Smart Data & Analytics

Taming Data With Science

By Adele Gritten

We’ve reached a pivotal moment in the application of data science: data-informed crisp flavours – Snack 2.0, if you will.

But without being flippant, the use of open online data to gauge consumers’ changing tastes – literally in this case – is a very smart use of technology for market research. Actually, what’s even more impressive is how Pepsico UK is now moving the story along from product development, by using open data to inform sales and marketing strategies.

This is just one example of how data science, used intelligently, can answer business-critical questions and why it’s increasingly coming to feature in boardroom conversations. The analysis of multiple information touchpoints is nothing new – it comes as no great surprise there’s a correlation between cinema ticket/VOD sales and poor weather, for instance.

However, integrating the seven-day weather forecast with a targeted ad server informed by regional data mined from social media channels moves the dial a bit further. This is a fairly basic scenario, but combining multiple online data sources with legacy (i.e. offline) research methodologies and predictive analytics to offer actionable insights can be transformative.

We all recognise that outdated notions of demography no longer cut it, especially in isolation. Consumers are complex, multi-faceted and often fickle, while brand affiliation is becoming a thing of the past and modern businesses need to speak to individual needs. The time for personalisation at scale is upon us.

Research methodologies need to be as multi-layered as their subjects and contextual data is key. There are now simply too many variables for human research teams to track effectively, especially on such a granular level in a real-time environment. So, realistically, machine learning is the only way the sector can develop. However, as I’ve addressed previously, we aren’t doing ourselves out of our jobs. Without human intuition to interpret the metrics, much of the data will remain meaningless.

So where to start? As per the Pepsico example, the first step is to identify exactly what the business objective is and work backwards from there. Often, there’s some existing data that offers the germ of a hypothesis to build upon.

In this case the requirement was to identify emerging taste trends, but other typical business questions might include: how do I break into new vertical sectors, how do I minimise customer churn, where can I find new subscribers, how much should I budget for advertising and promotion to reach a particular audience, etc?

Internal business data taken from – as appropriate to the sector – CRM, POS, EMR etc. can be aligned with traditional brand tracking and both structured and unstructured customer experience data collection from open sources, and in real time if necessary. When interpreted correctly, data drawn from a mix of sources allows you to identify the motivations of existing customers and/or prospects and puts you in the best position to respond to those needs.

However, what’s really powerful is being able to overlay these findings with predictive analytics to interrogate the “what if” scenarios. For instance, we’ve recently worked with a client in the healthcare sector to identify what impact different levels of promotion would have on driving subscriptions; and what would be the optimal times of the year for making that investment.

Open online sources – notably social media – mean data is now readily available and continuous digital data collection stands as little more than a hygiene factor. This doesn’t mean we can become too reliant on data, however. Although the value of the standalone ‘survey’ becomes further diminished with each passing year, the legacy qualitative skills that underpin the sector still have a key role to play.

Combining data science with human intuition offers the best of both worlds – the speed and efficiency of the machine with the contextual understanding that only a human can offer.

The value lies in interpretation and consultation to offer balanced conclusions that will maximise the impact on the bottom line.

There’s little doubt that a data-informed approach will open up new opportunities, and we’re likely to see some exciting and perhaps even unexpected outcomes across multiple sectors. As we gain an even better understanding of the nuances of customer demand, we’ll identify fresh markets, audiences and revenue streams. It seems highly likely we will also see some unexpected pivots in strategy from long-established businesses.

In a similar vein, there’s little doubt the incumbent research firms will need to become much more data science-savvy if they are to retain market share. However, I believe our experience in traditional skills, gained over decades, will give us a clear advantage over emerging and purely data-centric competitors. The future of this sector will belong to those that are able to combine both data artistry and science.

By Adele Gritten, UK MD of Future Thinking

Leave a Comment

* By using this form you agree with the storage and handling of your data by this website.
Please note that your e-mail address will not be publicly displayed.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Related Articles