shopping.com clicks part III – ROI tracker

Based on the sage advice from our account rep, we decided to implement the ROI tracker.  If anything, it would be interesting to see the data they display.  Of course, we haven’t forgotten about our original mission of overbilled clicks.

So, the ROI tracker consolidates some info that is pretty helpful.  It shows the clicks by product, conversion and your cost per order.  I could have already gotten that info already, but it was nice to have shopping.com make that available.  The info that was really interesting, that I couldn’t get was the sitewide stuff.  Under the competitive Analysis tab of the Reporting, there was a column for ‘Sitewide Clicks’ and ‘Your Clicks’.  That is useful because it lets me know where I stand with my products vs competitors.  Based on our Amazon metrics, I know that we complete with merchants on about half our skus, so based on where we want to be in bidding, I may get 50-75% of the sitewide clicks.  If we bid higher, we’d probably get a higher %, if we bid lower, then we’d probably just get clicks on the 50% of products where we have no competitors.

In optimizing our account, we decided to cap our spending at $500 per analysis segment.  Once we hit $500, our account shuts down and we are able to have a clear closing period for our analysis.  Also, when we start it up, we can monitor how fast traffic comes back to those urls.  So, for our $500 spend in Nov, we got 575 clicks.  But, the sitewide click count was 379!  How can we possibly be charged for 575 clicks if the shopping.com network only generated 379 clicks for those skus?  Our weblogs recorded 318 clicks during the period, which is definitely closer to the sitewide number than the ‘Your Clicks’ number.

So, of course, I contact shopping.com to “understand” these numbers better.  It turns out that there is a known bug.  I agree, however, shopping.com thinks that the bug is with the sitewide clicks number; but I think the bug is with the Your Clicks number.  Of course, shopping.com doesn’t want to lower the Your Clicks stat because that’s how they get paid.  Here’s the comment direct from shopping.com:

“With regards to the ”your clicks” and ”sitewide clicks” discrepancy this is a ”known bug.”  The sitewide click column is not being captured accuratly and is only intended for a directional purposes.  The number of clicks you are being charged are the “your clicks” column.”

I figured that out, but I wasn’t going to let it slide this time.  Here’s their follow up:

“Our system and algorithum counts clicks differently from your internal tool and it might be not attributing clicks to the entire Shopping.com Network.  The bug is on the actual number being calculated for the column “sitewide clicks”.  We currently do not have an ETA as to when this will be fixed, however we are hoping to have it resolved soon. 

I understand where you’re coming from however you always have the option to pause your campaign within the Merchant Account Center.  I have informed my colleague to reach out to you once the shoe category is enabled for the Rev Share Program (cart) as I feel this might be a better opportunity for you.”

So, it was suggested that we stop using shopping.com until the CPA program is live in my category.  Maybe sometime in late 2008.  However, I am using basic weblog info to record clicks.  Afterall, the internet has been around long enough to be able to record clicks to a webserver correctly.  If shopping.com was really interested in solving this issue, they would have shared the ip and click info for MY ACCOUNT only with me so I could see what was happening.  I have no choice but to think that there is massive overbilling for fictitious or fraudulent clicks and it will continue to go on until they get sued or enough merchants leave the platform.

So, in conjunction with my new blog, I kicked off a new test of shopping.com.  We had our account turned off for most of December.  I must be nuts to shut off one of the top sites during the holiday, but the traffic was just that bad.  I turned the account back on January 1.  I have also been recording our web traffic each day.  I plan to post a running log count for each day of January to see how the traffic is shaping up. 

I am curious if I’m the only one who has had this experience, or are there other merchants that have seen this.

A conversation with shopping.com about clicks

To continue my earlier post, I was able to schedule a conference call with an account manager and her supervisor to review our clicks issue.  In response to their request, I forwarded our web logs showing the shopping.com traffic so they could identify the clicks that never got to our webservers.  I asked them if they could give me a list of the clicks with the time and ip address.  I wanted to match that traffic to my weblogs to see if I could find a pattern of which clicks aren’t getting to us.  They declined.  I figured they would.  They said they would research the clicks issue and get back to me.

After 3 weeks, I decided to follow up with an email to see what was going on.  Here is part of the reply:

“Hi Jonathan,

I apologize for the long delay!  I hope you find yourself well! 

I have finished investigating the traffic discrepancy between your report and ours. 

The reason for this discrepancy is due to Shopping.com not tracking by unique click throughs and more than likely your tracking system is not capturing the traffic from the entire Shopping.com network. 

Please let me know if I could be of further assistance to you.”

That’s not a very helpful answer is it? We submitted a report to shopping.com that included total clicks and unique clicks by ip address.  Our system was purposely setup to record this stuff so that shopping.com couldn’t give us this type of answer.  So, I followed up with a phone call and explained that we submit a very specific, unique url with our products.  I asked shopping.com if that url is used for all clicks.  They said yes.  So, since our weblogs are recording all the traffic to that url, we must be capturing a record of all traffic from the shopping.com network, right?  Hmmm, she acknowledged that it made sense that we were recording all the traffic.  I followed up with an email and here’s part of the reply:

 ”With regards to unique user and the clicks – Shopping.com does not charge you for the users visit in which happened within minutes and seconds of their previous click or clicks.  However if they return 2, 4 hours or days later you are charged for that click and again if they click once again within minutes or seconds you will only be charged once. 

The URL that is sent within your feed is the one used for the entire network.  Our list of partners has been definitely ramping up as is Value Based Pricing.  Shopping.com is highly aware that some of our partners are sending low quality traffic and as a result which is why Value Based Pricing was released and launched, which gives merchants either a discount for that low quality traffic or you’re not charged at all depending on the partners conversion ratio.  this moves our partners to work on driving quality traffic to us.  The way you’ll start seeing this help you in your campaign is when you see that your average CPC has decreased. 

The best way to track your campaign progress is to implement the Shopping.com ROI Tracker as this will definitely track all the incoming traffic and can potentially increase your discount on value based pricing as merchants this already installed are seeing a bigger impact on their campaign vs. the ones that don’t.  I have attached the ROI Tracker user guide for your convenience.  If this is something you need help with let me know and our technical support can definitely help you in implementing.”

So, it sounds like I’m capturing the traffic correctly.  Now, I’m hooked as far as figuring this thing out.  My company submits feeds to a half dozen different engines and shopping.com is the worst performing one, by far!  I wasn’t excited about implementing another ROI tracker, but I had to walk down this path, if only to take another excuse away from shopping.com.  Let’s see where that takes us.

Are shopping.com click numbers wrong?

I am an online merchant that uses a number of channels to drive traffic.  One of them is shopping.com.  In the past six months, I’ve been working to optimize the ROI we’ve been getting on all the channels and the one that jumped out (badly) is shopping.com.  So, I turned to our weblogs to see what shopping.com traffic is doing when it hits our site.  To my surprise, I found that I was getting only half as many clicks as shopping.com was billing me for. 

My first stop was contacting shopping.com.  Of course, after getting ignored the first couple of times, I shut off our account and gave up on them.  But, as a coincidence, I ran into a shopping.com person at the PESA Summit in Australia and she said she’d get someone to contact me.  She followed up and I started a dialog with shopping.com to get to the bottom of the clicks issue.  Since its a long story, I’ll be posting it in pieces.  Stay tuned!