The solutions
1) The major problem and what changed now, that brought big data and analytics into sudden limelight, is that many executives are starting to embrace analytics and use it in decision making added to their conventional knowledge.
A simple example, an organisation does market research and launched a new product targeting a particular segment, let’s say: Femalebetween 18-35, with secondary school education.
Based on the assumption and research, you expect the target audience to come and yes they were attracted to the product but just 20% of them are in the target segment you want, 80 are not. On the average and in the first month, this segment formed the highest number of adopters and like it happens in the conventional wisdom, we think the product was hitting the right target. But alas!, a year into the project ,your explanatory analytics showed that 60 per cent are Male, aged 40-65 working-class men.
In a situation like this, when the brand intents to target the segment directly, conventional wisdom is that we have a five-year plan and we must roll it out (more advertising and share of voice), but with this explanatory result, the brand needs to change course. So what will happen if the brand has already committed hundreds of millions of Naira to the project? Will the executive team be willing to pull the plug or change course or will they stick to conventional wisdom and wish this away?. If you are not ready to take analytics insights and use it as it is meant to, it’s better not to invest in acquiring the tools, and tech people to support your quest, because the result will just be the same or even put your organisation in a worse financial shape. If the your results remain flat and you have invested in technology and human resource, then the cost of the technology and resource put you at a negative(worse position than you were the period before the project).
Deeper analytics could actually show that larger percentage of men actually buy the product for the women, although they are not the target, their factor of interest ( a variable that works independently of other variables to achieve a desired result, confound) makes them buy the product although they might never have seen your ad or campaign before.
If the goal for the campaign is revenue this is fine (since the revenues are coming in), but you must now make efforts to communicate to this men where they are and how they like it, while making you top of mind.But if your goal is to increase female customer base so you can upsell them later with other offerings, then you need to make a decision on the campaign.
If executives are not ready to take the risk, then they shouldn’t or else what they will get will be data analytics project that will provide an illusion of actionable insights instead of actionable insights itself.
2) For brands to succeed in analytics projects, there is need to get the role of an analytics translator or interpreter in place. An analytics translator takes business goals and turns in into simplified information for the data scientist. The scientist in-turn gets the information requested and sends back to the translator. The translator then matches the information with the specific business goals and present to executives the recommended experiments, based on explanatory and predictive analytics done.
This experiments will be used to validate the possibility of achieving the desired outcome from the project. Very few people can fill this role, because it is rare to find people with the deep business skill/qualifications, an in-depth working knowledge of data science, critical thinking, analytical and a knack for solving problems.
This individual simplifies what needs to be done(What looks so big gets broken down and made actionable),gives clarity to the team, pointing them to the exact drivers that needs to be focused on to get the desired outcome.
Lack of individuals with this skill causes a high rate of analytics project failure.
And before I forget, every executive making decision today must be armed with knowledge to know what good analytics looks like and what bad analytics look like, so that a bad analytics report is not presented as good. Decisions made on bad analytics data can’t come out good.
–––