penalty, and given that there was a penguin update just before that drop, it is pretty clear that the site took a hit from that update, then saw an improvement when he completed the little bit of work he did.

What Got Hit?

Knowing now for sure that both sites took a hit, the next step is to figure out exactly what was hit. With the site I manage at my day job I will usually run a category report to find out which terms were affected so that we can evaluate and establish a plan to recover. I didn't have time to set up the categories of terms to run the report (it takes days to categorize terms, but since that has already been done for the site I manage - it takes me just a few minutes or so to categorize now) so this time I grabbed the number of words in each term. If the phrase has just 1-2 words then it's safe to say it is a broad term, and if it has 3-5 words then those are more exact. Penguin tends to focus on sites that have optimized for long tail terms, and less on the broad terms. So, this is a faster way to get a similar understanding.

I ran a comparison to see how things were in his big traffic months compared to the recent months that there was a drop. I took 4 weeks in the high traffic month from a Saturday through the last Sunday and compared it to the last 4 weeks Saturday through the most recent Sunday. This would give me an exact day of week compared to that day of week and reference the beginning of a month to the end of a month. Ideally it should compare to the same time of year to reflect searcher behavior for the day of the week, time of month, and the time of year, but in this case the day of week and time of month was going to be good enough.

Example of Keyword Data with Count
Note: Terms and traffic are not representative - I changed them to protect the client.

The table above is what the data looks like. Do note, I changed the top keywords, and numbers to protect the client - but this gives you an idea of what I was working with. From there, I created a pivot table and played around with the data to give me more insight into what was going on. Number of words in a phrase, visits, pages/visit, etc. It all helped me understand what was going on before and after the update.

Keyword Count - showing how many words in a phrase were driving traffic from high month compared to low month.
Note: numbers and details have been removed/changed to protect the client.

The 3-5 word terms dropped from the high month to the recent months. This shows that the longer tail terms were hit, which is pretty indicative of Penguin.

Looking at the Sites

Having spent the first hour of my time running the reports and pulling charts, I spent the rest of the time looking through the sites now that I know what to look for. The sites were once optimized for long tail terms, but something happened that they lost that traffic. As I dig through both sites had categories for the two word terms (such as "area rugs) with links to individual pages for each item that fit in that category. The first that I noticed is that there is a URL hierarchy (something the website I work on lacked). So he was good there...

I started noticing as I was looking through both websites, that they were structured exactly the same way. I also saw that the navigation was the same on the left linking to different URLs, but the content appeared to look very similar on the pages. I grabbed a couple of the URLs that were focused on the long tail terms and pasted them into copyscape. The report kept not only pulling the other site as the first match, but other sites that sold the same products. This is a very common issue with eCommerce websites - since they don't have the time to write their own copy for each product, they tend to pull it in dynamically through syndication. With not enough unique content on those pages, then the site appears to be duplicating what all those other sites have. It's not a majorly serious issue, as Google tends to understand syndicated content, but if a site doesn't support the content with something unique they just won't get rankings as well as the ones that do.

Martin's sites had a bit more of a issue though since he has two sites with the exact same content, and the exact same structure. When I compared the terms that both sites got visits from during the high month, then I noticed that not only were a lot of the terms the same, but there were a lot of the terms with site #2s domain, and name in there.

Compare Terms from both sites with visits from organic traffic.
Note: the domain name has been changed, and so have the numbers to protect the client.

I think at this point it was very safe to say that the site took a hit by the Penguin update to the long tail terms due to both sites duplicating each other.

My Email and Recommendation

After spending a couple of hours on the site, I drafted this email and attached the excel document I used to analyze the sites (note: the email below is changed slightly to protect the client):

Hey Martin -
So I dug into both sites and the Google Analytics to see what’s going on. I’m attaching my excel doc if you want to see my work, but it looks like both sites definitely took a hit of some sort.
 Moz.com keeps a list of when updates happen so you can keep an eye on things: http://moz.com/google-algorithm-changeThere was an update at the end of January then another big update in March that may have led to you losing your rankings. I've seen this drop in other sites that are built very similar to yours – so I dug into the analytics to make sure that my assumptions are correct.
 What I found:I first compared visits from organic traffic (SEO) against your average position the past few months (webmaster tools only goes back 90 days, so I couldn't go back to January unfortunately).
 Site #1 definitely saw a decrease in traffic along with the drop in conversions (pasting the charts below for you to look at). With Site #2 there was a drop in traffic, but the average position seemed to not drop as much. Usually this would be a sign that people aren’t searching as much, so I wanted to check your keyword count and impressions week over week. If the number of terms drops seeing impressions drops from one week to the next it is usually an indicator of a penalization or hit by a panda or penguin update. I’m not pasting those charts in here since they are really raw, but you can see them in the excel sheet. The terms that have 2-5 words in there took a huge hit, while the one word, and longer tails appear to be sticking around. I toggled from keyword count to visits from the keywords and those sets stay pretty consistent in dropping.
 What this means is that you most definitely took a hit in rankings from the updates. Not just rankings dropping, but a bulk of your 3-5 word terms dropped out of the index completely. Those 3-5 word terms are also the bulk of where your visitors come from – those longtails are higher converting terms and can really affect revenue if they drop off. It looks like both sites are build very similar and have a lot of the same content. I compared the top referring terms both sites saw in your highest traffic month and both refer traffic for “your domain” which isn't good. They both get traffic from “broad term” but site #2 has site #2 beat there. They also both have several long tail terms that are the same.
 When I run a report on copyscape.com to check for duplicate content – the site #2 along with a few others come up (included link directly to copyscape) The “product” rugs page on both sites is exactly the same – almost word for word.
 It’s kinda fun to have two sites show up for the same terms, since you could get double the traffic. In fact that’s what my company does – and what I manage. We have dozens of “portal” sites to grab leads to sell to car dealers. But if Google has any idea that both sites are connected then both sites get penalized. I think this might be what has happened here.
 Your first solution would be to get unique copy on all of the pages of the site. I know it’s tough writing for all of those pages, and copy writers can be expensive. There is an alternative called TextBroker (http://textbroker.com). They have writers that bust out copy pretty quickly (2-3 day turnaround) at a pretty reasonable rate.
 I would recommend getting an account set up and start asking them to write for your pages. Even your homepage content – while there is a lot of it, but looks to be pieced from other content on the web.
 The order I would have them do it in is in order of the pages that had the most traffic in your highest month, and then work down from there.
 Once you get them going on that – I can do a full keyword analysis, check to see where the opportunity might lie, and get you a complete plan.
 The excel doc is attached – let me know if you have any questions.
 Hopefully this was helpful J

All in all it took me just a couple of hours since this is what I do for our executives regularly, so I didn't charge Martin for the work.

SEO for Good!

Martin was so excited and appreciative of the work I had done, and what I had found that he asked me if he could pay me in some way - "..if anything to help the school in Nicaragua". I gave Martin the link to donate to the school, and he did.

The money immediately went to help buy supplies for my Husband's students we are bringing with us. Since they had to pay for their immunizations out of their paychecks, don't have sleeping bags (we are loaning bags to them), and anything else they need they pay for themselves, I wanted to help them so that they could focus on helping build the school and not stress that they have everything they need for the trip.

In the end, I helped Martin with his websites because I like to help small businesses succeed, Martin returned the kindness by helping the students with their supplies, so that they in-turn could help build the school for children in Nicaragua.

Everybody wins!

Anatomy of the URL and Stuff

I'm sure you are looking at the URL above and thinking to yourself; "Wow, I never realized that all that stuff meant something." Oddly enough it actually does... As the world wide web has changed into a search friendly, user interactivity playground, the formation and meaning of the URL has evolved considerably in to a very significant factor in not only search engine compliance but in how people use websites. Lately I have been helping clients understand how their website's are structured and how servers to browsers to users work. It's something us search optimizers view as something so simple yet can be so complex to someone who doesn't understand how it all works. So here is the URL broken down piece by piece and explained.  

First - What is a URL? 

A Uniform Resource Locator is a website address that holds very important information between each "." and "/" much like an address to your home contains a house number, city, state, country, etc. This allows the browser to connect to a specific website, directory/path, and/or file in order for the user to see what it is on the page you want them to see. A URL consists of the following:

Hypertext Protocol Established by English physicist Tim Berners-Lee in 1990, hypertext protocol is a request/response standard typical in which the client is the application (User on a web browser such as IE, Firefox, safari, etc) and the server which hosts the web site itself. The client submitting HTTP requests is typically referred to as a user-agent (or user) with the responding server—which stores or creates resources such as files (html, .asp, .php, css, etc) and images—referred to as the origin server.*

  WWW (World Wide Web) or "sub-domain" 

The WWW is typically placed before the main domain of your website URL, referencing the World Wide Web. Remember the game you played in elementary school where you could start your home address with your house number, street, city, state and then go off as far as your country, continent, and even earth. The WWW is the address starting with "earth". In some cases, what we call a "sub-domain" can replace the WWW in your URL, which references a whole new website within your existing domain. Search optimizers can use this as a way to target certain key terms. For example, a real estate agent targeting a specific city will use http://city.domain.com and thus will have a leg up when ranking for anything within that city. In most cases the sub-domains will link to the main domain and, since they are treated by most search engines as a domain all it's own, then it will count as an external link credit, boosting the rankings for the main domain it is linking to. It is highly recommended that you avoid this technique as it is only tricking the search engines and in the end will hurt your rankings rather than help. 

  Domain Naming System (or DNS) 

The domain naming system was established so that the common user can understand in simple terms the location of a web site. A web site's files are usually stored on a server that points to a specific IP address (much like a phone number directs someone's call to your phone). In order for the general public to understand where to locate a certain website and it's files, the specific domain name resolves to that particular IP address. In addition, the Domain Name System also stores other types of information, such as the list of mail servers that accept email for a given domain (such as you@yourdomain.com). 

Top-level Domain Extension 

The domain extension originally consisted of the generic gov, edu, com, mil, and org. With the growth of the internet, the addition of country extensions and other such categories have come into play. The most recognized of the extensions is of course the .com. If you are optimizing for a specific country and language, then the best route to take is to register your domain with that specific country's extension. This will help the search engines recognize that you are targeting that particular audience and will rank that site accordingly. Be sure that your country specific site is in the native language for that country to avoid any duplicate content issues. Do also be careful of linking from that domain to your main domain as once again the site will be penalized. 

Directories and Files 

Here's where the fun stuff comes into play. Just as your computer organizes your word doc, excel, and other such files into folders, a server structures your website files in the same way. A "directory" or "path" is much like a "folder" is on your computer. In standard (old school) html development (before the days of creating dynamic websites powered by databases and user interactivity) a file would be created and named "index.html" or "default.html" and placed either on the main domain folder (in which the DNS resolves to on the server) or placed in a named folder (in order to help the webmaster organize the site's files). As the technology grew and more ways to develop websites with user interactivity and database driven websites advanced, the structure has pretty much stayed the same with the addition of "parameters" that reference a part of the database and returns content and such on a page based on those parameters. (have I lost you yet?) Let's go back to the basic structure of the static html files and go from there...

A Dynamic website is one that has a few static pages (in other words the pages are coded and are only editable by a developer) that have parameters that will pull in content or trigger specific actions from a database. The basics of a dynamic page is one that pulls words, images, etc from a database and can do so creating multiple pages with different content from one basic page. A more complex dynamic page (or site) is something like Facebook, or Twitter in which they recognize whether or not you are signed in with a username and password and will show you either your profile page (if you are signed in) or a "please sign up" page (if you are not signed in or don't have an established username).

In order to help understand this let's talk about how a database works. A database is essentially similar to that of an excel spreadsheet or table in a word document that has a unique identifier for each line (or row) and holds different content for each line item. Example:

	Email	First Name	Last Name
Sujo234	bob@bobsemail.com	Bob	Sujo
Forjill23	jill@jillsemail.com	Jill	Forman

Username

In this example the username is the unique identifier with the email, first name, and last name as different parameters for that username.

The content will be different on each page. With dynamic content the possibilities are endless as far as how many pages you can create from developing and design just one file. A great example of how a dynamic page is created for search optimization purposes is on usedcars.com - If you search for "used cars in oslo mn" you see the "UsedCars.com Oslo MN" page in the results. Look at the URL in the address bar when you go to that particular page - http://www.usedcars.com/browse/mn-24/oslo-163.aspx. In this case the page is pulling in the unique ID that is equal to "OSLO 163" and "MN 24", just as the username is the unique ID in the above table.  

SEO Friendly URL 

In order to make your dynamic URL friendly for search engines you must use a rewrite. A great resource for rewriting a URL is the Apache Rewriting Guide. Some open source content management systems (such as Wordpress, Drupal, etc) already do the rewriting for you and all you have to do is enter what you want the URL to be (be sure to include your key terms separated with dashes "-" and not underscores "_" for search happiness) Who would have thought a URL could be so complicated? But when it comes to search optimization and understanding basic website development it is very important to understand how the URL works, how it is structured, and how to make sure your site is URL and search engine compliant. *http://en.wikipedia.org/wiki/Http_protocol

Facetwitetiquette - How to Suck at Facebook and Twitter

Facebook

What not to do on Facebook

	The constant Gardener, Mafia Hitman, Virtual Pet Owner, or Whatever the updates and invites never end...
	The Bad Marketer become a fan, join my group, comment on one of my million updates...
	The Crude Photo Tagger tagging everything from that party last night you don't remember to taking shots off of a girls belly button.
	The Rash following you around commenting on all your posts and liking every photo...
	The Unfiltered they post everything and anything...
	The Most Popular Person EVAR inviting everyone to everything...
	The Twitterfied linking twitter updates to Facebook flooding their profile with meaningless tweets...
	The Bored Quizzer taking every quiz available...
	The Passive Aggressor posting well thought out updates without mentioning names...
	The Annoyingly Proud Parent using their child as their profile pic...

Twitter

What not to do on Twitter

	The Unproductive Tweeter tweets updating every minute...
	The Retweeter retweeting everything they see...
	The Conversationalist bouncing back and forth between one person...
	The Untweeter Tweeting from Foursquare, tweetmeme, or the like...