What Should my Cookie Length be
From Wiki
Affiliate programmes: what should my cookie length be?
The cookie length is the amount of time that a cookie will remain on a user’s machine after they have clicked a publisher link to an advertiser’s site, unless it is overwritten by another publisher ’s cookie.
Because clicks are an imperfect measure of the influence a publisher has over a user’s decision to buy, the fact that there is a cookie present is not proof that a customer’s decision was made on the basis of visiting the publisher site and clicking through to the advertiser. Nevertheless, after examining data from sales tracked via the Awin network, we have found that the vast majority of publisher sales occur within 24 hours of the cookie being placed.
Knowing this, how should advertisers set their cookie length? On one side is the argument that you can reduce your cookie length on the grounds that most publisher sales will be captured regardless. But on the other hand, is this not just as much an argument for extending them?
Some large programmes, most notably Amazon, use a 24 hour cookie. But is it really fair that if a publisher provides the deciding factor in the user’s decision to buy, that unless they do so within 24 hours the publisher gets nothing? The same question can be asked, to an even greater extent, of programmes that only offer single-session cookies. In recognition of this, the industry standard in the UK has always been 30 days, and often this is accepted by default.
Moreover, some advertisers choose the maximum cookie length – 999 days – likely as a gesture to publisher s that they are on their side - after all, many people do not keep the same computer for this long! - but with the knowledge that they will rarely pay out on anything older than 30 days. If however they do, there is an argument that this customer can only be said to have a very tenuous link to the influence that the publisher has played. This is why some advertisers reject long cookie periods on principle.
Ultimately though, we cannot know everything about a user’s decision-making process, and who should be correctly attributed the sale. After all, we cannot cookie the user themselves, only their devices, of which there are a growing number - be it their home or work PCs, their mobiles, or tablet devices. As we noted above, the fact that the publisher ’s was the last cookie present does not tell us anything about whether or not, or to what extent, they successfully persuaded the customer to purchase. Commissions paid out on cookies older than 30 days only show that that publisher was recognised as the last referrer by the advertiser before the customer transacted. But if the advertiser is happy to reward the publisher on the basis that no other actors can be attributed the sale, then so be it: the publisher should get the commission.
So despite the debate, the final decision is therefore the advertiser’s, and publisher s can accept it or vote with their feet by choosing to promote a competitor. But three factors should be considered when deciding on cookie length.
1. Think about the typical user journey onsite: is this complicated or fairly straight-forward? Cookie lengths should recognise how the on-site conversion compares to the click-through rate from a publisher ’s site. If the OSC is far below the CTR it might indicate a problem with your site’s ability to convert the good traffic that publisher s are sending. So even if the publisher has successfully convinced the user to buy, the cookie length might be increased in recognition that OSC is not perfect. This might be due to onsite design, or simpler things like stock levels. A longer cookie period might therefore be set in recognition of the value of their traffic or the limitations on the site in encouraging conversions of that traffic
2. What is the average order value (AOV)? Low-cost transactions that satisfy an immediate need, from groceries to payday loans, might be seen as justifying a lower cookie period. On the other hand, longer cookie periods would certainly be appropriate to considered purchases like holidays, expensive electricals, or seasonal events with long ‘lead-times’ like Christmas, where the user is likely to begin shopping long before they choose what to buy.
3. Look at what your competitors are doing. The publisher market is increasingly competitive and the cookie length could be a deciding factor in spite of better commissions or EPCs. Perceptions count for a lot in this industry, and as we have seen with advertisers who choose to offer a 999 day cookie, the ‘positive PR’ aspect of cookie length demonstrates commitment to recognising publisher s’ worth.
The bottom line is that it is very important for advertisers to think about the product they are selling, and how the customer might behave before buying, when considering cookie length. Whilst there are better indicators of the way that publisher s have engaged their users and converted them to customers than the click-based cookie, these will differ across publisher s. The golden rule is to be completely transparent about cookie length with your publisher s. This extends to de-duplication and the use of post-view cookies (which drop on impression) within the publisher channel, and whether these overwrite click cookies. So whatever you decide to set your cookie length as, make sure it is clearly advertised in your programme’s description.