It is becoming both more important and more possible to segment audiences into not just demographic personas but also behavioural personas.
Action and reaction
It’s a infrequently made connection that the phrase “be careful what you wish for” is a guiding principle of effective deployment of predictive analysis. The reason is as follows:
- you may well get good results from the analysis to a question you deem important; but,
- have you considered a priori if your organisation is capable of acting on that answer? and,
- have you considered if acting on the answer will produce the reaction you desire from the selected audience?
The math may show a strong correlation between an action and an outcome but unless that math was applied to the segmentation to be used for the action, and the organisation is capable of making segmented offers of value, then the math is worthless.
Creative isn’t what it is cracked up to be
In the recent CodeFuel Online User Behaviour Study (November 2016) explored motives for clicking on ads. It found that only 17% of people gave the primary reason for clicking on an ad as “it being creative”. On the other hand 39% would click on “a good offer”. A full 52% would click on a brand that they liked or if the content was relevant and valuable.
The study also found that 30% of people almost always skip a video ad, and 69% skip always, most of the time or sometimes. What makes people watch a video, more than creative content or brand relationship is “if it has interesting content” at 41%.
Need for analytics before predictive
This all means that before running to heavy hitting solutions incorporating such things as predictive analytics that you must know your audience and their behaviours.
Just “creativity” with non-video online ads isn’t going to generate good value for money. In fact, even worse than the stats above was the CodeFuel finding that 43% of people will leave a website if it has pop-up ads. Therefore a pop-up “creative” ad is of very little value.
The initial analytics that are needed are those which enable you to understand and segment your audience. This is from a combination of your own CRM, marketing automation, ecommerce data, 2nd party data, social listening data, campaign response data, engagement performance and perhaps social personality insights for example.
Within the “traditional” persona defintions for your business you need to then overlay the ad type e.g. interesting offer, relevant content, brand affiliation, creative etc.
Then you can run various campaigns – both owned and paid – which are classified and tagged against the full range of persona characteristics. When sufficient performance and response data is available against each “cell” of your “campaigns versus personas” matrix THEN you can task predictive analysis to explore how to optimise your ROI.
Include personas which inlcude ad clicking behaviour
What’s clear from the report is that creative isn’t the answer. The answer is using personas which include “ad clicking behaviour” with intergated analytics to be able to send the right ads to the right people at the right time.
And a part of this answer is the ability to aggregate customer experience and social and campaign data in an enterprise CX platform such as Sprinklr.