Data Analysis When Buying Websites: “The Funny Looking Kid”
One of my undergrad finance professors was discussing methods of analyzing corporate financial statements when he summarized one technique as spotting “the funny looking kid” in the crowd. Sometimes when looking at data, you may not know exactly what may be wrong, if anything, but you have to dig somewhere.
As I’m currently on the hunt for this month’s website purchases, I am asking the usual questions to the website sellers when it comes to “funny looking kids”. I may be missing a piece of information, it may be just some unexplainable phenomenon, or there may be a story behind why a certain piece of data stood out from the crowd. Maybe that story will make me move on to the next prospect, maybe not.
Here is a current example:
This is volume data provided by the website seller for the month of March. See the Funny Looking kid? It’s pretty obvious that March 8th was an unusual day. The tougher thing to determine is why? It may be something, it may not. Probably just one of those internet traffic flukes we all experience…an unexplainable pop or drop. No big deal…just another issue to be aware of when buying websites.
Let’s take a look at the adsense data provided by the same website seller, for the same month:
Your typical, unpredictable, unreliable, high-octane variance here. Classic Adsense data. But you know what is unusual? As strange as adsense data can be, I would at least expect to see a bump in page impressions on March 8th to correlate with the huge March 8th traffic day shown above. Wouldn’t you?
The first thing that comes to mind: What if this data isn’t real? What if the adsense is from another site? Or the traffic? What if it’s photoshopped? When you are laying your money on the line, it’s good to be suspicious. Hopefully there is a reasonable explanation (I’ve asked and haven’t heard back) but who knows.
There is a reason why the ROI is quite high on buying website investments relative to other investments. It’s pretty damn risky stuff. Even if you have lots of good historical traffic and revenue data…and you are comfortable with the source of the traffic and the accuracy of it all….it could still just evaporate one day, and you have no control over that.
Until next time…
