Over the past year we have been interested in our website statistics and our growing traffic. I have blogged previously about Alexa and was challenged to review the Compete statistics. After growing in rankings for a few weeks we removed the Alexa widget and saw our rankings plummet. We then installed the Compete widget and saw ourselves go up the rankings quite dramatically before removing that widget and seeing our rankings decrease. Meanwhile, our own website statistics have shown consistent month to month growth with an average of about 4000 unique users per day at present (As shown in the figure below).

stats.png

Bottom line, based on our observations, neither Alexa nor Compete give anywhere near valid statistics. At the SBS conference in St Louis this past week I asked the audience, about 60 people, how many in the room knew of or had heard of ChemSpider. ONE hand went up…and that was someone I had informed many months earlier. As I expressed to the audience…this was not disappointing news to me…it was quite exciting to know what the potential growth is as people are informed of the service. I expect the growth to continue, especially after the visits to the ACS and the SBS.

Stumble it!

4 Responses to “Alexa, Compete and Website Statistics”

  1. David Bradley says:

    It’s been quite well known among SEOs for quite some time that analytics such as Alexa and Compete are generally not accurate ways of measuring the success of a site. Outside, SEO circles though, no one really worries about them. Google Analytics or a similar tool is really the only way to go, however, presenting those validated stats publicly is not as viable as displaying google pagerank, compete or alexa values.

    db

  2. Chris Singleton says:

    Will Google Analytics enable us to see which users are accessing which molecules? If there are several that are queried frequently, then we may push some of our deposition efforts towards those. If there is a readily available chemical that is looked up a great deal then we could get the real spectra for that compound, and maybe provide a bit more info than I would normally deposit for an ‘average’ compound. I still have access to an NMR and a high res MS for some stuff, as well as several common chemicals, so that might be worth checking out.

  3. Antony Williams says:

    We did add Google Analytics for awhile but it was causing us issues with some of its own bugs as well as giving us high traffic load that was adding to server burden. It may be time to go back and revisit the use of the system now we are out of beta. I’ve examined the stats in a rudimentary way and don’t see any specific patterns that are obvious…they may be there but I just don’t see them. The spikes such as “Xanax” that I see are there because it’s one of my examples…same for Taxol and Quesnoin. If you have common chemicals that you want to run anyways I say go for it!

  4. David Bradley says:

    It’s interesting you should mention server load with Google Anal, I ended up having to switch host for one of my sites because some script or other was overloading the server and the host wanted me to upgrade without explanation. It could have been GA…never occurred to me that they’d be leeching so much bandwidth.

    db

Leave a Reply