LONDON: Big data is the driving force behind modern marketing, but it can seem overwhelming at times, so marketers should adopt six key steps to help them make the most of the mass of information, according to a leading digital specialist.
Verity Gill, Digital Director at Ebiquity, an independent marketing analytics firm, says it is imperative to design any marketing data solution around a clear framework based on marketing objectives.
It is also essential that all parts of the business are familiar with this framework, which should include measurable KPIs across a consumer’s purchase funnel or decision journey, she explains in a new WARC best practice paper.
Entitled How to build a data strategy, Gill’s article goes on to advise marketers to make use of data management platforms (DMPs) in their planning, which will help to link first, second and third party data sources into the overall system.
This also helps to build a complete picture of the consumer to be targeted, as long as marketers only use datasets that can be trusted.
“Datasets can tell you very different things if they have been set up differently or for different purposes,” Gill writes. “Again, this comes back to what you want to achieve.”
Her third piece of advice concerns the deployment of effective methodologies for extracting useful consumer insights, which might include econometrics (or marketing mix modelling).
However, she warns there is little point in adopting expensive methodologies if staff have not been trained sufficiently. “Make sure you have the in-house skillset to implement insights, once identified,” she says.
A fourth consideration for marketers is that they shouldn’t use data just for the sake of it, but instead – like Amazon, American Express or Netflix – should loop insights back to the consumer to really engage with them and meet their needs.
Marketers should also recognise that there are limits to the insights that can be obtained from data. For example, correct insights can be rejected because not enough statistical proof has been found to support the insight.
Another common mistake, known as a type II error, is that incorrect insights can be erroneously accepted when data correlations do not logically represent the reality of what is happening.
“There are still many blind spots in data, caused by walled gardens and bad execution, and it is important that you don’t expect too much when data simply hasn’t caught up with objectives,” Gill says.
Finally, she states that data should never be used in isolation. “Data is pivotal in making investment decisions, but the importance of the human interpretation of this data is just as crucial,” she says.
“By adding a human layer of interpretation, marketers can protect themselves against replication and invest in areas that have significant and real potential.”
Sourced from WARC