Whose commission is it anyway?

Alas! With performance-based marketing, where you only pay a commission on actual sales, it seemed like we’d finally found the ideal ROI-focused means of customer acquisition. Things were going fine until it began to dawn us that we were paying more than once for the same customer. How come?

The problem occurs due to the duplication of counting across acquisition tracking solutions. For instance, if I’m doing affiliate marketing, online advertising and paid search, each with its own tracking solution, each with its own performance-based metrics (and commissions paid in some cases), then it is possible, when a sale or other conversion event occurs, that each will claim the scalp, and subsequent reward, for that customer.

And it gets even more complicated when you factor in post-impression sales or ‘view-through’ sales. We have come to acknowledge that a significant proportion of conversion events occur in a window of time after the initial click, view, interaction or other exposure. So multiple sources, multiple time frames. And then there’s offline – how many sales are actually the result of offline marketing but still get attributed to online, most likely as a post-impression sale?

There’s also spyware and “cookie stuffing” where the unsuspecting merchant again risks being charged multiple times for the same sale, or where false pretenders are usurping the genuine claimant to commission.

These problems are not inconsequential. Talking to a large retailer recently they reported an instance where they were being charged eight times for the same end transaction.

So what to do? Currently the industry is grappling with the problem in three ways:

1. Set your policy and stick to it
You might choose to pay your commission to whoever first referred your customer. Or whoever referred the customer for the session they converted. Or you might want to split the commission. Whichever you do, you are likely to upset someone, but you’ll just have to adjust your commercial incentive structure accordingly and they’ll have to live with it. Different sorts of businesses will suit different policies. The message for now is to measure and analyse the commercial implications for you and devise policies accordingly.

The more advanced are looking at policies which are different depending on the nature of the campaign or conversion event. For example, where there is a lot of accompanying offline marketing activity you might want to stipulate a much shorter cookie expiry window so you only pay out on online referrals which convert within a day, or even hours.

2. Integrate your tracking technologies
Easier said than done as there are very few online marketing platforms out there which truly integrate all acquisition methods with a ‘universal tag’ to ensure consistent reporting. The web analytics tools can do it but you’ll still have to grapple with inaccuracies in uniquely identifying users, especially over any length of time as multiple device use and cookie clear-outs muddy the picture. It is possible to "reconstitute" unique visitor and lifetime value records once a user does log in (purchase, register etc.) but still quite a messy business.

3. Live with the problem and fudge accordingly
Most of those we’ve talked to say that, at the moment, they’re just having to live with a tolerable level of inaccuracy, which, as it equates to over-payment, means they’re adjusting their commissions downwards to make up the difference. Not ideal, perhaps, but the current reality. This approach is refined by treating different partners differently: for example, you will probably work with your super-affiliates on a different commercial basis than others.

Most agree that the best solution would be a mixture of clear policy setting and ‘universal tag’ technology which would work across all tracking solutions. The likelihood of all the providers agreeing on this unified tagging approach seems slim at the moment. I believe we’ve only just agreed what a page impression is, so plenty to keep us busy for a while yet…

Any thoughts / experiences in this area?

 I think the messages here are right - "set your policy to match your business objectives and implement it" the only real problem is "how"...

We have been undertaking this kind of analysis for some of our customers for some years, without really understanding the magnitude of the problem, until we listened to the comments from attendees at the excellent Masterclass e-consultancy ran last week!

speed-trap’s approach is quite simple. We "classify" each (and every) visit to the site according to "campaign" which generated it. That classification is typically on the basis of landing page (deeplinks), query string parameters, cookies, or referrers. By using a first-party cookie, you can extend this “per visit” analysis to provide a "life history" for each visitor, recording the campaign which originally acquired them and the campaign associated with each subsequent visit, (without getting into the third-party cookie minefield.) We use this same classification process to add behavioural attributes to each visit or visitor (e.g. looker, buyer, searcher etc.) and other attributes can be added (e.g. total purchases etc.)

Prophet can then use this data to implement reporting to support the customer’s chosen "business model" - e.g. "pay the affiliate who send them on the day they purchase" or "pay the guy who first acquired them" or "split it between them" etc. Clearly this kind of approach allows them to more equitably reward the affiliate who in their view played the key role and reward VALUE - and remove the uncertainty...

This same approach is typically extended to cover other types of "campaign" (e.g. e-mail, banner, sponsored search, free search, off-line, promo etc.) and we typically provide customers the opportunity to  implement  "cost per click" costs for each campaign , or "cost per campaign" so that they can see at a glance which campaigns are really acquiring the visitors who matter and realise realistic cost-of-acquisition and lifetime-value metrics, or just assess campaign profitability.

Certainly we have customers who use this data on a daily basis to optimise their campaign spend and tell us that they have acheived between 30% and 100% improvements in Advertising Budget ROI... so it seems to work for them.

I guess we are effectively implementing the “universal tag” approach Ashley recommends – and our OneTagDoesAll™ approach makes that simple for the customer…

Sorry for the commercial – but it does seem to work!

Malcolm Duckett – speed-trap limited…