Behavioural Targeting (BT) is one of the hottest things in online marketing right now, with advertisers and publishers trying to get the best response possible out of the billions of banner ad impressions served everyday across the internet.

For advertisers it means spending their advertiser dollars more effectively and for website owners it means getting paid more for their available inventory; but for consumers the jury is still out on whether it means more relevant ads, or an invasion of privacy by an unseen ‘big brother’ looking over their shoulder.

 So what is BT and how does it work? BT works by using a user’s browser cookie to track their activities across a website or network of websites. The BT engine then decides, based on the pages a user visits (their "behaviour”), what attributes should be assigned to that user (cookie).

The attributes can take the form of demographic type states like "probably male in their 20s” or "probably female in their 40s”, or in some systems "Personas” are used, for example "green fingers” or "early adopter”. BT systems can also guess at a user’s "Intent” based on their behaviour, for example "interested in buying a new family car” or "planning a wedding”.

These attributes are determined using a range of technologies and are most effective when deployed across a network of hundreds or thousands of sites. The most popular technologies used include ‘rules-based’, where simple hard coded logic is used to set an attribute, then there are self-learning systems that bucket people with similar behaviour together, including their likelihood to click on a particular category of ad or complete a specific type of action.

The self-learning systems can also be combined with the third most popular technology, ‘profiles’. Profile targeting captures information directly from the user through site registration or online surveys. This control group of users can then be used to model the behaviour of all the other users tracked across the network.

 BT is important, as it supports the desire of advertisers to be more targeted in their online campaigns. Marketers are moving away from "buying sites” and moving towards "buying audiences”.

The second major shift in online advertising that is driving BT is the movement away from CPM (cost per ad impression) pricing and the move towards performance-based pricing (cost per click or action). This shift means that website owners no longer get paid for showing an ad (impression); they only get paid if someone clicks or clicks and converts. This means that it is vitally important to put the right ad in front of the right person.

As way of illustration, here is a real-life example from the New Zealand ad network TPN.co.nz. Advertiser A is a cheap flights website. Advertiser B is the publisher of a bestselling book.

If Advertiser A (cheap flights) is shown a fixed number of times to an audience aged 20-30 then it generates 8600 clicks. If the same ad is shown the same number of times to people aged over 50 then it generates only 1220 clicks!

Conversely if Advertiser B (book) is shown to the 20-30 age group then it generates only 2100 clicks; however when shown to the 50+ age group it generates 6300 clicks! Therefore it can be seen that TPN would get paid for 29,800 clicks if the ads were all targeted at the right audience and only 18,220 clicks if no targeting was used. That's a 61% increase in ad revenue.

 It has also been shown that in many cases the likelihood to click also increases the likelihood to convert, so the advertiser gets more clicks from the available inventory and more conversions from the clicks they buy.

So who are the big international players using this type of technology? Well the largest and most powerful is Google of course. Via their search, video, email, Adsense and Ad network they cover most of the users on the planet and can, and do, collect a wealth of information about users’ behaviour that can then be used to target and optimise ads across their networks. To see some of the information Google knows about you, visit www.google.com/ads/preferences.

One of the other most interesting companies is Facebook, which has access to an unprecedented amount of profile, relationship and behavioural data. Facebook even allows advertisers to target users based on very specific information. For example you can target 140 single men who are 25, live in Auckland and who are interested in cricket.

So all this sounds great: advertisers can reach their target audience, website owners can get paid more for their valuable inventory and users get more relevant ads.

However there is a range of issues that arise around users’ privacy and the topic is currently being debated at the highest levels in the US and Europe.

The main principle BT systems adhere to as a means of avoiding breach of relevant privacy legislation is to refrain from storing any information (without specific permission from the user) that identifies the person, for example their name, email address or full postal address. But it is also wise to store only the information you need to store and to aggregate or delete data when it is not needed. It is also best practice to encrypt the cookies and any stored information to ensure an extra level of data security, both when the information is transferred across the internet, or stored on the user’s machine or on the BT ad server.

The other main principle is to offer an opt-out option and there are a number of programs that have tried to co-ordinate all the different BT systems into a single opt-out – www.networkadvertising.org is the most popular in the States. The only issue with opt-out systems is that if you clear your cookies then the opt-out is lost and you have to opt-out again.

For the diehard privacy user there is a range of ad blockers and internet security packages that provide more permanent protection. However, because most users are unaware of the capability of BT and other targeting technologies, privacy advocates continue to take offence at the current state of affairs and the issue is likely to remain controversial.

Admittedly this type of targeting technology can be easily abused to breach someone’s privacy as soon as they register on a website. For example if you advertised a job on Facebook using their targeted ad service, then you could break your campaign into age groups and then send each age group to a different landing page and capture that information, along with the job application data. You would be breaking quite a few laws in the process, but the fact that it is so easy to breach someone’s privacy like this is a real concern for many and it highlights the need for in-depth understanding and data protection when dealing with user data on the internet.