[Haiko de Poel, Jr.]

Stephen,

Since the merchant sets the cookie and usually needs to track the performance of their site, cart and other good stuff like returners and clicking, most use some (or many) form(s) of analytics like Omniture or the like. How does linkshare ensure that these analytic programs do not destroy, or otherwise corrupt the merchant's cookie so that all the sales that should be credited in the LSTrans is correctly uploaded?

Also, I understand that linkshare doesn't use it's cookies to track sales, but it does use merchant's cookies ... so wouldn't "We don't use cookies" be a "misnomer"?

[Haiko de Poel, Jr.]

Wanted to add

If I go to Overture (not affiliate cookied), shop, get to the cart and then proceed to look for coupons in google (an affiliate link) and click on it to open in a new window and then click on view cart in that new window, my cart contents get trashed so I am assuming that the 2O7 cookie which is an integral part and querier of the OSTK cookie gets corrupted with the linksynergy link, therefore, I am postulating that the data that it sends back to LS in the LSTrans should also get corrupted.

[smesser]

Jeff,

This thread was actually started because of a discussion in the CJ forum that called into question the CJ cookie tracking accuracy. In that thread Todd, in response to a question with regards to our tracking without cookies called that approach LinkShare propaganda. To his credit he has since retracted that statement as he was misinformed about the way we do things and thus I do not believe it was malicious. That lead the posters to ask for a clarification between the two methods as most affiliates did not either believe that we did it differently or did not understand how what we did was different and its implications. The thread has never called into question our accuracy as far as I can tell.

Now to your questions, to be clear, our installation guidelines say that a merchant has to store our tracking information into a database at entry (what we call a landing page) that is then associated with a shopping cart or session ID (depending on how the merchants ecommerce system works) and a merchants cookie (for return days). That way at the end of a sale we can associate the user’s session with a LinkShare ID (via a database match) and pass it along (via a secure channel) without having to go to a cookie. They will of course still set a cookie as I mentioned above (and are required to do so), one as a redundancy method in case of database failure but secondly we require them to do so if they want to offer return days. That cookie will also be a merchant cookie and not a network cookie because of the risks we have already gone through in detail. So based on our guidelines, the cookie is not supposed to be used during initial tracking other then for redundant purposes. So to answer your question, you are not correct in understanding how our system works.

That being said, since there are so many different shopping carts, we sometimes have had to make an exception to this rule, but in each case the issue is reviewed by a technical group to make sure that the approved method has every possible safe guard against loss as possible. In the end, this is the exception and not the rule and sometimes it is a necessity.

So once again, to answer Jeff’s question, the cookie should not be used for the initial tracking activity and only as either a redundancy method or for return days. Jeff, I hope that clears up any misunderstanding.

Best regards,
Stephen

[smesser]

Haiko,

I think my post to Jeff should clarify some of your issues. With regards to analytics, we do work with all the providers and merchant partners to make sure that the analytics engines do not cause any problems.

Best regards,
Stephen

[Jeff_Converseon]

Steve,

Thanks for clearing up how the thread started, my bad.
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>...our installation guidelines say that a merchant has to store our tracking information into a database at entry (what we call a landing page) that is then associated with a shopping cart or session ID... <HR></BLOCKQUOTE>
I did understand that this is required for a merchant to track the LinkShare sale. My point is that this can usually only be accomplished by writing a session cookie. Unless the merchant is passing the session through the URL on every click (like Amazon), the only way to associate a database entry created at click with a later shopping cart action is to use a session cookie.

In other words, cookies are not just used "for return days," they are an absolute requirement for the majority of your merchants to track the intial click to purchase.

I would say that my initial understanding of your tracking is correct. Though a LinkShare cookie is not used to track a sale, a merchant session cookie is almost always necessary to do so.

[smesser]

Jeff,

As I mentioned there are exceptions, but I would disagree with your statement. Most of the merchants that we work with use database driven sites and thus either pass hidden variables in the page, the URL or keep track via the database. Thus a cookie can be used but is not typically the way database driven sites manage the user session.

Best regards,
Stephen

[ToddCrawford]

First, I still feel this whole cookie issue is LS propoganda. I admitted I was wrong in thinking that LS used its cookies to track return days. I haven't retracted anything - I just admited I was wrong.

Advertisers using CJ have more than one option for tracking transactions. They can rely on our cookies and tracking pixel, they can set their own cookies and batch data or they can do it the way LS does it and capture values during a session and track transactions without cookies.

Most of our advertisers have internal tracking as well as our tracking integrated. In the event that a sale is not tracked (for whatever reason), they can still recover the data.

I think there is an easy way to test the accuracy of LS vs CJ tracking. Let's select a current LS client that has a reliable history of tracking transactions and launch them on CJ (free of charge) for 90 days in hidden mode (not visible to the entire network). We can ask for volunteers from current affiliates to switch out their links for 30 - 60 days. At the end of the test we can see if there are any differences between what the affiliates earned via LS and what they earned during the test period via CJ.

[MichaelColey]

Stephen,

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by smesser:
Most of the merchants that we work with use database driven sites and thus either pass hidden variables in the page, the URL or keep track via the database. Thus a cookie can be used but is not typically the way database driven sites manage the user session. <HR></BLOCKQUOTE>Todd's (admittedly unscientific) test of Linkshare merchant carts indicated that over 70% of the merchants appear to require cookies for their shopping cart to work. That directly contradicts your statement that most pass the session information in hidden variables or the URL. Just because they use a database doesn't mean that don't use cookies for session tracking.

[ToddCrawford]

Jeff,

The value is stored in the HTTP header and persists as long as the browser stays on the same site. Cookies are not needed to track and capture this information. If a sale occurs, all the data gets written to a file on the merchant's server and then is batch over to the network.

Jeff, I would remove your email address from your sig unless you are looking for more spam.

[MichaelColey]

I would like to make one major clarification to this whole discussion.

This all boils down to two different technologies for affiliate tracking:

1) Network Cookies / Image Tag. This is the technology used by most CJ merchants, some (most?) BeFree merchants, and many/most merchants at other networks. The upside is that it is VERY simple to implement (and hard to accidentally mess up). The downside is that cookies (and/or the image tag) can fail some percent of the time.

2) Merchant Tracking with Sessions and/or Merchant Cookies and Batch Reporting. This is the technology used by all Linkshare merchants, some BeFree merchants, some CJ merchants, some merchants at other networks, and most independent merchants. The upside is that it has the potential to track more sales. The downside is that it's far more complicated (and thus more prone to failure at many different points, and that it still relies on cookies most of the time. It is especially vulnerable with custom links, as EVERY potential landing page on the site must properly log affiliate variables and initiate sessions.

My contention remains: When you weigh all the factors, NEITHER technology has a significant advantage over the other. They both have problems. In the overwhelming majority of cases, EITHER will track fine. In theory, the second technology (the one Stephen is defending) should be more accurate. In practice, I've seen more problems with the second technology, although it hasn't been significantly worse than the first.

Also, I want to point out that the ONLY person claiming that either technology has a signficant advantage is Stephen.

[Jeff_Converseon]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Most of the merchants that we work with use database driven sites and thus either pass hidden variables in the page, the URL or keep track via the database. Thus a cookie can be used but is not typically the way database driven sites manage the user session. <HR></BLOCKQUOTE>
Thanks for the clarification.

Just to be clear, however, just because a merchant uses a database driven site doesn't mean they don't use session cookies. A database entry and a browser session must be linked either through the session cookie or the URL.

If you are going to continue to claim that "LinkShare doesn't use cookies to track sales," then you'd need to verify that all of your merchants pass the session in a URL. Otherwise, it is misleading.

It would be more correct to say "LinkShare doesn't use cookies to track sales when the first click leads to the sale AND when the merchant passes the session through the URL. Otherwise cookies are necessary."

Normally I would say this is splitting hairs, but because you are so quick to point out the "flaws" in cookies (thereby bashing competitors in the process), and insist that LinkShare tracking is superior because of those flaws, I think it is important for you to be very clear about this issue.

You did indicate that it is an exception to the rule, perhaps you can be more accurate about how much of an exception? Todd's testing indicated that it was actually the norm -- 70% used cookies to track the session. If that number is incorrect, what is the correct number?

[Jeff_Converseon]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>The value is stored in the HTTP header and persists as long as the browser stays on the same site. Cookies are not needed to track and capture this information. If a sale occurs, all the data gets written to a file on the merchant's server and then is batch over to the network. <HR></BLOCKQUOTE>
Again, if the session is not kept in the URL string, then there is a cookie being used -- a session cookie that expires if the user leaves the site, as you said.

There are only a couple ways to track any visitor as they go from one page to another -- either a session cookie or a session variable in the URL. Otherwise all you know about that person is their IP address and other information that is not unique to that user.

[abestwebas]

He's not the ONLY one. I've said before during one of my many times of CJ bashing that I've had consistantly better results with LS and still think so.

[MichaelColey]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by web_novice:
He's not the ONLY one. I've said before during one of my many times of CJ bashing that I've had consistantly better results with LS and still think so. <HR></BLOCKQUOTE>But are you saying that the difference is due to tracking?

[abestwebas]

QUOTE]But are you saying that the difference is due to tracking?[/QUOTE]

I'm sure there are multiple reasons. Perhaps tracking is one of them.

[ToddCrawford]

Jeff,

A simple JAVA script within the page can store tracking values durring an initial visit without holding the values in the URL or in a session cookie.

We're splitting hairs here but I wanted to point out that it is possible to track an initial visit without using cookies or passing the values in the URL.

I believe this is how LS claims to track session based sales. Although Jeff you are probably right that some (most?) actually use a simple session cookie to track sales via LS.

[Kellie aka Ms. B]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Advertisers using CJ have more than one option for tracking transactions. They can rely on our cookies and tracking pixel, they can set their own cookies and batch data or they can do it the way LS does it and capture values during a session and track transactions without cookies. <HR></BLOCKQUOTE>

Todd is there any specific criteria set by CJ which determines which method an Advertiser uses or is left up to the Advertiser to decide?

[ToddCrawford]

We work with each advertiser to determine which tracking method will work best for their business model. Most opt for the pixel integration because it requires less technical resources to implement and gets their program up and running faster.

[Jeff_Converseon]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>A simple JAVA script within the page can store tracking values durring an initial visit without holding the values in the URL or in a session cookie.
<HR></BLOCKQUOTE>
I still believe I am correct that you would need to pass those values from page to page with a cookie or through the URL -- a JavaScript can't keep a session from page to page on it's own. However, you are correct in saying this is really irrelevant as the topic here is what portion of LinkShare merchants use session cookies to manage cart sessions.

[Affiliate Ian]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Jeff_Kowabunga:
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>A simple JAVA script within the page can store tracking values durring an initial visit without holding the values in the URL or in a session cookie.
<HR></BLOCKQUOTE>
I still believe I am correct that you would need to pass those values from page to page with a cookie or through the URL -- a JavaScript can't keep a session from page to page on it's own. <HR></BLOCKQUOTE>
I on the other hand likes to split hairs... just kidding. Just trying to understand all tracking issues to better our own system.

I would think javascript tracks a user by pushing a value to a session cookie, cookie or URL. By itself, I am unaware that javascript can store a tracking value. Maybe I need to learn more? What am I missing here?

[smesser]

Todd,

Nice one, you should be in sales. I have an idea, why don’t we take your top customers and sign them up for LinkShare ;-). Anyway joking aside I do think that accuracy is a big issue for our industry and our hard working affiliates. We could have easily gone down the lower cost (cookie) route years ago and saved a ton of money. Because of whom our customer base was (fortune 500 brands) and the reputation that we wanted we did not want to take on the risk that cookies presented. We also felt that this known liability was too big to take on.

Cookie tracking would save us a huge amount of money. I do not think this is splitting hairs by any means as you will see below. Our method, while more expensive however is more accurate and thus allows us to avoid some of the discussion threads on your boards such as:

Take a look at these tracking failure examples (just since August) in the CJ forum from Abestweb. This is just a taste of the extent of problems that we are talking about:

CJ and Norton blocking over a year now..... (http://abw.infopop.cc/eve/ubb.x?a=tpc&s=548608979&f=106607689&m=806105195)
CJ sales WAY down (http://abw.infopop.cc/eve/ubb.x?a=tpc&s=548608979&f=106607689&m=380109676)

Out of 63 Unique Clicks CJ Has Tracked 3! (http://abw.infopop.cc/eve/ubb.x?a=tpc&s=548608979&f=106607689&m=340100786)
*.befree.com and *.bfast.com added to restricted sites list : (http://abw.infopop.cc/eve/ubb.x?a=tpc&s=548608979&f=106607689&m=852102296)
CJ Flunks Test. (http://abw.infopop.cc/eve/ubb.x?a=tpc&s=548608979&f=106607689&m=251108166)

qksrv.net/Cj affiliate links never work on my computer (http://abw.infopop.cc/eve/ubb.x?a=tpc&s=548608979&f=106607689&m=985103156)
Todd, qksrv.net cookies = pests (http://abw.infopop.cc/eve/ubb.x?a=tpc&s=548608979&f=106607689&m=123103376)

All of which are complaining about missed sales, or cookies issues. In addition, as I mentioned earlier it is clear that with adware removal programs, Norton, and other issues around cookies we discussed, the accuracy problem will only continue to get worse. I have many more potential problems around cookies, but I think I have made my point (hopefully)

With regards to how many people are shutting off cookies (the first question in the logic tree that I posed): Rather then take Todd. Jeff’s or my words on what we see, read what the Pew internet study found.

“Pew found that 84 percent of Internet users in the United States are concerned about businesses and strangers getting their personal data online, but 56 percent did not know about cookies.”

”More notably, 10 percent said they took steps to block cookies from their PCs, Pew found.”

Ok that is one neutral (very reputable) party and not someone who is protecting their business.

Now how about how many people have cookies from third party advertisers (do not read third party cookies) shut off or destroyed.

Spyware/Adware removal apps (such as Lavasoft, Spybot, etc.) comes with cookie blockers/sweepers that as we all know disable advertiser cookies (read CJ and other cookie based trackers) cookies as a default setting. (Lavasoft alone claims to have distributed its Ad-Aware app to over 80 million home and corporate users.) That is a pretty large base of users that are not being tracked everytime the cookies are swept.

And finally the third failure, well you would know what that is so I will not touch on that.

All of these failures multiply on each other to create an exponential effect (Each feeds on the other causing the number of lost sales to keep growing as you go through this process). In other words, each failure adds on top of the prior one to create a lager base of problems, taking away from your commissions.

That is why I keep saying that its time to sunset cookie tracking. Advertiser based cookies as a tracking device should not be used if you want accuracy. I know you need to protect your business, but the flaws that we have already uncovered should not be ignored because it will only cause the affiliates to start to question the entire space.

I am not here to say that we are perfect, far from it. But when we find out an issue, we don't go hiding our head in the sand about it. When we find out about an issue, we take a proactive stance to make changes to our methods to minimize the exposure for both our merchant and affiliate clients. For example, we use cookies in our log-on interface where affiliate and merchant can access their accounts. Non-mission-critical from a tracking perspective, but an annoyance nonetheless. With all of the cookie blocking going on out there, the inability to log on is one of the fastest growing areas we get questions and complaints. So, we are redesigning our own site to eliminate cookies because of this issue.

As upset as I am to hear that you are still calling what I am saying "propaganda," all it proves to me is that you don't want to address the three major tracking failures that you yourself agree exists. Ultimately, it's up to merchants and affiliates to choose which systems they want to rely on, but when there are more accurate alternatives out there to choose from, affiliates should be wondering whether, by working with a merchant that has chosen to use a less expensive cookie-based tracking system, you are leaving money on the table.

Todd, we can go into the scale of these issues further if you really like, but I am not sure that it is worth spending more time on this issue if all you will do is refuse to acknowledge this reality, poo poo the risks by throwing out unsubstantiated numbers as diversion, and bury your head in the sand. On my side, I think its time for me to get back to work now.

Best regards,
Stephen

[bcwaller]

This has been very entertaining and educational. I just have to pop in with a comment.

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>With regards to how many people are shutting off cookies (the first question in the logic tree that I posed): Rather then take Todd. Jeff’s or my words on what we see, read what the Pew internet study found.

“Pew found that 84 percent of Internet users in the United States are concerned about businesses and strangers getting their personal data online, but 56 percent did not know about cookies.”

”More notably, 10 percent said they took steps to block cookies from their PCs, Pew found.”

Ok that is one neutral (very reputable) party and not someone who is protecting their business. <HR></BLOCKQUOTE>

But this does not answer the question of how many people have cookies turned off. This only says how many people in a self selected survey (they chose to answer) "took steps". Was this step installing IE6? Altering the security settings in IE? Installing the Google toolbar?

The only data that would be impartial would be from a third party tracking company such as WebSideStory or WebTrends that has lots and lots of data that they could show over a wide variety of sites to come up with these kinds of stats.

I know that we have not done any extensive checking, but our estimations (yes, not absolute numbers) are under 10%, but how far under we don't really know. BTW, we've been doing URL encoded tracking since the beginning of time (OK 1996), as well as cookie based session tracking.

[Jeff_Converseon]

Brad is correct, the Pew numbers are a poll, not a measure. Here is measure:
[QUOTE]Web site audience analysis service Web Side Story found in a review of more than 1 billion page views that cookies were disabled just .68 percent of the time.[QUOTE]
And again, all of this debate over CJ vs. LS tracking is moot if LS relies on the majority of its merchants to use session cookies for tracking.

[smesser]

Jeff,

With all due respect, you keep trying to point to the grand canyon (cookie tracking)and say its no different then a pot hole (LinkShare tracking).

I am not sure what more I can point to. There is an old saying "No one is as blind as he who refuses to see"

Steve

[Kellie aka Ms. B]

Whew, I'm getting dizzy from all the spin in this thread. http://abw.infopop.cc/infopop/emoticons/icon_biggrin.gif

There is quite a bit of good info here though for both affs and merchants who are willing to pick out the important points. A couple that are probably deserving of more attention than they received also. http://66.132.187.8/smilies/browsmiley.gif