In an earlier entry we focused on the definition of what we call “attribution” (that great unknown concept) in this article we want to go deeper into the use we make of this concept in Kimia.
Let us briefly recall that attribution is successfully matching all traffic characteristics, sources, sub-sources, internal attributes… of our publishers to the final statistics, that includes the revenue generated.
In Kimia, what we understand as “characteristics of origin” which we attribute to the traffic, could be divided into three:
- Fixed Features:
- Publisher: who sends us traffic.
- Channel: the website, application, or sub-source of fixed traffic defined by the publisher.
- Ad-code / Zone-id: this corresponds to an integration where the publisher sends us the traffic (a banner of a certain size, a redirect, a pop up, …)
- Variable characteristics: In the URL that we give the publisher to publish our integrations we allow values to be sent to four parameters that will also be stored in the attribution. Usually these parameters are used by our publishers for information regarding the position of the creative, the page in which it appears, or other data related to the sub-sources that you want to assign. These parameters are: af, mk, ext1 and ext2. Later we will talk about them when we go over statistics.
- Click identifier: Some of our publishers have their own attribution system, so they prefer to send us a click identifier that we will send them back by postback, in case a payment event is confirmed. Through this click identifier, the publisher will be able to make its own attribution.
Both fixed and variable characteristics will be stored in our statistics, so that our publishers can make queries, filters and groupings for those characteristics and can see the attribution assigned to them in terms of prints, clicks, conversions and generated revenue. In addition, traffic characteristics, such as the country of origin, operator and device, will also be associated.
In general, our publishers choose our dashboard to check their statistics. With these stats, publishers are provided with all characteristics of the attributions, except for the variable characteristics ext1 and ext2.
Kimia’s platform dashboard
Some of our publishers prefer to request the stats in CSV file format, to integrate later in their own statistics. These reports also contain all attribution features, however, they do not contain data for the af and mk variables.
The click loss between the affiliate link and the offers final landing is totally out of a publisher’s control and not only that, you can’t really measure it, as the clicks you see on your affiliate network stats are the ones that reached the affiliate link, not the clicks that reached the offer landing page. The affiliate network has no way to count those (the offers landing is not theirs, so they don’t have any kind of stats for the clicks that reached it, unless the offer/partner reports them back, which is unusual).
You need to separate the click loss between the different jumps. As mentioned before, you can measure and try to control some, but others are totally out of your control and you won’t even be able to measure them.
1) The first and biggest jump is the publisher/ad network to your landing (comparing clicks counted by the publisher and the ones received in your landing). Here it will depend a lot on the way you get the traffic:
– Redirects: second on click loss % after pops. Similar to pops, two reasons for click loss: hosting and page loading speed, and then the bounce rate because the landing angle/design. The main difference with pops is that the original publishers page is no longer available, so they might stay in your landing if it’s attractive.
– Banner clicks: First click loss again happens on hosting, but here the bounce rate should be way lower if your landing is really related to the banner. They already know more or less what expect from the first click, so they have more patience waiting for your landing to fully load.
2) Once the surfer has decided to stay in your landing, of course the bounce rate and the ctr of your link to the offer are the key here, it’s all about angle/design.
3) Once the user has clicked on your button to the offer, then it’s all about the affiliate networks hosting and tracking software speed.
4) After the surfer has been redirected to the offer (if the offer is rebrokered through other networks, then step 3 applies again.) Then its all about the offer landing loading speed, which is something totally out of your control and out of the affiliate networks control, even in terms of trying to measure it, as I wrote in the intro.
All these jumps apply both to mobile and desktop, but the issue on mobile is greater for one reason. Some devices/OS, specially old ones, have issues handling several redirects and tend to hung up after a few. So when you receive the click from the ad network, it might have been through several redirects before, either at the publishers site itself or if the click is being brokered through several ad network, that is pretty common in redirects/pops, not as much for banner clicks. Therefore the click you received might have gone through 5 or 6 redirects before reaching you, and it might have frozen up before loading your landing.