INside Performance Marketing
Top 50 Industry Players from 2017
Splitting Commissions in AM: Your Views (2 of 3)

Splitting Commissions in AM: Your Views (2 of 3)


After Part One of our 'Splitting Commissions' article, we have the next installment for you.

Joining us for part two are affilinet and dgm.

What are your thoughts on Value attribution?

Tina Judic

Tina Judic (Artemis8): Perceived value is very different from actual value. If an advertiser can prove that one click led to another click that culminated in the user becoming a customer, then I can see the value of this. However, there has to be a significant supporting strategy around Value attribution. If certain partners are ear-marked as sites that support the commencement or continuation of the purchase path journey, and are actively encourage to do so, then I can see how value attribution can work. However, for the partner who has invested to ensure their click led to a sale and, therefore, received 100% of the commission, we could potentially see a reduction in sales volume with the value attribution split; If the profitability of the action is reduced considerably through the reduction in commission awarded.

Helen Southgate

Helen Southgate (Sky): I think this is a tough one, how do you attribute value, can you say that a display ad for example on facebook had less impact than a listing on an affiliate site? I would say it is arbitrary to try and assign a value to each touch point and really tough. eBay have obviously done this but from what I can make out it is a bit of a blackbox as to how they attribute value and I guess it would be completely different for each client. I would worry that as an affiliate I would not be able to get a clear understanding of my ROI.

Patrick Lynch dgm Patrick Lynch (dgm): The last click model has served the industry well over the last ten years, however as the industry has developed and new affiliate channels have come to the forefront this system is starting to look dated. In the future I would expect that all campaigns will have some form of attribution so that all the channels that affiliates are working in are rewarded based on their value within the purchase process. That value may change on a campaign by campaign basis and finding the right mix of attribution will be the challenge for affiliate account management teams of the future. At dgm we have had the ability to run attribution based campaigns and are currently working with some clients on modeling out the effect this will have on their campaigns.

Clarke Duncan

Clarke Duncan (PaidOnResults): If it's about rewarding all parties in a fair way then I think it's worth looking in to more, if it's about how can we pay Affiliates less as someone values another channel more then I can see issues leading to that, as it stands most Merchants have some idea as to what they can afford to pay different channels so really we are talking about 2 systems, one that covers all marketing and one that covers elements inside of marketing within a specific channel.

Peter Rowe

Peter Rowe (affilinet): In my opinion Value Attribution is actually the key element in the multi attribution debate. My concern with multi attribution is that it doesn’t necessarily reward value. A blanket split between the various touch points seems too simplistic and doesn’t take into account the value added by each touch point. Just because a consumer clicked a link does not mean it was of value and we need to dig deeper to understand the buying process.

Commercially and operationally this is easier said than done. I read with interest Part 1 yesterday and agree with many of the comments that we are some way off a solution to the numerous problems of multi / value attribution. With so many tracking options available online there is a technical solution for many of these problems but we face many more issues with standardisation, affiliate buy-in and making these models commercially viable for all concerned. Overall, whatever route the industry chooses I do believe we should work collaboratively on these broader issues.

Hero Hero Grigoraki (Webgains): Value attribution models can be beneficial, but not as industry standards, more for individual merchants. There is no “one size fits all” value attribution model, as each merchant is different – so it needs to be evaluated on a per merchant basis. Of course, affiliate networks will have to support merchant decisions, provided affiliates are protected, rather than shortchanged.

Kevin Edwards

Kevin Edwards (Affiliate Window / IAB Chair): Value attribution is something I’ve been advocating for the best part of a year. I presented my thoughts on what a value attribution model could look like at the a4uexpo in 2009 (last third of the presentation). I think we need to educate merchants about how they look at affiliate marketing. Networks have a responsibility to encourage advertisers to start creating an ongoing cycle of information sharing. They need to identify what’s important to them (new customer acquisition, contribution in the sales journey, engagement, average spend, persistency and so on) and pass this key performance data back to affiliates not only so affiliates know more about their traffic but also so merchants know where value is being driven.

Jason Dale

Jason Dale (Loquax): I think it's possibly a more sensible approach than split commissions. I'd like more merchants to assess what it is an affiliate is doing and reward them accordingly.

Perhaps the value attribution approach will help some merchants remove the rigidity of their %age commission structures and pay on leads for example - if appropriate of course!

Julia Stent

Julia Stent ( Any payment model where merchants are looking in depth at how they reward affiliates and putting these learnings into practise is a good thing in my book. Value attribution can take a number of different forms - from a fully automated algorithmic model like eBay's Quality Click Price model down to simpler things like incentivising new v returning customers. In it's simplest form value attribution can be one of the easiest solutions to achieve - for example, giving appropriate commission increases to affiliates who consistently drive higher average order values would require no technical development for most merchants. Value attribution is a way of encouraging quality rather than just quantity, which can provide great returns for merchants who adopt the right strategy for their programme.

Duncan Popham

Duncan Popham (Total Search Solutions): Although I think that a move towards an increased focus on value is imperative, I do have concerns on the level of transparency being provided. As payment is clearly the key driver for all affiliates, the methodology for determining payment should always be clearly understood by both parties. I think Ebay's quality click pricing is an extremely interesting move, I don't think the industry should collectively move towards a less transparent system.

Alison Guise

Alison Guise (Commission Junction): Whilst the last click model has provided a reliable industry standard for years now, the continued dominance of certain channels such as voucher codes and cashback sites means that other channels are consistently appearing to under-perform. For performance marketing to work ‘fairly’ then content and comparison sites (for example) will need to be paid for their efforts.

Creating a world of value-attributed payments will incur many hurdles. It will be very challenging to create a standard and provide transparency on why an affiliate was paid their piece and why – however it can be done! Even more challenging though is how to deal with certain customer-facing propositions such as cashback sites (as they display the payout often well in advance of the shopping process). So does this mean that cashbacks would have to remain the same?

All of these issues (and doubtless more) will have to be answered and for that reason I think we are a while away from moving off the last click model but as an industry we should certainly start moving in that direction

Is there room to move towards a Cost Per Engagement model?

Hero Hero Grigoraki (Webgains): CPE as part of a value attribution commission scheme could potentially work for certain types of merchants, especially those who pay fixed bounty commissions. It would however require a very detailed analysis of the customer path, assign value to each touch point as well as evaluate the lifetime customer value; data which the majority of merchants at this stage do not have. It would be very difficult to apply it to those merchants who pay a percentage of the basket value, as it would overcomplicate things, especially for affiliates such as loyalty schemes. Would the industry then have to switch to bounty commissions, like the US market?

CPE however could potentially be the metric that will enable FMCGs to finally enter the online, and the affiliate, arena, which can only be a beneficial thing for everyone. It does not however need to either replace the CPA model or the last referrer method; they can work alongside each other, as they define a different type of activity from the consumer.

Helen Southgate

Helen Southgate (Sky): Yes, but again the difficulty is in how you would attribute a value to this. From a client point of view moving to pay on engagement in addition to acquisition means more cost unless you can somehow split, but assigning this again would be quite a challenge.

Clarke Duncan

Clarke Duncan (PaidOnResults): Depends on if the market actually wants it or if we all trying to be real clever and come up with "new Coke" when the current drink tastes ok. But hey even Coke and Pepsi have found room in the market for them so I am sure some people will love it and others will be happy to stick to a simple easy to understand model that is working.

Kevin Edwards

Kevin Edwards (Affiliate Window / IAB Chair): Cost per Engagement is value attribution by a different name in my opinion. It references ‘value’ beyond an arbitrary click. I think it’s inevitable advertisers will look to better target their marketing spends. As all online matures suppliers will need to be more visible about what delivers ROI to specified goals. There are two interesting by-products of CPE or CPV models. They are greater buy in from merchants as they have increased confidence that they are spending more wisely and the potential introduction of non-transactional advertisers who are looking to engage with consumers. Affiliates have this traffic available and are well placed to market themselves and build strong affinity relationships with these advertisers.

Alison Guise

Alison Guise (Commission Junction): As often discussed, merchants should be looking back down chain to see where they can put tools in place to drive customer interaction, and pay for it. The first stage of online commerce enabled merchants to quantify their marketing/PR budgets and optimise for value. Current technologies can now offer them the ability to clearly assign value and track the delivery of the engagement; so it is to some, like me, a natural next step.

Patrick Lynch dgm

Patrick Lynch (dgm): I think moving to a cost per engagement model may take the affiliate channel away from it's core strength which is return on investment based performance. Affiliate Marketing has always been firmly focused on delivering sales or leads and this transparency and simplicity has been a key selling point for many merchants. The Cost Per Engagement model could lead to greater cost to the merchant if not linked back to an effective CPA for each unique sale.

Jason Dale

Jason Dale (Loquax): I think there's room to include CPE within affiliate marketing. For a site like Loquax the opportunity to have advertising that would engage our audience and earn us revenue would be nice.

Perhaps the issue isn't trying to search for a one size fits all model in affiliate marketing but for each merchant to devise a set up that is appropriate for them. Then provided there's clarity about the program, good data to make comparisons between what they do and what a competitor does etc... then that could be ideal.

It seems that there's a hunt on for The Holy Grail of attributing commission in AM, but what's good for one may be useless for another.

Julia Stent

Julia Stent ( Online advertising is awash with a thousand and one different metrics for payment. They can get as weird and wonderful as an advertiser or agency would like (the latest to come onto my radar was CPRV - Cost per Returning View), but in my opinion there are those metrics that fall within the 'affiliate' pot and those that don't. I've seen affiliate marketing grow massively over the last few years and I believe the key to its success is the minimal risk to an advertiser - having worked merchant-side I've definitely seen this in practise. For most merchants who sell a product or service, I'm not sure a Cost per Engagement model falls within the realm of what would traditionally be thought of as 'affiliate' activity and an industry shift this way could be detrimental to the core of our success.

_There are a lot of affiliates who are starting to work with advertisers on mixtures of payment models, such as CPC, CPE, CPV etc and there is no reason at all why an affiliate cannot also act like other online partners the merchant might work with and advertise for them in this way - we do it here at However I have a suspicion that a lot of the time when this kind of activity is taking place and the advertiser is intending to use it to generate more sales, non-CPA models are being used because the affiliate's activity is key to customer journeys yet doesn't drive the last click. Other models such as multi-attribution could be a very valid alternative while still keeping the core principle of centring around a cost per sale. _

Tina Judic

Tina Judic (Artemis8): Yes, potentially, but only if there is a solid strategy defined. This will be highly reliant on the expectation of each partner and the deliverability of each partner to truly work.


Maria O'Flynn (TradeDoubler): Yes – especially for emerging affiliate sectors where they are showcasing client brands and products without the need for an actual visit to the client site. Technology and tracking are key here, and areas where we are investing significant effort.

Continue the conversation

Got a question or comment – tweet Chris chrisjohnski or comment on Twitter, Facebook or LinkedIN.

Chris Johnson

Chris Johnson

Content Director for PerformanceIN. Based in Bristol & London.

Read more from Chris