Underestimation of Architecture Stability in Affiliate Tracking
Along with cost and features, serving Stability tops the list of concerns that Affiliate Networks usually pay attention to while choosing tracking platforms. In this article, we will try to draw some attention to different aspects of this factor to give you a deeper understanding of how back-end processes can dramatically affect your business, no matter you are using 3rd party service or developing your own in-house platform.
Brief Overview of Tracking Process
Once the user sessions occur, the clicks go to the tracking system to be recorded with a click ID, this click with its ID would be passed through to the advertiser as well. When one click converts, a postback would be passed to the tracking platform through the postback URL. In each single case, the conversion bears different actions – install, sign up, registration, or purchase.
What actually do we mean when we talk about Stability?
Let’s dive a little into this issue so you could understand all the sophisticated expressions the developers put together.
For Clients, system stability is reflected in two major constituents:
1. System Uptime –What exactly does it mean? In short- it’s the capability of a system surviving emergencies. Under the system stability, we mean three components: hardware, network and software, where first and second are maintained by server provider and software is managed by platform in-house developers.
Basically, tracking systems base their servers on a public cloud server like AWS (Amazon Web Services), Microsoft Azure, Google Cloud – these big 3 providers are controlling more than 65% of a public cloud of Iaas and Paas market. FuseClick is based on AWS, as the most reliable one (with a broad number of big clients such as NASA, Airbnb, BMW, Docker trust AWS).
Now, you might think: “Everyone assures me, I will get 99,99% reliability of his tracking, what’s the difference between all platforms then?”. Yes, presumably, all big platforms work with one on the same providers, the server architecture impacts a long way and it’s really where the SaaS software distinguish from one another. Here we are coming to the next question about tracking speed.
2. Click Time – or tracking speed, is how much time a user could be directed to the landing page. Click time can variates from 40MS (if traffic comes from the same area as server deployed) to 3-5S (if traffic comes from the far region). In practice, this parameter is closely correlated with click loss rate: the longer the click time is – the higher the click loss a Network will bear.
Click loss is always very sensitive especially for big Networks, so let’s elaborate on what time is spending for and how can platform improve the click time.
Well, basically, click time is a cumulative amount of transmission and processing time, when click got recorded by the server.
Network transmission time is a part of click time, that needs a click to reach the Tracking Node (Server) and come back to the user.
Obviously, as we said before, this parameter depends on the regions of coming traffic and location of server deployed. The more servers are available for a Network, the less time it will take for traffic from any country to reach the nearest node, ergo transmission time can be reduced from 3-5S to 40-50MS, with significant click loss average decline by 15% (even by 30% in some cases).
So what is the problem? The problem is that most of the platforms on the market, their architectures have been developed with only 1 or 2 nodes. This basic decision they made in the beginning of their development stage was based on financial and technical capacities. These initial investments increase with the platform complexity, that has been put as a baseline for further development.
Once platform architecture is done, it’s almost impossible to make any changes in the source code to add more options for tracking infrastructure. Now it demands much more sources and energies for secondary coding and testing, – so it’s easier to start a new platform from scratch then.
Why clients lose? Losses come when Network doesn’t have enough leverage to manage a transmission click time for traffic, coming from far regions. In this case, all traffic will go to a single node, no matter how much time it takes. Potential threats are that click loss rate can be so high and absence of backup tracking server can freeze traffic processing, in case this single node goes down.
In FuseClick we developed the code in a dynamic way that allows us to add as many nodes as we need in few hours. By embedding multi-node tool we can do traffic processing for our clients without any headache.
Processing time – time needed to process and record the click in Tracking Node. The main importance of this is a direct correlation with click loss: the shorter the processing time is – the less the click loss network will bear.
This parameter is fully depended on the algorithm optimization applied by in-house developers. FuseClick has improved this speed from 5MS in 2016 to 1-1.5MS in 2017, that lead to about 3% further click loss decline.
Don’t miss a chance to test it for free!!!
As you are probably aware, currently, many SaaS developers (including tracking platforms) provide a free trial period. Utilize your free trial to test all these things!
Get done with set up in 1-2 weeks and test it fully for the remaining time!
Firstly, it will not cost you anything. Second, if your expectations about these parameters haven’t been missed, you still have a chance not to sign the contract. Check your contract point for things you are really taking care about, don’t hesitate to request a tracking speed statistic, platform server capacities- don’t buy a pig in a poke!
What’s the value for different scale businesses? / Does Stability matters for small and mid-size companies?
Despite the fact that it’s always a big player who usually concern about the server stability the most, small and middle-size business also pays attention to this factor even if it may seem not that much significant point for them. Probably, for last ones it may be more relevant – they are more sensitive to technical failure. Unlike big players who has technical backup servers to support smooth functioning even if provider’s servers fall down, small companies have to rely on their providers fully. In the worst case, they will bear both financial and reputation losses.
