Cookies are used (among other things) for measuring performance, targeting advertising and personalizing sites. But how reliable are cookies and more importantly, how does that impact our data and decisions?
As a way to identify an individual, third party cookies are unreliable. However, most advertising applications for cookies assume they are used to identify a person consistently over time.
How Reliable are Cookies?
Using Google Ad Planner, you can see the difference between the number of estimated cookies and actual people for a site. The difference makes it clear there are significant limitations to using cookie-based data.
Here is comparison data for a range of sites (all stats are for US visitors through Google Ad Planner):
- Yahoo: 190M unique people, 980M unique cookies
- YouTube: 170M unique people, 1.1B unique cookies
- MSN: 120M unique people, 450M unique cookies
- CIO.com: 180K unique people, 470K unique cookies
- Convince & Convert: 15K unique people, 36K unique cookies
- Jeff Bullas: 35K unique people, 93K unique cookies
- MarketingProfs: 120K unique people, 320k unique cookies
On average, cookie-based unique visitor metrics are 3.67 times Google’s estimated number of users. Even more notably the results vary significantly, from 2.4 times up to 6.5 times, across the sites listed. There isn’t a simple rule of thumb you can apply to correct your data for issues with cookie-based tracking.
The Impact of Using Cookies
Rather than fixing cookie-based data, marketers need to understand the impact cookies have on various aspects of their marketing activity and measurement. Here are some of the implications:
1. Online Marketing Measurement
Cookie-based tracking understates online marketing performance. The amount that is missed varies based on a number of factors, but the time it takes from someone to buy, convert, register or otherwise act is a critical factor. If this activity generally happens within minutes, cookie tracking will have a far smaller impact than it if takes weeks.
For online advertising, cookie deletion means reach reporting is often overstated. Note some companies have addressed this by estimating reach based on cookies that are at least a certain number of days old. This is a viable approach where sampling is appropriate and should be applied across a broader range of online marketing metrics.
2. Attribution Modeling
Because cookie deletion happens over time, marketing activities that are generally early in the process will be under-reported relative to late-stage touches. This will increase the reported efficiency of branded search and decrease the efficiency of category search or banners for most marketers.
Attribution modeling here refers to attributing marketing results to multiple marketing touches over time, rather than historical standard practice in online advertising of assigning all credit to the last touch.
3. Audience Targeting
Nearly all audience target relies on cookie data, and both cookie deletion and the use of multiple computers or browsers limits its effectiveness. Audience targeting will be biased towards the most recent activities rather than profiles built up over a longer period of time.
For some marketers this is a positive impact since targeting will be based on the audience’s most recent behavior. However it limits the potential reach of retargeting ads.
4. Site Personalization
Recognizing repeat visitors is only one possible use case. Marketers need to consider what happens when repeat visitors are treated as first time visitors to a site and if there needs to be a path to re-identify or authenticate.
Sites that recognize users should always provide a way for users to authenticate, even if it is not required, so personalization is not lost.
How do you handle cookie deletion or its implications on your marketing? Do you have other questions about the impact of cookie deletion? Leave them in the comments or share them with me on Twitter (@wittlake).