The following notes on SEO were taken at the author’s visit to SMX West 2010.
Diagnosing Technical SEO Issues
Adam Audette (Audette Media) was the first one to speak. Adam spoke about SEO site audits, and reminded us that they should be a collaborative process.
SEO Site Audits
When multiple people are involved in a site audit, it can get confusing. It’s very important that everyone involved is on the same page; otherwise, you may find that you are giving your client conflicting advice/suggestions in the site audit. For this reason, he suggested the use of a project management tool like Basecamp.
Everyone does site audit’s differently, but they should all include the same core elements. Adam identified these as being…
- On-Site: Domain(s), navigation, sections and categories, pages, and media (images/video/etc)
- Off-Site: Backlinks (quantity, quality, frequency), social media signals, cache dates/crawl frequency/indexed pages, and toolbar pagerank
After identifying the core elements of an SEO site audit, he said that the “big 4 factors” are URLs, site architecture and navigation, deep pages (pagerank dispersion), and site latency.
Gabe Gayheart (Razorfish) took the stage next, and continued the discussion on site audits.
Gabe recommended that you do a full site crawl and identify each page on the site. He also suggested doing “search engine simulation crawls”, which are individual crawls where you simulate each major search engine, and “browser crawls” where you view how the site is being rendered in different browsers. I don’t recall Gabe mentioning this tool, but for the browser crawls, I would highly recommend using a tool like browsershots.
When you’re finishing up a site audit, how do you prioritize the recommendations that you’re making? Gabe recommended that you prioritize in terms of impact (how will this effect the client’s business?), ease of implementation (how difficult will these recommendations be to execute?) and readiness (how quickly will the client be able to implement these changes?).
SEO Tips For Images
Brian Ussery (Search Discovery) was next. He went over some of the things that you can do to make sure that your images do well in Google Image Search:
Use good filenames. If your file is titled 120483.jpg, how is Google supposed to know that it’s a picture of a dog? Call it fluffy-white-pomeranian.jpg instead. Make sure that you’re using descriptive alt text/anchor text (when applicable).
Don’t put text in images. The search engines aren’t going to read the text in your images… I may be speaking too soon on that one, but regardless, you’d be better off with placing the text right next to (or below) your image. It’s a good practice to provide information on your images anyway, because the text can be used to determine what the image might be about, and that’s more copy for your page.
Specify image width and height. This one was news to me. The browsers I’ve used have always displayed an image properly regardless of the inclusion of width and height, but I’d imagine that there is a benefit to including this if he mentioned it.
Provide as much meta data as possible. Data such as EXIF, tags/labeling, and location info can be used to help your image perform well.
He also mentioned using a favicon with expiration to avoid 404 errors. Using a favicon is a good idea for obvious reasons… But did you know that you’re generating 404 (not found) errors each time the favicon is requested if you don’t have one?
CSS Sprites
Patrick Bennett (ModernBlue) was the last to speak. One of the things that he mentioned was CSS sprites. I had heard of these being used before, but have yet to try and implement them myself.
How do CSS sprites work? Instead of having 80 different images for your site, you can put all of the images into 1 “master” image, and then use CSS to display the parts of the image where necessary. That’s a very brief description - For more information on CSS Sprites, check out this article from A List Apart.
Domain Names, Parameters, URLs, and All That Jazz
The second presentation on my schedule was another technical one. Google’s Maile Ohye was the first one to speak, and the only one that I found myself taking notes from.
She cleared some confusion on ccTLDs (country code top level domains) when she revealed that they are automatically geotargeted by Google. So if you own a .fr domain, Google will know that your site is targeting people in France. If you don’t own a ccTLD, you can manually geotarget your site in Webmaster Tools. If you do this manually, it takes about a week to go into effect.
Industrial Strength SEO
I wasn’t sure what to expect from an “industrial strength” SEO presentation, but I liked the name. It actually ended up being about SEO for large websites.
Marshall Simmonds (Define Search Strategies) spoke about his work with The New York Times, and the challenges that he was faced with while working there… 22 million documents, a subscription “wall”, rigid IT, an enormous company ego, resistance to change/getting them to “get” search, a limited CMS, massive duplicate content, and crawl barriers were just some of the obstacles that he mentioned.
Marshall talked about the window of opportunity for content creators when breaking news happens. When the US Airways flight landed in the Hudson River last year, that window of opportunity was about two hours long. In that time, you had to do keyword research, come up with good content, and do whatever you needed to do to get your content published and ready so that it could be indexed by the search engines.
The New York Times missed out on that opportunity, because they didn’t do the necessary keyword research. Their headline used the word “land”. While technically the plane did land in the Hudson, everyone was searching for things like “plane crash hudson”. This is an example of a dilemma where a company has to choose between editorial integrity and organic search traffic.
-John Vantine
Originally from Wpromote, Inc. PPC Management. Cross-Posted with permission.
