ShopMyLikes Shopping Sites

I used to work for Pronto.com, which is/was a shopping search and price comparison site. In the black-arts of search engine optimization (SEO) and the derived exploitation strategies that were bandied about at the time (circa 2009), Pronto.com decided to launch a handle of "vertical" sites, which meant just focusing on a particular category, such as fashion or electronics. The idea was that more sites and more links and everything interconnected could drive more traffic. Secondary to this was tailoring the web site appearance to appeal to the particular demographic which that category would attract.

Being somewhat dubious of this as a strategy, and being somewhat dubious of the SEO "best practices", I decided to play around with all the various SEO ideas to see what I might learn. Using Pronto's API, I built a series of 12 different web sites, trying to minimize custom coding, and make it very easy to roll out new "verticals".

One other feature that I was playing around with was the idea of searching over user-selected "favorites", rather than a full product index with brands and stores the user might never be interested in.

These sites no longer work, but there are screen shots here to show what this looked like.

The table below shows the vertical sites name/logo, what its theme was and the inspiration for the color scheme. Click on the logo to see a static image of what the site's home page looked like.

Each site was based on a theme, and within the theme there were categories and within those specific items. For instance, for ShopMyTeams, the theme was sports, the categories were the types of sports and the items would be specific sports teams. Each themed domain used a semi-manually curated set of items that provided the items in the index which allowed browsing the site. The items were selected as those that had a least a handful of products matching that item in Pronto's product index.

I built the site using Google's AppEngine, and in particular, wrote the code in Python. The product results were driven by Pronto's API, which also happened to be something I created (designed, developed, tested, deployed and maintained). I wrote the code for these entirely in my free time at night and on weekends. I made no money off these sites personally, though I probably did drive a little traffic to Pronto.com during the period it was working. Not likely more than a few tens of dollars total though.

I used Google Analytics and their Web Master tools to track things directly, and I also had access to the Pronto API log files and reports to see what the traffic looked like from that end. After a while I put Google's AdSense ads on the pages to see if that would generate any revenue. In total that yielded less than $20, none of which I actually collected due to Google needing a minimum of $100 to cut you a check.

For a time I had 12 different Internet domains, one for each site and the umbrella domain shopmylikes.com. After a year, having not seen that much traffic, I did not renew all the domains and moved the unrenewed ones to be sub-domains of shopmylikes.com. I kept a few of the better ones like shopmyteams.com, shopmylikes.com and shopmystyles.com. A year or two later, I shut the Google App Engine hosting down and now only retained a few of the domains.

I am not terribly good at coming up with color schemes for web sites, and with my needing 12 of them and wanting them to be different, I needed a strategy. I decided to "borrow" the color schemes from NFL teams. I found a web page that gave me the RGB colors for all the NFL teams, and I picked form those to define the color scheme. The mapping from web site to NFL team is shown in the table above.

Among the various details I used for SEO and web "best practices:

• carefully constructed URL paths and organizations
• XML site maps (on site and uploaded to search engines)
• Use of HTML canonical meta tag
• Automated content generation from templates to provide unique content.
• Index pages to aid crawling