I’m no fan of EBAY, and in fact, with AMAZON Kindle, I’m probably more interested in what I can do on Amazon.com than anything that Ebay.com has to offer me; but I didn’t care for the blog post in the New York Times Bits blog – EBay’s Traffic Drops Amid Identity Crisis where it was written that …..

….. But eBay’s traffic began to decline sharply last fall, and it dropped below that of Amazon in November, based on numbers from Nielsen.

Aside from the Nielsen numbers, which don’t appear to be substantiated anywhere else, EBay.com traffic is still well above Amazon, in the US, and Worldwide.   That shows just how misleading numbers can be – and the report cited by Ina Steiner at AuctionBytes.com appears to only look at Nielsen (see below)

Look at Comscore, where the blue line corresponds to Ebay and the Yellow to Amazon.

Compete.com, where I have a Pro account and can see 2 years of data, doesn’t show Amazon ever having more traffic than EBay.

Alexa doesn’t show Unique Visitors – so I grabbed pageviews, instead

I didn’t look at the other metrics but what I got out of this – someone already knows they that EBay is losing ground and just looking for whatever will support it – instead of just looking at enough information to make an “informed” decision.

Personally, like I said earlier, I don’t that much which of these sites is leading – but studies like the one in AuctionBytes.com need to be taken with a grain of salt.   I don’t do posts like this that often anymore, because it’s obvious that the data will differ between Nielsen, Comscore, Compete, Alexa, Site Analytics, etc – and while it’s necessary to chose which version of the “truth” to follow – it’s also important to be fair in looking at this data and to find support of it in other numbers (in this case, outside Nielsen – which wasn’t done).

My approach to competitive analysis – is to find at least two, but better to have 3 sources of data that, more or less agree, in this case – I’d have wanted Comscore, Nielsen and Compete to show the same thing, more or less, even if the numbers differed.

What do you do when they don’t agree – withhold judgment while looking for at least one additional source that shows the same basic trend.