In the past weeks I have observed some strange Technorati search results. When I search for links to my weblog, about a third of the times Technorati shows a first search result that doesn’t make any sense. It is always only the first returned result that is wrong. It shows a website that contains no link to my site, and the highlighted link is to another site altogether. I have added two examples from the past weeks. Anyone else seen this behaviour?
In recent days I saw two sudden drops in the number of sites which Technorati says link to me: from 157 to 111. The number of linking sites disappearing seems roughly in-line with new links (not necessarily new linking sites) being added at the front. The total number of links has dropped slightly, not much, and certainly not as dramatically as the number of linking sites.
This morning I took some time looking into it: Technorati seems to only take into account the first 199 links of the 267 links they report. Those 199 links indeed come from 111 sites. But older stuff is unreachable, and results in the already all too familiar Sorry, we couldn’t complete your search because we’re experiencing a high volume of requests right now. Please try again in a minute or add this search to your watchlist to track conversation. So I cannot reach stuff in Technorati that is older than 866 days or 29 months at the moment.
Every time a new page with new links appears at the front, a page of 20 old links is dumped so it seems. Those are mostly links from blog rolls from when I started blogging 3 years ago and that are still in place, so they hit my stats on how many sites are linking more than they hit the total number of links.
Now the number of links is not very relevant by itself, but that to me it looks as if Technorati is loosing data at the back door sucks.
Anybody else experiencing this?
[UPDATE] David Sifry is quick in pointing out in the comments that Technorati is NOT loosing data at the back door, and are working hard on improving their service. Certainly David’s respons time to my posting (90 minutes) is fast!
[UPDATE 2] The drop in sites linking to me as reported by Technorati is caused by a new way of evaluation. Instead of counting the total of all linking sites in the now almost 3 years of Technorati’s existence, they now look at the figures of the past 6 months. That favors dynamics in blogs more than seniority, which figures given the nature of blogs, doesn’t it?