I don't see how you could tell if this worked unless you had 2 stores with almost identical items. 

 

One thing I always find is that the the daily traffic numbers are not updated regularly when sales are down. I used to avereage 100-150 page views per day, now 50-100 per day.

Wouldn't you be better to do it on just a few of your listings so you can see if those particular ones sell more than the others, eg. just do one or two pages of your Active selling list?

Almost certainly it's just another data point, not a magic bullet.

 

What I mean is that the cassini search engine will look at many things to decide how to rank your items, and the best result will be from ticking as many boxes as possible. Some examples:

 

Feedback score

Shipping options such as express

Refund policy

Stated handling time

Number of photos in a listing

Number of characters in a title

Number of sales for an item

Number of sales for a seller

Number of similar listings

Number of buyer-seller interactions (messages)

Response to and outcome of refunds/cases

 

The list goes on. What we will never know unless there is either a code leak or a deepthroat type person who leaks the details is how the different data points are weighted. And if that happened, ebay would change the algorithums to prevent everyone gaming the system. Since you had just come back from holidays, your listings had probably been flagged as stagnant, and a blank edit gave you a temporary boost. I recently did real edits on all of my listings to update postage information, and it made very little difference. There was a boost, but it was within statistical anomoly, so it could have been a coincidence.

 

The point is that your best bet is to make the best listings you can, and put some time into ebay every day.

I just wanted to add that the google search engine works in a similar way, comparing many data points to decide ranking. Ebay have basically copied their idea because it has proven effective.

Sorry, just a correction, I meant to say standard deviation, not statistical anomaly.