We are lucky to be working in the digital industry as it provides the most targetable and measurable medium yet in terms of marketing.

With the ability to combine first and third party data, bring offline data into play and combine it with granular behavioural and audience data (such as pages visited, searches made, content viewed, and time spent on certain activities), very specific audience segments can be defined, built and reached.

If you want to build an audience of wealthy families who live in Scotland, are interested in curling and are currently reading regularly reviews on executive cars, as well as looking for quotes on car insurance for example, these people can be identified and targeted accordingly.

Data Management Platforms (DMPs), or Audience Management Platforms (AMPs) if you prefer, provide the technology to collect, store, interrogate and then build audiences at a hyper-targeted level to support the needs of advertisers by providing the specific target audiences they wish to reach.

All this sounds very positive, but when looking at this in the cold light of day, is this really what brand marketers need and is it helping to improve advertising?

Is it beneficial or are we just becoming more efficient in reaching a cookie pool? We can offer them hyper-targeting but can we support this with the metrics they need to highlight success to them? 

All too often, with the wealth of targeting criteria available online, it feels as if we are simply throwing data at marketers on the basis that something may stick, rather than going back to basics and understanding what they need. 

However, just because there is the ability to do something, it does not mean it is necessarily the best course of action. If we go back to the fundamentals of direct marketing, then any segment needs to be distinct, identifiable, able to be reached and economically viable (i.e. large enough) to be worth targeting.

If you cannot execute a segment efficiently at that level, it may be best to ignore it and focus on defining something more broadly. 

Take for example Experian, a company with more than 25 years in developing consumer classification systems across the world.

Mosaic UK, built from over 850m input sources and more than 400 variables, breaks the country’s population into 67 different demographic types that make up 15 broad lifestyle groups. While they could create many more segments, they recognise the fact that too many options will be likely to put marketers off.

While people prefer some degree of choice, too much choice can be demotivating and unnecessary. Any good restaurant will have a relatively limited menu to make decisions easier (and to help manage stock control and cost at the same time).

Meanwhile, those which seem, on the surface, to be more attractive by having a larger menu and more choice actually make it much harder and more confusing for diners to make a decision. It’s exactly the same with audience segments – too many become unwieldy, confusing and often unnecessary. 

When it comes to Mosaic, the important thing is that segments are robust AND can be applied across multiple channels, be this direct marketing, email marketing, online or TV.

From a brand perspective, marketers want to have consistent target audiences that they can reach cross-channel: while online can drive the development of hyper-targeted segments, it becomes very difficult to translate and replicate these same audiences across other environments, making this benefit less attractive to a marketer.     

When it comes to targeting, online can learn much from the offline world about how online segmentation should evolve and be packaged into simple, yet scalable, audience segments.

Marketers should be targeting segments that are selectable and viable (i.e. profitable) while ensuring their desired audiences can be reached across all campaign mediums they use. If we can address these needs to help drive online targeting, perhaps we can make online a more relevant environment for brands and thus further encourage them to embrace digital advertising.