This is an oldie but a goody. We all know the 30% / 70% rule of on-site vs off-site SEO. We talk about every day internally. And it comes up with seo reseller demonstrations regularly. I've also been blogging about onsite and offsite SEO for years (see also here. It comes up every year in the SEOMoz survey of ranking factors (see my interpretations here). Today I happened by an old blog that I maintain and the data told the story so well, I thought I should post it. Let me explain.
Today's data is for a blog that I was regularly posting fresh content for over a year. I was, of course, also building links to this blog. The search engine rankings were moving and all was good. Well, somewhere back a few months (April, according to my records) I fell off the on-site SEO wagon and stopped posting content. The link building continued in earnest, but the blog got good and crusty as nothing new went on the website for over 90 days. Take a look at what happened to the rankings in the screen shot from the Semify SEM Reseller Portal below:

Take a close look at the thumbnail graphs above - with the big red arrows. I am not showing you the actual keywords, but they are not necessary. Those thumbnails are a small version of our search engine rankings time series graph that is on several screens in our system. Pretty much without exception, we lost ground on search engine rankings here after a long period of inactivity on the blog. Here is a close-up of one of the keyword terms. Notice how it dropped in nicely (read: moved UP the search engine results) and then how we blew it completely by not posting for 90 days. Then the trend reversed itself sharply.
So if you ever wanted proof-positive that on-site content needs to be updated at least at some regularly interval, here is a great little experiment. Of course, every good SEO strategy review tells you that you need to have high quality content as part of your link building strategy. I have many blogs where I didn't fall down on the job, and their thumbnail SERP graphs look much better than this. This doesn't mean that you need to be blogging every day. But I think at least once every two weeks is required to avoid turning into this case study.