125 New Posts Today

[orfeo]

First of all, thanks everyone for the wealth of information you are sharing here. I have been reading the forum on and off for years, but this is the first time I am posting.

We have a fashion site based on datafeeds and we use product links through out the user experience.

We have a procedure where we test each and every product link before it is activated and periodically we test each product again.

This was a necessity since some of the links simply sent to the home page of the eshop or send to a page informing the visitor that the product is out of stock.

What I am asking is quite straight forward. Is there a way to test the links without messing up with our total clicks, our EPC and convertion stats?

The question goes for both CJ and Linkshare.

Another network (local one) gave as the option to use another affiliate ID when we do the testing.

Anybody has any experience on that regarding Linkshare (or CJ)?

thanks in advance.

[orfeo]

So nobody has any idea how to test products but avoid having it destroy your EPC?

[teezone]

Nope... each click will be registered to your account...

I found out the hard way last year when I ran a script to test 404s..

[orfeo]

We test various aspects (404, 500, out of stock, wrong prise, one size left) and we are very thorough with our testing. There should be a simple way to do a test.

For now we use subid tracking so at least we can see what is really happening and the real conversions. Problem is that merchants probably think we are the worst sellers ever

I will contact Linkwise and hope for a timely and meaningful answer (not very optimistic about it).

[gcarson]

Originally Posted by orfeo

Another network (local one) gave as the option to use another affiliate ID when we do the testing.

Interesting. So you can use like a subid that's just for testing purposes? I wonder how the networks feel about it though?

[orfeo]

I dont see why they should have any problem.

Instead of making 30.000 clicks per month from my account and make everyone look bad (me and the network) they gave me a way to make the clicks without getting any commission.

e.g.
the link was
http://www.affnetwork.test/click/affid1234/linkid1234/
and I use for testing
http://www.affnetwork.test/click/affid1/linkid1234/

no commission of course, but also no excess clicks showing up to my account.

[teezone]

I agree that it makes sense... no reason the network shouldn't accommodate. If you have a system or script in place that can adapt, it sounds great.

For me, I'm already stretched thin enough.. and can see accidentally mixing things up

Are you able to test a complete datafeed in a reasonable time..?

The reason I ask is that a developer pal built a script for me that tested response headers for product links, but I stopped using it as it took too long (or perhaps I was too impatient?!). And this is on a server that has decent capacity.

[orfeo]

Checking headers is completely useless.

Only situation it can help is when the network (Linkshare has done that) returns a 500 error or 404(more rare).

What we need to do is make sure that the product is in stock and it has enough sizes.

You can do that using scraping, but when you have around 100 eshops, this is more than a full time job.

I have developed a platform that makes it very quick for a user to check a product and give feedback to our backend. But still there are some of our people that do that every day for a couple of hours each.

I don't think this is reasonable for one person. I would go insane looking product after product for anything more than a couple of hours.

[Chuck Hamrick]

So nobody has any idea how to test products but avoid having it destroy your EPC?

Doing a half dozen test orders shouldn't kill your EPC unless it is fragile to begin with. You should be working directly with LinkShare tech to find a way not to effect the EPC. Or you can ask to have the orders and clicks cleared after testing, I have done that in the past.

[teezone]

I don't think this is reasonable for one person

Yes, I came to same conclusion myself.. good to know I'm not alone in that view!

The problem here, Chuck, is the networks are quite unresponsive (overall) to the issue of poor datafeed quality - and having to do extra work to facilitate an affiliate doesn't seem to be at the top of the list (even though it will actually help sales!). This is my biggest complaint - I even had one merchant tell me the GAN process to update his feed was "confusing". So it's a crap feed.. and has been for 6 months.

orfeo, I hope you get that extra account for testing from both CJ & Linkshare.. it does sound like the best solution here.

[Chuck Hamrick]

For some reason I started read at post #4 and thought this was a merchant. I now see the #1 post and understand the question better. At the last agency I worked we had a system that would check product pages for out or stock. They never made it commercial and the CTO who designed it left (to become an affiliate). Not sure what something like that would cost and if it would resolve the EPC issue.

[orfeo]

Hello Chuck,

I am talking about 100.000 tests per month. So EPC will go way down.

Unfortunately the merchants have proved that these tests are necessary, at least for out business.

There are 2 problems that made as move away from automated checking through some kind of scraping.

A. We are talking about A LOT of eshops. For now around 70. In six months probably around 200. Most of them, don't just say "Out of stock". Some do, some others have an image, some use a css class somewhere, some use complex JS and AJAX just to display "0" somewhere on the page. Doing such a mechanism for that many eshops is a huge task and even bigger is to maintain it and test it. Also it doesn't solve B and your EPC will still be crap.

B. Stock availability is not the only problem. If only it was.
- Some times the prices change.
- Other times the photo we have is a green dress and on the product page it says that there are 0 green left and 98 pink versions. Still in stock, but wrong product.
- Other times there are 34 XXXL available and 1 S. Practically out of stock.

For now my biggest problem is Linkshare.

For CJ I use the full link in buy buttons and I extract the product link from the URL in order to use it for all the testing.

[Chuck Hamrick]

I work with an affiliate who runs a team of 12 that scraps sites for product information. Would be happy to introduce you, send me an email.

[JoyUnltd]

Originally Posted by orfeo

For now my biggest problem is Linkshare.

This situation might be related to the scenarios you described. I'm certainly not at the level to do the testing you are, however, recently found a similar situation with a LS merchant and going through this for the past week trying to sort it out.

As sales with this merchant gradually tanked over the past few months, then took a nosedive, I started checking links & found many that went to the product page but were actually noted as being discontinued or out of stock on the page. I sent a bunch of links to the AM who verified that these and similar discontinued items were being pulled from the feed on their end.

She found out the error is occurring on the LS end & told me that LS is attempting to resolve it. Perhaps this might apply to more LS merchant datafeeds?