Want To UNDERSTAND Tracking? Read This Now.

If you want to really understand how tracking works, you only need to understand three things.

These three things will let you understand any tracker - not just affiliate trackers like Voluum or Prosper, but also traffic source tracking and affiliate network tracking like Cake. They all work the same way!

This tutorial's pretty long, but stick with it!



I've split it up into three sections, plus a wrap-up at the end. There are handy self-study questions in each part that you can use to make sure you've understood it before you move on to the next.



Thing The First: Database IDs

The first thing you need to understand is how databases work. All tracking runs on its own database, which is basically the same as a massive spreadsheet - columns of data.



Whenever a visitor enters any tracking system, whether that's an affiliate network tracker, an affiliate tracker like Voluum or a traffic source tracker, they are assigned an ID.

Then, the tracker associates every other piece of information it gets about that visitor with that ID. Their mobile carrier, their browser, which lander they visited, whether they converted, whatever.

That way, you can find out all the info you want from a tracker. For example, if you want to know what the conversion rates are for a mobile carrier, the tracker searches for all IDs that used that mobile carrier, then checks how many of those IDs converted and gives you the result.

EXAMPLE: You're running a campaign in the UK and want to know how Orange is doing as a carrier. The tracker looks for "Orange" in its "carrier" column, makes a note of the IDs of every visitor who has "Orange" listed, then goes through all those IDs and counts how many are listed as having converted, and how many aren't.

These IDs are called different things on different systems.

Nearly all affiliate trackers call them either a "subid" or a "clickid".

Traffic sources call them different things. Decisive, for example, calls this id a "bidhash". If you want to send information back to a traffic source - for conversion tracking - you need to know what the name they use is.

So if you see something called a subid, a clickid, or something similar, you know that's an identifier for an individual visitor in a database.

But why do you often see those with square brackets, curly brackets, or hashes around them? Well, we'll get onto that in Thing The Third - but first, you need to know about how information is passed from web server to web server.

Question to ask yourself to make sure you've understood this bit:
When you search for the CTR of visitors who landed on a specific lander, what does the tracker do to find that?



Thing The First-And-A-Half

One other thing you need to know about databases is that they have fixed "slots" for data associated with each ID, with standard names.

Many trackers let you store any information you like in those slots, but the slots are always named the same. For example, Prosper has slots called c1, c2, c3 and so on. You can store anything you like in them, from Decisive bidhashes to POF age and state info - but the slots are always called c1, c2, c3 and so on.

Likewise, Cake gives you slots called s1, s2, s3, s4, and s5. We usually use s2 to store the subid we send from our trackers to Cake, but that's just convention - you can store anything you like in there, from names of your favourite animals to guesses on the winner of the next Superbowl. Storing subids is probably more useful, though!


Thing The Second: GET requests

If you controlled both the traffic source and the offer for your campaigns, tracking would be waaaay simpler, and you'd barely have to do any setup at all.

That's because the same system would track all your visitors from arrival to exit. The thing that makes affiliate tracking tricky is that there are at least three systems involved: traffic source, affiliate tracker, and affiliate network.

Those systems have to pass information back and forward.

  • The affiliate network needs to pass information to your tracker when a visitors you've tracked converts on an offer.
  • To do that, your tracker needs to have passed the affiliate network the ID you were using to track them.
  • The traffic source has to pass information to your tracker about which ad the visitor who just arrived at your tracker clicked on.
  • Sometimes your tracker has to pass information back to the traffic source about which visitors converted, using the traffic source's ID rather than the tracker's.


And so on.

Fortunately, there's a standard way of passing information back and forward on the Web - it's called a GET request. That sounds pretty intimidating, but the fact is you see hundreds of them every day.

A GET request is added onto the end of any Web URL in order to pass information to the web server running that URL. You can spot the GET request pretty easily - it's everything after a "?" symbol anywhere in a Web address.

So, for example, here's a URL with a GET request: https://www.google.co.uk/search?q=boobs

In this case, the GET request is q=boobs. So what does that mean?

GET requests are comprised of two parts: the name of the information being sent, and the value of that information. In this case, the name of the information is "q", standing for "query", and the value is "boobs", because that's what I searched for on Google.



If you want to send multiple pieces of data, you send them after the question mark separated by an "&". So, for example, here's what a standard affiliate URL from Cake breaks down into:



You use these GET requests all the time in your affiliate campaigns.

If you're creating Facebook ads, for example, you just send the name of each ad in your GET requests to your tracker.

On Prosper, for example, you'll send "t202kw" as the name, and the name of your ad as the value, giving you t202kw=Ad_1_Blonde or similar at the end of your URL.

And most of the time a tracking system will generate some information automatically.

For example, Prosper auto-generates a campaign ID so that it knows which campaign to send visitors to. That's the "t202id" variable in your Prosper URLs.

Question to ask yourself to make sure you've understood this bit:
If your Facebook URL ends in ?banner=banner1&subid=123 , what does that mean?

Thing The Third: Tokens

Often, you won't want to send the same information every time in your GET requests - you'll want it to vary by visitor.

For example, if you're running an offer on the Cake platform, you'll want to send the subid of EACH visitor to Cake, rather than sending the same subid every time.

That's simple enough to do - you just need a piece of text that the tracker software will recognise and replace with the right information at the right time.

These are called "tokens", and they vary across all trackers, but mostly look roughly the same.

They start and end with at least one symbol rather than a letter. That can either be { and }, [ and ], # and # or something else. Sometimes tracking systems even use multiple symbols, like CPVLab's {! and !}, or Decisive's {{{ and }}}.



See https://stmforum.com/forum/showthread...-Source-TOKENS for a very complete list of tokens!

Why do they do this? To signal to the tracker software that this is a piece of text that needs to be replaced, rather than just copied.

If the tracker just replaced anything called "subid", for example, it might replace the name in some GET variables - so &subid=subid would end up being &123=123 rather than what you intended, &subid=123.

So, wherever we want the tracker to put appropriate info for each visitor, rather than just the same info for everyone, we use a "token". You can usually find a list of tokens on the documentation pages of the relevant affiliate network, tracker or traffic source - if all else fails, check IAmAttila's massive list of them here or Google!

Question to ask yourself to make sure you've understood this bit:
If you know that you need to pass a subid to your affiliate network with the name "s2", and your tracker uses the subid token {clickid}, what should you put at the end of your offer URL?


Putting It All Together

You might feel like you're still confused!

But actually, you've now got all the pieces of how tracking works.

Let's look at running a campaign with Prosper, Facebook, and a Cake-based network.

Offer URL

You start out with the Cake URL, which will look like this: http://ubernetwork.com/?a=1431&c=2114&s1= .

Given everything we know above, we know that this is passing information TO "ubernetwork.com" - your affiliate network. It's passing the information "a" (affiliate ID) and "c" (campaign ID) to the network.

We know that we need to send the ID that our tracker uses too, so that Cake can pass that back when people convert, and we know that the best GET variable name to use is "s2", so we change "s1" to "s2".

Now we enter this into the "offer URL" section in Prosper. We put it in Prosper because Prosper is the part of the tracking system that needs to send information and visitors TO the affiliate network.

We know that Prosper uses the token "[[subid]]" to represent the place where we want it to put the visitor subid. So we add that as the value in the name-value pair s2=[[subid]].

When someone actually arrives at our offer, Prosper will replace "[[subid]]" with their subid - so if their ID on our tracking system is 42, they'll end up going to http://ubernetwork.com/?a=1431&c=2114&s2=42

URL for Traffic Source

Next, we tell our traffic source where to send people when they click on our ads. Prosper creates this URL for us, but it will end up coming out like this: http://www.blahblah.com/LP/Anast/CLa...d=6976&t202kw=.

We know that Prosper takes the "keyword", which is usually used as the banner name, as "t202kw". So we add our banner name for each banner to the end of that address, completing the name-value pair t202kw=Ad1 for Ad1, for example.

That means that when people click on Ad 1 in Facebook, our tracker gets sent the information that they clicked on that ad. We don't use a token here because Facebook doesn't have any tokens!

So, we've got two of the information connections between tracking systems sorted out, as we discussed in Thing The Second. We have a URL in Facebook which passes information to our tracker, and we have a URL in our tracker which passes information to our affiliate network.

URL for Postback

All we need to do now is to let our affiliate network know where to send information when someone converts, because we want to send that information back from Cake to our tracker.

We do that using a Postback URL, which we get from Prosper, and input into Cake in the "Postback URL" section of the offer.

Cake has a handy list of tokens at the side, and we know that we want to pass "s2" back to our tracker as "subid" (because we sent "subid" to cake earlier, which stored it as "s2").

So we take the Postback URL that Prosper provides, which looks like
Code:
http://track.tracker.info/tracking202/static/gpb.php?amount=&subid=
, and we complete it by adding #s2# at the end (because this URL will be processed by Cake), completing the name-value pair subid=#s2#. That means that whenever someone converts, Cake will now send the subid of the person who converted back to Prosper, which can then make a note of that in its database, so we can then search to find out what landing pages are working best, which carriers are working best, etc.

And that's how tracking works!

I figure you may well still have a question or two - if so, ask 'em in the comments!

Komentar

Postingan populer dari blog ini

The Perfect Mobile Banner Ad. (Template inside)

Need A Geolocation Script To Redirect Users Based On Their Location? Here's One!

How To Boost Your ROI With Photoshop