It's well known that the affiliate channel works on a cost per acquisition (CPA) model, meaning that when the desired action is completed, be it a sale, lead or quote,  the affiliate is paid.

To be able to attribute the sale to an affiliate there needs to be an agreed metric by which a sale can be attributed.

Typically this is a 'last click wins' model: if the last click is attributed to a particular publisher then they receive the commission.

The problem with last click

There isn't necessarily anything wrong with this and I am an advocate of this model, which has worked for 15+ years. After all, it is the same approach that is widely used to attribute sales across all digital marketing channels.

However, such a model presents unique challenges to the affiliate channel, as unlike other disciplines, affiliates (and the network) are only ever paid once a last click sale has been made.

At first glance this might seem fair enough and it would be if all digital disciplines were held to the same rules. But they're not. Other channels such as display and search are paid up-front.

So whilst they might be judged on the same criteria to assess the effectiveness of their activity, other digital channels are rewarded regardless of the outcome.  To underline the point, the affiliate channel is not. 

I am increasingly starting to think that by working to a pay per performance model over the years we may have created a rod for our own backs.

Advertisers want sales, therefore they try and drive their channels to acquire those sales, so everyone, not just the affiliate channel, is pushing to be able to deliver that 'last click'.

The 'last click' therefore is becoming increasingly squeezed and the emergence of more sophisticated targeting and retargeting is threatening the affiliate channel's ability to secure the much sought after 'last click.'

In fact our share is in danger of shrinking as other disciplines, especially retargeting, seek to recapture customers that have already been directed to the advertiser by a digital channel, thereby winning the last click and despite the affiliate’s involvement in that customer journey, they end up with zero. 

Revenue not attributed to the affiliate channel

So, the difficulty occurs when the affiliate channel is assisting the sale but isn't the last click. In this instance affiliates aren't being rewarded for their contribution to that sale - whereas other digital channels would have been.

When I worked advertiser side the data I looked at showed that 70% of times the affiliate channel would win on a last click basis, but there was an additional 30% of times where the channel contributed but didn’t win the last click 'race'

That is almost a third and represents a significant amount of revenue that is not attributed to the channel but also recognition of the wider role that affiliates have to play in digital marketing also goes unrecognised. 

Whilst you could argue that the affiliate channel has also benefitted due to the growth of voucher and cashback activity, which as a rule of thumb will win last click the majority of times, it is the affiliates not in this space that are losing out.

Cashback and voucher sites have a key role to play in driving successful affiliate programmes, but perhaps their success has meant that we haven't been as quick to notice that as a whole the affiliate industry is feeling the pinch when it comes to winning that last click.

So, what's the answer?

The last click for me is still the right attribution model to use and is still the best way in my opinion of assigning a sale to a channel most accurately.

There is talk of 'cutting up commission payments' and attributing them according to where the affiliate contributed within a sale. This for me is a lot of work for little return. 

Rather than complicating matters, I think we need to simplify them. To ensure that affiliates are getting the recognition they deserve we need to help advertisers look beyond the last click and understand the role that affiliates play higher up the path to purchase funnel. 

This potentially also means considering different payment models, outside of cost per sale, like other channels currently do. We often say that affiliate marketing is a microcosm of online, in that it crosses all disciplines of digital marketing, yet it doesn’t share the same payment models.

The data for this unfortunately sits outside the affiliate channel, so in many ways it is in the hands of the advertisers and how they want their affiliate programme to develop.

Influencing consumers further up the purchasing funnel

Increasingly, more advertisers want to engage affiliates that influence consumers further up the purchasing funnel. To do this they need to start looking at all of the touch points where different affiliates are adding value. 

Advertisers then need to share that information with their affiliate network and work together to define new payment models that will benefit all parties and ultimately increase ROI, drive more sales and have more eggs in many more different baskets. 

This for me is where affiliate marketing should already be going. However, it can't simply be about obtaining more money within the channel (even if this is fairer); it has to be about using different payment models to drive more sales for advertisers and ultimately increase ROI.

The channel is built on the ability for advertisers to have reach across rich, content driven sites but the last click model no longer supports that. For our industry to grow, we need to start thinking about and proactively pushing forward alternatives.

Helen Southgate

Published 22 October, 2013 by Helen Southgate

Helen Southgate is UK Managing Director at affilinet and a guest blogger on Econsultancy.

5 more posts from this author

You might be interested in

Comments (17)

Save or Cancel
Neil Sampson

Neil Sampson, Head of New Business at DC Storm

As the first attribution provider to split affiliate commissions across all contributors within the path according to true contribution, DC Storm has successfully implemented this new attribution model for ResortHoppa and we are now implementing similar models for multiple major UK brands across the main networks. The results have so far been able to speak for themselves.

Customers today are no longer exclusively ‘voucher code customers’ or ‘cashback customers’, in the same way as they are not restricted to either online or offline channels, or indeed individual devices. A good attribution model will look at performance across all of these elements in order to calculate the statistical influence of each individual touchpoint within the customer’s journey, whether it’s a voucher code click, a banner view or an interaction with a mobile app. It’s only when you take an holistic ‘non-channel’ view of the customer journey that you’ll be able to see which channels, campaigns and marketing messages work well, and which don’t. By focussing your budget on the parts of the mix that work, you’ll see better results for either the same or even less spend.

Our experience so far shows that the effect of paying affiliates according to a multichannel attribution model is that a greater number of contributions are recognised for the affiliate channel, even if the commission for each is smaller. In addition, more affiliates are encouraged to promote the program because they suddenly have a way of generating income outside of the saturated “last click” space.

ResortHoppa experienced a clear increase in effectiveness of the channel (more affiliate contributions rewarded and a natural uplift in overall commission). What’s more, not a single affiliate left the program after the change.

almost 5 years ago


Duncan Popham

Great piece Helen (and a good response from Neil). I've always felt that the biggest barrier to attribution modelling in general is an industry (digital, not necessarily affiliate) to accept that perceived wisdom is at best, inaccurate and at worst, wildly misleading. Last Click Wins may well have been the "best worst" model for most of our collective careers, but that doesn't mean its remotely fit for purpose today (or even that it ever was).
I'm not sure if complicating the channel through either differing payment models, or indeed, complex attribution payment systems. For solutions to be widely adopted, they need to be widely understood, and perceptually beneficial. I don't think either of those models particularly welcome that (although I concur that they can be very effective for very engaged merchants).
I'm starting to come round to a simpler system using assist clicks as a secondary payment trigger - its adaptable and could be adopted cross industry if the will was there. Filtering on affiliate type for strength of payment would allow for strategic influence on rewarding the most relevant partners.

almost 5 years ago


Peter Dickenson

Splitting the commissions is fundamentally flawed as it still skews the payments in the wrong direction for the merchants.

Attribution is very much about understanding paths and the touch points in the buying process. Things are very much past splitting click paths to give attribution values. Some companies are too busy working towards a static cost per acquisition and forgetting that we're in a marketing environment.

The merchants who are using it properly are massively increasing sales, and using affiliates with a variety of payment methods now. They are also controling consumers with the data. Last click is and should be a part of every mix, it's still a reduced risk strategy which works well with unbranded affiliate sites.

We live in a dynamic world now but the basics of marketing still continue - Price, Place, Promotion.....

almost 5 years ago


Chris Garner

I think this is a fantastic article and such a topical debate globally. I am really interested to hear how DC Storm uses attribution modelling for affiliate programs like House of Fraser.

What are the rules of attribution?
How does the technology work?
Is it cookie based (which is flawed anyway because of the huge deletion of cookies)?
How do cashback sites work on attribution of splitting commissions across the customer journey?
And loyalty sites for that matter?
How do bot clicks further up in the click stream of the customer journey get constrained so they are not being rewarded commission for not really contributing to the sale?

Who else is running attribution on the affiliate channel with any success and what tech are they using?



almost 5 years ago

Helen Southgate

Helen Southgate, UK Managing Director at affilinet

I don't think attribution is necessarily the answer if we are talking about splitting commission. There are lots of difficulties with this as you point out Chris, particularly around cashback, and I've always had an issue with how you can assess which part of the journey is most important - first, last, middle, end? It's all arbitrary in my opinion.

Also, my general feeling is some affiliates will win, some will lose out and you'll end up more or less in the same place. So possibly a lot of work for no additional return. I think a good first step, and simpler process would be to look beyond the CPA model.

The key thing is every Advertiser will be different so there isn't a one size fits all solution, but then that's what we're all here for, to look at each advertiser individually and assess what is going to be most effective and cost efficient for them. We are a channel that specialises is acquiring sales however I think we could be even more effective if everyone broadened their way of thinking about the channel.

almost 5 years ago


Peter Potter

Some great points and completely agree with Helen 'Some will win and some will lose'.

almost 5 years ago


Matt Lovell, Head of Customer Data, Insight & Analytics at Eurostar International Ltd.

For me it's really interesting that this debate has been banging around for a few years now and yet, largely as a result of the difficulties Helen flags (in terms of both the cashback front where you can hardly say to a customer - sorry you don't get as much cash back as there were 6 other channels you used before you came through that affiliate and in terms of understanding how valuable an affiliate actually is.

I used to work on a client that as a result of allowing brand bidding for such a long period of time, now had two affiliates who appeared in natural search immediately below them with very similarly named websites. As a result, while a high volume of users interacted with their websites before sale, it was really difficult to actually confirm that they had helped influence the sale as if they hadn't been there, the user would probably have gone to the Brand's own website instead (given 90% of their traffic came from branded searches).

For me, the best way I've found for better rewarding affiliates who are actually delivering value to the brand (judged by looking at those that drive new customers, provide the first touchpoint for a user or add value in terms of relevant content, for example tourist boards and sites with content about destinations in the world of travel) is to understand how many sales they are influencing and then offer them rewards based on driving incremental traffic based on the understanding from your own path to conversion data that this in turn will drive additional sales.

This avoids any complex attribution logic yet at the same time, allows you to identify and incentivise relevant affiliates.

The only other point I'd add is on the voucher codes / cashback websites front, it's a really interesting one for me. Where a brand offers a comparable product to others I can see this potentially being essential (or people will defect) but where you offer something with a point of difference, by appearing on these sites are you not encouraging customers to use them?

almost 5 years ago

Peter Leatherland

Peter Leatherland, Online Sales Manager at Ethical Superstore

I think cashback sites should play a big part in this debate, with them specifically relying on the last click model to pass the commissions back onto the shopper. You mentioned retargeting potentially taking the last click from the affiliate skewing the balance away from them. However most merchants will see this the other way with the affiliate channel being attributed too much with many shoppers arriving through other channels but using a cashback site one their decision has been made and giving the affiliate channel the last click.

Affiliates who aren’t cashback sites would be the ones potentially loosing out but I wonder if many merchants will just bunch all affiliates together and see the channel as a whole being attributed too many sales, even through many content affiliates will be getting less commission than they truly deserve.

One real issue with a more complex attribution model which Matt points out is with shoppers using cahsback sites who may or may not receive the ‘full’ commission depending on where they went before the cashback site, and will actually be encouraged to wipe cookies to avoid losing cashback – not a good thing for any affiliates!

Perhaps a better way to reward affiliates is to change your commission structure to reward sales from new customers more. We run affiliate programmes and have recently changed the structure so we offer a much higher commission to affiliates if the sale they send to us is from a new customer, and a lower rate for an existing customer. This skips the need for complicated attribution models and effectively rewards affiliates who are helping customers discover our site rather than just obtaining the last click. This should have the effect of one the whole giving content affiliates higher commission rates.

I appreciate this doesn’t completely address the problem you talk about but it does redress the balance towards content affiliates who probably feel they are losing commissions which are attributed to cashback sites who got the last click.

almost 5 years ago

Neil Sampson

Neil Sampson, Head of New Business at DC Storm

Attribution modelling should be everything but arbitrary. At DC Storm our attribution models consider the incremental effect that each touchpoint has on the customer’s propensity to purchase. The world is changing and the customer’s path to conversion is becoming longer and more complicated, crossing online and offline channels, and multiple devices. The way things currently stand the affiliate channel is being left behind. For one of our clients after joining the customer journeys we saw over 30% of sales were made across multiple devices. Without joining these journeys together there is no doubt that the affiliate channel is being undervalued.

From an advertiser’s perspective, the affiliate channel is attractive because their investment is directly related to sales. To move the payment model away from CPA would put this position at risk. Yes, there are going to be winners and losers and certainly cashback sites should be treated separately because the end consumer is directly affected by any change. However, at the end of the day an advertiser will be reluctant to spend money on an affiliate that doesn’t add incremental value.

almost 5 years ago


Jenny Barnes, Head of Digital Marketing at Legal & General Investment Management

We use behavioural attribution modelling, so that we do not have to guess which click is the most effective. The algorithms calculate the value of each clicked based on the behaviour the person shows on the website on that visit.

Since implementing this a couple of years ago we quickly discovered the value from Affiliates was greater than we had thought from a last click basis, but we are yet to implement this as a method of rewarding affiliates due to reservations with our affiliate partners, which are stuck with the last click mentalilty.

In the end clients want to spend more on the media channel that works, so if affilaites have nothing to hide they would embrace this fairer method.

almost 5 years ago


Matt Bailey

I think this is a fascinating debate and whilst it is clear that something needs to be done, it is clear from the different viewpoints on offer here that there is no right answer.

One thing I think is key to point out is that this situation has emerged because the last click model has encouraged affiliates to evolve their model (or come up with new models) in order to give themselves the best opportunity to gain that last click. Would a switch in the model result in a more equitable playing field where the current 'losers' are getting rewarded? Or would it result in a shift in behaviour to take advantage of the new rules but provide no added value to the client? For example, cashback users deleting cookies (as Peter mentioned) or clicks being forced so as to feature in the chain?

My other question is around providing visibility to publishers. It is all well and good incentivising publishers to drive specific actions/levels of engagement but without giving them any insight into what they have done I don't see how they can optimise their activity to encourage this.

I feel that this will become an increasingly important question as the momentum towards performance and the need to connect data across channels and partners grows. Brands will want to understand the interactions consumers make and understand the role that different partners/channels play.

almost 5 years ago

Peter Leatherland

Peter Leatherland, Online Sales Manager at Ethical Superstore

@Jenny - I don't think affiliate networks would want to leave the last click model, they would potentially lose many of their big cashback sites who would not want to move away from last click.

Look at it from the cashback's perspective, they deal with many cashback claims and disgruntled shoppers already who have not got their cashback straight away. Explaining to them why it can take months before they get the cash is not easy and even more hard to explain why it may not have tracked in the first place.

They don't want to have to say "you only got half your cashback because you visited some other website first" - Firstly shoppers wouldn't like the idea of there steps being traced and secondly would probably block cookies or use a separate browser for cashback purchases to avoid this and potentially wiping out any insight a merchant would get on the origin of a sale.

almost 5 years ago


Matt Lovell, Head of Customer Data, Insight & Analytics at Eurostar International Ltd.


The other problem with behavioural attribution is that it works on the logic that all customers behave in a similar way when they visit the site and therefore what they do on the website can be used to deduce how valuable that specific visit (and hence the source of that visit was). It also works on the assumption that the source that drove that visit induced that behaviour from the user (rather than something external which resulted in them wanting to visit the site anyway).

As nice as some of the solutions the likes of Qubit, Visual IQ, ClearSaleing, Artemis etc. are, the concern it always seems to raise is whether a little black box can really accurately unpick which channel really influenced a customer's behaviour - I've certainly yet to see something that is truly compelling...

almost 5 years ago

Helen Southgate

Helen Southgate, UK Managing Director at affilinet

@Jenny Interesting model that you are using and I'm surprised to hear that affiliates are not embracing it. Perhaps this is another challenge we have within the channel that there is a reluctance to change and test new models.

@Peter for me it's not about losing the last click model, as I think it's clear this works and has to stay for some channels, not just within affiliate. However, I think there is an opportunity to look at other payment models for other affiliate promotional types, perhaps such as content which are more likely to appear higher up the chain.

@Matt completely agree with you regarding the behavioural attribution model. This has always been my challenge with attribution in general, every customer is different and there are so many factors which will affect their purchasing behaviour very differently. However, perhaps it is a better solution than what we have; it reminds me a bit of the current finger print and cross device tracking solutions. They make assumptions from behaviour which won't be 100% accurate but it's possibly better than what we currently have.

almost 5 years ago


click here

I think lots of affiliats networks which provides good services for customrs and clients . SO affiliats marketing is one of the best ways to increase own business reputation and a biggest resource earning.

almost 5 years ago



Very interesting discussion, can't believe I didn't catch this one earlier.

For us at Affilae we tend to agree with Neil over at DC and also push the split commission rule based attribution model, tracking all touch points and channels contributing to a funnel and even allowing the advertiser to be dynamically included. While it's always evolving, we find keeping the attribution as clear and simple as possible at this point is what works best as the industry progresses in this direction. At this point we try not to push reliance on too many unnecessary added layers of statistical automation or analysis for attribution, it's still imho very important to know your affiliates true value by knowing your affiliates themselves. This is not something at this point I think you can really just purely fully automate and expect just attributions from your affiliates perspective.

In terms of coupon and cash back type sites I don't think they have anything to worry about, if anything a shift away from the last click model is not a negative thing for them anyway, it will ultimately permit them to become more attractive again and stop coming across as leeches in the conversion process, sure they may earn a bit less per conversion but it will prevent them from organically and unintentionally sabotaging all other affiliates in the process and ultimately restoring more credibility and interest among quality content providing affiliates in the whole CPA model which I think is a very important requirement coming into 2014.

almost 5 years ago


Robert Citelli, Principal Founder at @Sales & Marketing

Would love to see you update this story and talk to more Publishers.

The last click model is a scam, IMHO.

The Advertiser gets free advertising and the Publisher gets zero.

Advertisers promote their extended cookies which are meaningless in Last Click In. What's the value of a 60 day Cookie when the user established by a Publisher loses the sales commission to the next guy?

Further, the Advertiser or Portal Dashboard driving multiple Advertisers' programs has a minimum threshold to payout prerequisite. Since sales commissions are marginalized/sliced/diced Publishers don't see any ROI.

Adding insult to injury, the Affiliate Service Provider portals DO NOTHING FOR PUBLISHERS and will say so unabashedly when asked. Their emphasis is solely on keeping their Advertisers happy and paying them. Links are outdated, products served are prior versions, customer service is lame and any excuse is used to make sure no payout is made to the Publisher.

A CPC model would be much more equitable to everyone involved, even if it was a percentage of a penny per click because at least there would be some cash flow. We have generated hundreds of thousands of CTs in the past 12 months. The ROI has been less than meaningless.

Affiliate Portals and Advertisers need to compensate those assisting their sales. The flip side is for Publishers to pursue Direct Sales programs with the Advertiser and we are now actively engaged in doing that because Affiliate Marketing through the major Affiliate Service Providers is a useless waste of time.

over 1 year ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.