From a sample of four customers we attempted to find any correlations in ctc time across the different verticals of each client. I thought maybe this would give us some insight into user behavior, and it turns out I was right.
The results below display the percentage of users that convert at different times depending on the types of offers (campaigns).
Conversions Sample Size of 528,788
Software Tools Conversions Sample Size of 7,061
Gaming Conversions Sample Size of 772,677
Entertainment Conversions Sample Size of 370,620
You will notice that 3 out of the four programs we pulled data from have an overwhelming majority of users converting in less than 30 minutes. Why do you think that is? And if that is the case, why do all of them have a 30 day session lifespan for affiliates? What is the probability that conversions occurring outside of 30 minutes are fraud?
Think about it. If 99.7% of your visitors converted in the first 30 minutes (as shown in the Entertainment data above) and an affiliate sends you 100 conversions that happened hours after the click, something is not right (at least in a direct response model). When the vast majority of visitors have a very exact behavior, why is it that the visitors coming from a certain affiliate have such a drastically different behavior?
For example, visitors that come back to an offer and convert days after their click are almost always influenced by some additional advertisement. Perhaps another affiliate got to them and sent them back with a coupon code. How likely is it that the user suddenly remembered to return to the advertisers offer and signup, days after they clicked the ad, without being prompted by any other marketing materials.
Again, what is the point of having a 30 day cookie? Let’s say the companies we collected data from set their max conversion time to 1 day. If a user actually does come back to the offer several days later, it is still likely that the affiliate will still receive credit. The visitor can always pass back through the affiliate’s ad, but if they read a new article that changed their mind about the offer, should the original affiliate get all the credit. And what is the likelihood in that scenario that the affiliate is doing some sort of black hat cookie stuffing anyway?
Now, I’m not saying that every advertiser and network should go out and set their cookie session time to expire after thirty minutes, but this data certainly compels me to consider traffic that falls outside of the norm as suspect.
So why exactly do so many users (on direct response campaigns) convert in that 30 minute slot? It just so happens that Google also picked 30 minutes as the session cut off time in Google analytics. So then why is it that you see so many offers and campaigns with cookie lifespans of 30 days or more? Is it really necessary to give credit to the affiliate when all that time has passed. And if you do, what are the odds that user will just google your name, click your paid search listing and cost you an extra $1.25. Perhaps Google figures that it is more likely for a visitor to be impacted by some other factor if they’re coming back later, or maybe they just want credit for a PPC conversion.
It is time to start making some intelligent decisions about your conversion data. The first step is figuring out how quickly the majority of your users convert. Once you have this very powerful data, you can start to make quick judgments about the traffic that falls outside of that norm. Perhaps you’re seeing the opposite. Maybe it takes the user several days to convert. If so, why do you think that is? Are they converting because of some internal follow up email, and if that’s the case, do you think the affiliate deserves the sale? Also, take a look at the traffic you already know is fraud and consider the conversion times. Did the users convert in less than 10 seconds even though it takes half a minute just to physically fill out the form and click continue? This might sounds pretty basic, but it really is a piece of data that just about everyone ignores. I would love to see some comments below on your findings.
As we continue to develop new tracking technologies, we find it incredibly valuable to study patterns in data across our wide range of customers. Feel free to give us suggestions on the types of data you would like to see. Email email@example.com, or reach out to your account manager. I can already say that you can expect a lot more where this came from :)