Along with Big Data, terms like Data Science and Analytics, Machine Learning, Artificial Intelligence, NLP (Natural Language Processing) have come to the fore over the recent past together with scarcity of talent in these areas and thus top dollars are being paid to acquire such talent. Furthermore, according to IDC Research, digital data is expected to grow at a compounded growth rate of 42% through 2020. With about 1.7MB of data expected to be generated every second for every human being on the planet, digital data levels are expected to grow to 44 trillion gigabytes.

We take a look at the crucial role that analytics plays. Brands (and data) have been around for decades and there have been success stories even before ‘data science’ came into being as a discipline, but the playing field has changed. Whilst (structured) data has always been around ever since brands have been (call logs, sales volumes split by numerous parameters/segments, and lots more),  it was often ignored and remained in the backyard. Brands which have not been able to keep a close watch on the hugely increasing and overflowing streams of data and resulting insights have found it tough to survive, sometimes falling to inexperienced but smarter players or even ‘start-ups’.

Big Data (best defined by the four key characteristics or 4 V’s of such data which are Velocity, Veracity, Volume and Variety) has added great complexity to the role of analytics. Which means we now have huge ‘volumes’ of data measurable in scales like petabytes (1 Petabyte = 1,000 TB) from a ‘variety’ of sources directly from the consumer (hence ‘veracity’) and in realtime (at high ‘velocity’).

These massive data sets generated from multiple sources are invariably ‘not clean’. Precious nuggets of information to be derived from the data are often either hidden amongst not so relevant bits, or incongruous fields since its from varied and unstructured sources (i.e social, internet of things, etc). This forms the starting point and the most time consuming part of data analytics or a data scientist’s job, almost 80% of time being spent towards preparing and managing data for analytics.

Josh Wills (former director of Data Science at Cloudera) once said : “I’m a data janitor. That’s the sexiest job of the 21st century. It’s very flattering, but it’s also a little baffling.”

From Data to Insights

Data by itself, and especially ‘Big Data’ can be overwhelming but it’s the insights that must be drawn from it which make it so valuable and thus the role of data scientists is a priceless one.

For Singapore-based e-commerce company, Lazada, its big data journey is “just starting” says their Singapore based CEO and co-founder, Martell Hardenberg. “The data that we collect allows us to better understand the shoppers’ profiles and general habits. From here, we are able to streamline how shoppers find products through targeted marketing and offers, better product recommendations, and more relevant and interesting products and product mix on display; which results in a better user experience,”

The big data team of Lazada is responsible for handling data science, data engineering, as well as search and data tools. It matches people with the products they are looking for.

Relevance to Marketing

The left brain vs right brain theories have been researched in some quarters and proven to be a myth. Analysis (left-brain dominant) is arguably the most important piece of any marketing campaign, even though marketers often need to appeal to emotion and empathy (right-brain dominant).

Unless data is analysed well, it won’t lead to meaningful insights or conclusions that would aid better marketing. To do this, marketers need to practice the following:

  1. Seek the answers by asking the right questions and probes e.g a display campaign may be getting very high impressions and even good CTRs, with terribly low engagement levels and high bounce rates (even though low download times), making it likely therefore that the ad is appealing, but not to the right audience who are quickly exiting.
  2. Aggregate and extract from multiple data sets or sources, vs relying on one data set e.g quantitative data reflecting overall increase in volumes (across SKUs) may pose a picture of comfort for a brand but a deeper look at the most profitable SKU reflects steady and gradual declining sales owing to quality issues (reflected through sentiment analysis picked up by social listening tools) may indicate an alert.
  3. Specify and measure based on the right metrics and not get carried away by the numbers .e.g many believe that ‘Likes’ on Facebook are a direct indicator of ‘popularity’ of a brand, whereas in reality this number by itself says nothing towards the disposition of fans towards the brand.
  4. Test the right hypothesis and generate actionable conclusions i.e delve deeper into the conclusions to come up with tangible and feasible actionables as takeaways.
  5. Avoid being a perfectionist It is great to have the most optimal output but the quest for perfection may delay taking necessary action which could also prove costly. Hence, best to make gradual improvisations based on learnings by continuing to keep a finger on the data pulse and gauging from consumer reactions.

Insights to Ideas, and Beyond

Ideas don’t simply germinate in a vacuum, and behind every successful idea lies an insight, which could be lying buried in an ocean of data.

Traditionally, insights were gathered largely through market research. Today with Social Media and some of the analytics tools developed including Social Listening which use Image Recognition and NLP (part of Data Science), the data if properly analysed have been put to great use towards creation of better content or product innovation.

We take for example the origin of Instagram Stories – humans are wired to listen to stories, love stories, and tell stories. Berkeley scientists discovered that the body produces oxytocin when exposed to character-driven stories.

An article in the Havard Business Review explains that storytelling prompts the body to produce the neurochemical oxytocin which is a “feel-good” hormone, also released when people demonstrate trust or kindness toward one another.

If your content can prompt the body to produce oxytocin, then you know you’ve done something right. Instagram has exploited this insight in the creation of Instagram Stories which has over 250  million followers, and growing.

Marketers and Left Brain vs Right Brain

Growing volumes can only make the task of ‘data janitors’ (aka data scientists) even more complex and time consuming. Though technology in the form of adequate tools including hardware and software is keeping pace, there appears to be a crying need for this breed.

For content creators and marketers, success will in a way continue to be measured by how well we appeal to the emotions. Data flow and opportunities to leverage it towards insights through analytics (right-brained) will keep growing even faster.

This doesn’t imply that marketers must make a beeline towards becoming data scientists, but now have an opportunity and need to derive even better insights through analytics, and to avoid being drowned by the data avalanche.

(To read more on Salim Khubchandani’s writings, sign up to the Marketing Magazine. The article was first published on Marketing Magazine on January, issue no 215.)

Leave a Reply

Your email address will not be published. Required fields are marked *