A few days ago, I offered to perform a SEO analysis on the website of one lucky winner– something at BlitzLocal that we normally charge between $6k-10k to do. There were 23 responses and I’m choosing BusSongs.com, which is by Keith Mander, a current Facebook employee and ex-Googler (not to be confused with the other Keith that blogs here).
Let’s first start with an assumed goal of the site– to make money from ads, as there are no products to be found. Keith is using 4 cleverly-placed Google AdSense units on each page, in addition to serving ads via Google Ad Manager (GAM)– a product that kills OpenAds and will be merged into DART DoubleClick (that’s the subject of another post). Notice how the links on the left blend in well with the orange.
The site has a Google Toolbar PageRank of 4 and a MozRank of 4.58– moderate juice is flowing to the site. The MozRank, as developed by SEOMoz is a more accurate view of link juice that is flowing, as the toolbar PR is rarely updated, plus there’s a huge difference between a low 4 and a high 4.
This nice level of juice flows through the rest of the site nicely, allowing 3,130 pages to be reported indexed by Google, such that even lower level pages are getting crawled and are ranking. Site that have a low homepage PR peter out quickly– there’s not enough juice left by the time the bot gets to the pages that are 3-4 links away from the homepage, so they don’t get indexed. To validate, just go to one of the lower level pages, grab a paragraph of text (maybe 15-20 words) and paste the whole thing right into the search box. That will let you know what’s being indexed.
Also try some of the terms the site wants to rank on. In this case, I searched on “99 bottles of pop on the wall” and see his site taking the first position.
Of course, search on just the domain and you see him first– if you’re not first on your own name, something is quite wrong or you have a generic name.
Not only does Keith rank #1 on his name, but he has 8 sitelinks, the maximum number of sitelinks you can have. While you can’t choose which links are sitelinks, it’s great to have them anyway. You have to be in the #1 spot for a search and also have enough “authority”.
I’d guess that Keith wants to rank on “children’s songs”, as that is the first searchphrase in his home page title. He’s #2 from my search here in the US, and the #1 result is PR5. Let’s go to SEOmoz’s LinkScape tool (requires a subscription, but well worth it) to dig deeper….
The #1 result has a higher domain mozRank (5.18 vs 4.31) and higher mozTrust (5.58 vs 4.49). They have 4,354 inbound links versus 848 on bussongs.com
. It’s true that quality is more important than quantity of links. In this case, the guy above Keith also has higher trust (more juice garnered from high-trust sites), so Keith’s better content doesn’t win the day. Rand Fishkin, CEO of seoMOZ, noted that he could create a crappy entry on wikipedia and an amazingly helpful article on a new domain– and the next day the wikipedia article will win.
Not fair, but says something about the power of inbound links.You can still beat guys that have more juice than you overall by selectively picking terms you want to rank on. If you search on “nursery rhymes”, you’ll see a completely different set of results than for “children’s songs”.
While some SEO pundits like to wax on about LSI (Latent Semantic Indexing) and how search engine theming can help you rank on terms that you don’t even have on your page– the more practical reality is that you want to have these terms on your page and for anchor text in sites that link to you. Note that in the above seoMOZ LinkScape report, that the #1 anchor text is “bus songs”. No surprise there, but “nursery rhymes” is #2 and “children’s songs” isn’t until #10, passing a piddly amount of juice from only a few domains.
Thus, Keith will have to decide whether it’s more important to go from #2 to to #1 on “children’s song” or try to get to the first page on “nursery rhymes”. It’s a question of big dwarf or little giant– which is bigger? Let’s find out how much volume is available:
For every 19 searches on “children’s songs”, there are 68 searches on “nursery rhymes”. Further, for every 19 searches on “children’s song”, there are 100 searches on “nursery songs”– the term you’d want to own if it were no extra effort. Note that Google’s Insights for Search tool doesn’t tell you the exact number of searches on each term– rather, they give you a relative figure, with the most popular term being indexed at 100 and every other term scaled against that term.
So how do you decide what term to go after? Let’s say that I was ranked #8 on “nursery rhymes” and #2 on “children’s songs”– good rankings on a highly popular term and great rankings on an okay term. Moving from #8 to #3 on the popular term would produce about as much increase as going from #2 to #1 on the okay term. As you get towards the top of the page, your CTR will go way up. I wouldn’t be surprised if moving from #2 to #1 yielded a 3x increase in clicks.
Of course, you wouldn’t do this in a spammy way, where overnight all your inbound links suddenly have identical anchor text of “children’s songs”. But you could kindly ask the 70 sites who gave you 81 links with anchor of “bus songs” to switch to something else. If you add “bus songs” to the list in Google Insights for Search, you’ll see it has a paltry 4 versus the 100 for “nursery songs”. I doubt you’d lose the #1 ranking on your name, largely because you get a boost from that being your domain, it’s not that competitive, and so many folks have already linked to you on that phrase.
Oh, and there are 26 inbound links that have BLANK anchor text– probably want to do something about that.
Finally, let’s take a look at bussongs.com through the eyes of a search engine spider, which reads text, not images. It looks like this:
- The 32% text to code ratio is excellent– we like to see over 25%.
- The 301 redirects from the www homepage and index.php to non-www is smart– it solves the most common SEO problem, called the canonical domain issue. Most people redirect to www, but as long as you choose one, it doesn’t matter.
- missing meta information– you should at least have the meta description, since you’ll want to persuade the engines to use your description when your results show up in SERPs (Search Engine Results Pages). Don’t worry about other meta tags– keywords, gaming H1 tags, and so forth.
- The alt text on images does very little, but while you’re at it, you should put your domain name at the END versus the beginning– change:
“BusSongs.com – Lyrics & Words for Children’s Nursery Rymes & Songs”
“Lyrics & Words for Children’s Nursery Rymes & Songs | BusSongs.com”
One word of caution, a few months ago, Google started changing search results to biased by whether you’re logged in, where you are geographically, and what you’ve searched on before. Thus, check your rankings when you’re NOT logged in and also via proxies. Every is getting different search results, so you don’t want to be led down the garden path.
And a few non-SEO items
- Funny that the #9 song is the Diarrhea Song– Kids…. what a sense of humor.
- If kids (and parents) like the site so much, where is your email auto-responder and newsletter subscription box?
- You should do the same on your Facebook page, which has 550 fans (of which I am one). Use the Facebook static HTML plug-in to put in that email box, a poll, and other interactive stuff.
- Maybe even install Facebook connect and Facebook Fan boxes– you do work at Facebook now, right?
- Leaderboards are a powerful concept– People are driven to do silly things in the name of popularity. Why not allow folks to submit songs, earn points as part of a community (provided they are old enough), and have “name that tune” games? On your top visited page, you show pageviews per day. If you shown total cumulative pageviews, the numbers would appear a lot more impressive.
- If I can nit-pick, you have some typos. “Angles” should be “Angels” here in the page title and text. By the way, I did that as an excuse to give you another PR5 link from my blog.
Keith, I hope you have enjoyed our SEO review today– congratulations on winning! You have a great site, as we’d expect from a former Google employee.
Readers, I’m considering making this a weekly review, so if you’re interested in seeing more of this, let me know by posting to my Facebook fan page at facebook.com/dennisyu.