Up to 25 Percent of Affiliate Transactions Involve Multiple Publishers

Last week an interesting Tweet came from Commission Junction University:

Less than 1% of affiliate transactions had another affiliate in the click path across @cjnetwork #cju2013 #attribution

— Trisha Lyn Fawver (@TrishaLyn) September 11, 2013

It stirred some interesting reactions (and discussions). The industry’s veteran, co-founder of Impact Radius, Todd Crawford, for example, replied:

@TrishaLyn was that a typo? I think it should be less than 25% not 1% of affiliate transactions had another affiliate in the click path

— todd crawford (@toddcrawford) September 11, 2013

I, on the other hand, couldn’t help but recall this Tweet (and data from Rebecca Madigan, executive director at the Performance Marketing Association) from Affiliate Management Days SF 2012:

10.4 average sources involved in any purchase decision according to Rebecca Madigan from the PMA. #AMDays >> attribution discussion anyone?

— Chris Bishop (@cpbishop) March 9, 2012

Later on Trisha clarified that the 1% example was a “1 merchant example.” But, regardless of this, in attempt to gather more empirical data on the subject (or how many affiliates, generally, touch a customer on his/her path to conversion), I’ve reached out to all major platforms and networks that work with affiliates. Those that had something to share emailed back, and below you may find what they have shared with me.

Brian Littleton of ShareASale commented:

I’m would not be terribly surprised if there were an individual Advertiser that was seeing a low percentage, it would all depend upon the makeup of their program.

In general though, 1% is likely a rare exception as opposed to the rule of thumb.   Most programs that have a full makeup of content, coupon, loyalty, price, etc… are going to see numbers much higher than that.

The area that sees the most action, of course, is the “Coupon Affiliate” who commonly overwrite, and are overwritten as shoppers seek out the best possible deal for their purchase.

All in all, I’d put the number closer to 20% though it does vary wildly with the makeup of a program.

I feel attribution technology within the Affiliate Channel is critical.   We developed a lot of technology, and dedicate a lot of ongoing resources to making sure that Advertisers have options available to them so that they can properly commission based on their own data.   I don’t know that we would dedicate as much time to it as we do, if it were only 1% of transactions.

Charles Calabrese of Performance Horizon Group said:

We are seeing an average of anywhere from 3 to 10 sites in the click path – varies by campaign and type.

At least 15% of transactions cross at least 2 publishers in the programs we analyzed.

Kevin Edwards of Affiliate Window wrote:

Obviously we mainly use UK data and the US affiliate make-up is quite different, but we typically see within the affiliate channel, single affiliate interaction sales comprise 80-90% of sales. I don’t think we’ve seen it lower than 75%. Trend wise we see:

* Price comparison ‘lose out’ the most – not surprising given the nature of their activity

* Cashback sites are most likely to be single interaction. They do prompt cookie deletion in their T&Cs but we estimate this happens in fewer than 5% of their sales

* Voucher code sites overwrite the most BUT this is partly down to their reach and they are mostly likely to overwrite other voucher code sites

* We have ‘soft click’ cookies in place for retargeting, toolbars or remarketing which means they cannot overwrite other affiliate links as we know this traffic is considered less incremental and viewed as ‘low hanging fruit’.

* The highest number of interactions is typically five or six and we usually find the average order value increases dependent on number of interactions (with single interaction recording the lowest AOV).

* There is variance between sectors with retail having the highest number of multi-interaction sales

* The data should always be viewed with caution as we cannot cross-device track. With the growth of mobile this creates additional lack of insight.

Dan Cohen of Tradedoubler shared:

Our click path analysis shows that performance marketing channels, often considered a last click channel, have become critical in the early research phase of the purchase journey.

Our internal data demonstrates that the majority of purchase journeys involve multiple performance marketing channels. Our recent award-winning study performance marketing: from first impression to last click underlines the fact that consumers are using performance marketing channels multiple times throughout the purchase journey and across our public network we’re seeing an average of 1.6 clicks per sale.

Finally, Sandra McDill of TagMan chimed in as well:

Across brands and products with longer and more complex order journeys it is very common to see multiple affiliates in the path to conversion. This isn’t surprising when you consider how many different types of affiliates there are at different points of the funnel such as content, price comparison, voucher and loyalty. What is interesting is how often clients assume voucher sites for example are only found at the end of the order journey on last click, but in many cases voucher sites start and assist in journeys they are normally not rewarded for. Having this visibility through a tag management system like TagMan enables clients to properly attribute value to these affiliates when making decisions about budgets and planning future marketing activities.

For every client and every journey the number of steps, and number of affiliates in each journey is wildly different so it’s very hard to guesstimate. A client across multiple affiliate networks using the full range of affiliate types might easily see 2-3 affiliates in some paths, but then have no affiliates appear in others.

Based on the above observations, it would be safe to conclude that anywhere between 15% and 25% of affiliate transactions involve more than one publisher in the click stream to conversion.

If you are tracking things like these internally, and have your own data to share, by all means, do chime in with your comments below.