There isn’t a marketing channel out there that doesn’t offer some degree of targeting these days. Each channel’s buying platform will offer advice on how best to use their targeting features but remember this; they can only advise what they can offer, and the more people you target the more money they get.
After 20 years’ managing audience/data sales teams I have experienced every tactic used to grow audience size.
The following posts will help you understand how to target an Audience better; what to avoid and what to look out for when buying audiences for customer acquisition or brand targeting.
Let’s start with some common audience types:
Declared – this is where a consumer has declared the actual information on offer. Most commonly this is an age or gender but there are surveys that capture a lot of other information direct from the consumer.
Derived – this is where an assumption is made about an individual’s demographics or interests based on an action they have made. For instance, deriving that someone is interested in a car because they liked a picture of it or assuming someone has pet because they have shared a picture of a cat.
Point of Interest – this could be grouped with derived as it assumes an interest based on the location of a device. For instance assuming that someone is interested in football because their phone location has been tracked to a football stadium. This is common in mobile and OOH marketing.
Transaction – this is where transactions are used to identify what people have bought, how they have paid for those goods, how much they have spent and when. An assumption can be made about interests, buying cycles, preferred buying channels and seasonal buying habits.
Modelled – this is where transactional, customer or declared data is analysed to understand more about the individuals. That analysis data is then used to find look-a-likes based on similar demographics, site visits or other behaviours.
All of the above have their merits and potential pitfalls and on the face of it Declared and Transactional would be the best and most accurate data to use.
These data sources however are the most difficult to obtain and in the light of GDPR the number of sources in this field has dropped significantly. You may want to check how the data is matched into the channel you are using and whether that is accurate enough for what you require.
I will go into matching techniques in subsequent blog posts.
Derived and Modelled can be very useful ways of growing niche audiences. With derived you should be able to find out what the data has been derived from and make a judgement on whether you think that is accurate.
You may have to accept with modelled that you will never really know how the audience has been built as the companies creating these audiences maintain that their “machine learning” and “proprietary algorithms” are confidential intellectual property (Facebook, Google etc). The one thing I would try to check is the accuracy and size of the original seed data.
If you need any assistance navigating the various audience types or you need a custom audience built for you email or call me and I will be glad to help.