The huge growth of online advertising owes much to the promise of providing more targeted reach than mass media. Unlike traditional newspaper, TV or radio advertising, online advertising can be directed at very narrow audiences to ensure minimal wastage in any campaign. Minimal wastage translates to higher CPMs for publishers, making personalisation a beneficial element from a revenue perspective. Much of this personalisation is done with the help of cookies; small text files that sit on the users computer to store information specific to them.
Some of the most tightly targeted ads come from re-marketing campaigns. These campaigns target users who have previously visited the advertisers website, as these users are seen as being more likely to convert. This is achieved by setting a cookie on the users computer when they visit the advertiser’s website, and then targeting ads only at users who have such a cookie. Because of the higher chance of conversions, advertisers are often willing to pay significantly more for impressions that reach those users. An example of this that many are familiar with is when they view a product on an online store and subsequently see ads for that same product. If this happens to you, then a store owner has set a cookie with an ID of the product that you were viewing.
I understand advertising cookies. What is cookie matching?
In order for this to work, and for the user to be served a targeted ad, the ad request has to be matched to a cookie from an advertiser. This is called “cookie matching” and the more cookie matching that can occur on your inventory the more opportunity there is to serve high-paying targeted ads.
The idea of cookie matching is quite a simple one, but in practice there are a number of challenges to getting cookie matches for every possible ad request, including:
- Users clearing cookies
- Browsers or additional software blocking cookies
- Users changing to another device
- Users browsing within an in-app browser
- Users opting out of cookies/personalisation in a website CMP
There is one other hurdle to cookie-matching and it is quite a big one. Cookies can only be read by the domain that sets them. This means that, for example, when Ad Exchange code runs on your site, that code has no way of accessing the cookies set by the thousands of advertisers looking to find a match for a cookie set by the particular Demand Side Platform they are using. The solution to this is called cookie-syncing. The demand-side and supply-side platforms periodically synchronise their data, attempting to match up the cookies that each of them hold. This is generally done as a batch process rather than happening in real-time. We’ll got into cookie syncing in more depth in a later post, but predictably it is far from exact science that will cause many possible matches to be lost. Be sure to sign up to our newsletter so as not to miss that post.
What can a publisher do to ensure more cookies are matched?
If you are a publisher and the topic of cookie-matching is new to you, then you are probably wondering how you can increase cooking matching rates and improve your revenue. As cookie matching is all handled buy-side, there is unfortunately not too much that a publisher can do to improve it. The biggest impact we can have on it as publishers is ensuring that we have varied demand sources bidding on our inventory. Introducing more and diverse demand partners increases the chances of any of them being able to match a valuable cookie to any given ad request. In fact, this is one of the key drivers behind the revenue gains that header bidding brings. Each SSP bidding will have it’s own set of cookie data to match on, so more partners means more chance to match. There are downsides in having too many partners in play too, but it is beneficial from a cookie matching perspective.
The same is true through Exchange Bidding, although that comes with some additional challenges. Cookie syncing is even more of a challenge in a server to server environment, so match rates tend to be lower. This is one reason why many publishers choose to run header bidding and exchange bidding together.