Who overwrites who in the affiliate channel? This question has been at the heart of some hotly-debated issues in affiliate marketing over the last few years, but little research has been done to answer the question directly.
When we scrutinise the value of incentivised traffic, question whether or not voucher code sites are ‘stealing’ sales that would otherwise have been credited to content sites, or wonder how different models of multi-attribution might work, the ability to inform these debates with insights in to what extent cookie writing is occurring, and amongst which affiliates, is surely vital.
We recently conducted new research on this question and the results were presented at the A4U Expo in London earlier this month, looking at ‘daisy chain’ data and revealing the path to purchase a customer takes from click to sale through the affiliate channel.
We looked at five brand advertisers in different sectors to see how many affiliates are typically involved in a sale, and the extent to which their cookies overwrite each other, and focused on the impact of promotions these advertisers ran to get a sense of how the picture changed during the promotion compared to before it.
There were three headline findings that we highlighted, and which we found to be the case for each of these five advertisers.
The vast majority of transactions have only one referrer. In most cases there is simply no other affiliate’s cookie being overwritten. Secondly, almost all transactions take place within 24 hours of the click, and of those the vast majority take place within the hour.
Thirdly, where an affiliate does overwrite another affiliate’s cookie to claim the sale, they are more likely to overwrite affiliates that use the same primary method of promotion. Voucher code sites were more likely to overwrite other voucher code sites’ cookies than they were to overwrite true content sites’ cookies, for example.
What was perhaps most surprising was that these conclusions were applicable despite the different sectors that our advertisers operated in.
Our sample comprised retailers from the fashion, mobile, and health & beauty sectors, and two from the electronics sector. Each advertiser was a household name, had a high street presence and ran large affiliate programmes exclusively with Affiliate Window, giving us a large enough dataset to justify our conclusions.
Our three main findings also held even where average order values were high: even in the cases of our electronics retailers, where average order values of £200-£300 would suggest more ‘considered’ purchases, the vast majority of transactions were still made within an hour of the click and involved predominantly only a single referring affiliate.
Furthermore, we found that these conclusions held true both for the period before the promotion as well as during it, and on both the programme as a whole and for just those transactions referred by the affiliate running the promotion. In other words, these three conclusions seemed to be relatively unaffected by the various exclusive offers that were running at the time: we can say they represent the norm for these advertisers.
One of the most intriguing findings was evidence to suggest that where an exclusive code had been offered customers spent more anyway. In two cases where the advertiser offered an affiliate an exclusive voucher code the average order value from those advertisers actually rose during the time of the promotion compared to before it.
What then can we extrapolate about the performance of the affiliate channel as a whole on the basis of this research?
The statistics on the high number of single referrer sales and fast transaction times demonstrate the success of affiliates in engaging and converting their audiences on behalf of the advertisers they partner with.
It paints a picture of affiliates winning the click-through and converting that user into a customer speedily and as a result of their own efforts. Although it could be argued that another affiliate, or another channel, would have made the sale sooner or later anyway, it is equally true that in that time the customer might have bought from a competitor.
This research shows affiliate activity in a positive light. Affiliates are revealed as click winners, not ‘click stealers’. By and large, advertisers do not need to worry that their affiliates are cannibalising each other’s traffic, and there is no evidence to demonstrate that incentive sites are overwriting the cookies of true content affiliates in any significant numbers, even if this assumption has been adopted by many in the industry.
There are also implications for the ongoing debate on multi-attribution. If the vast majority of transactions from the affiliate channel are sole referrer, surely the debate about how to apportion commissions across a number of affiliates involved in the path to sale is foreclosed? Multi-attribution models are premised on a longer daisy chain than exists in the vast majority of cases.
Ultimately, this research only looked at transactions coming through the affiliate channel. Further, similar research within the channel would be welcome, but outside it advertisers might wish to compare these findings to their own analysis of other channels and use affiliates as a benchmark to measure the effectiveness of their online marketing efforts as a whole.