In previous posts about getting traction for my Bandcamp recommendations project (BCRecommender), I mentioned search engine optimisation (SEO) as one of the promising traction channels. Unfortunately, early efforts yielded negligible traffic – most new visitors came from referrals from blogs and Twitter. It turns out that the problem was not showing up for the SEO game: most of BCRecommender’s pages were blocked for crawling via robots.txt because I was worried that search engines (=Google) would penalise the website for thin/duplicate content.
Recently, I beefed up most of the pages, created a sitemap, and removed most pages from robots.txt. This resulted in a significant increase in traffic, as illustrated by the above graph. The number of organic impressions went up from less than ten per day to over a thousand. This is expected to go up even further, as only about 10% of pages are indexed. In addition, some traffic went to my staging site because it wasn’t blocked from crawling (I had to set up a new staging site that is password-protected and add a redirect from the old site to the production site – a bit annoying but I couldn’t find a better solution).
I hope Google won’t suddenly decide that BCRecommender content is not valuable or too thin. The content is automatically generated, which is “bad”, but it doesn’t “consist of paragraphs of random text that make no sense to the reader but which may contain search keywords”. As a (completely unbiased) user, I think it is valuable to find similar albums when searching for an album you like – an example that represents the majority of people that click through to BCRecommender. Judging from the main engagement measure I’m using (time spent on site), a good number of these people are happy with what they find.
More updates to come in the future. For now, my conclusion is: thin content is better than no content, as long as it’s relevant to what people are searching for and provides real value.