Smart data: Creating a strategy that delivers actionable insights

There’s a growing consensus that the term ‘Big Data’ doesn’t fly.

Ironically, it was a marketing ploy coined for marketers, but now the marketers have woken up and want more answers. How are we meant to utilise this ‘big’ data? What’s important in among the droves? What's structured and unstructured data? The buzz-term has been replaced with a need for understanding.

The value of asking questions

According to professor Neil Woodcock and Nick Broomfield, chairman and director of customer management consultancy The Customer Framework respectively, the core values are ingrained in taking the right data and making it work. Turning your Big Data into smart data. But how do you actually do this? What approaches should you take to ensure your approach to data is led by a strategic end-goal or future insight?

“The starting point is for organisations to ask what questions they want to answer,” states Neil Woodcock. “What is the challenge they have? Is it around making the supply chain more efficient and effective by reducing inventory? Is it about customer retention? Is it about making brand content relevant? To understand what the purpose is helps you understand the metadata you require to then understand the sources of your data, so the volumes of data out there suddenly become smaller and more manageable.”

Extracting wisdom from data

For data-driven marketing, the idea of size is becoming polarised by the need for results. However, The Customer Framework recently undertook a survey in partnership with The World Federation of Advertisers (WFA) where 47 member companies - representing $35bn in annual marketing investment – were asked for their views on Big Data. It found that although 88% agreed that ‘Big Data’ is vital for current and future business decision making, a worrying 74% stated they were currently unprepared to take advantage of the opportunities presented.

Broomfield suggests in the early development of a data strategy, businesses should try adopting the ‘DIMKW’ model; where Data is turned into Information, Information is given Meaning and becomes Knowledge, and Knowledge is given Insight to drive Wisdom into decision making.

“The business needs are around deriving actionable insights from data as opposed to focusing purely on the data for data’s sake,” Nick states.

“So for us businesses need to both simplify and get smarter in their thinking and handling of so-called ‘Big Data’. Smart data is about the process and strategy of deriving actionable insights from the voluminous raw data that comes in. The data strategy must define the roadmap for using data in marketing with as much focus  - if not more - on the people and process requirements as the technology implications. Discussions on Big Data often mentions the three Vs – volume, velocity and variety, whereas smart data is more about thinking ‘what data do I actually need?’, ‘where will I get my data?’ and ‘how will I use my data?’. It’s those fundamental insight-driven questions that can make data smart, creative and usable so that the aforementioned opportunities presented can in fact be realised.”

Accurate, actionable and agile data

One focus marketers now take into account is the evolution of loyalty, driven by digital and social channels. As a result, CMOs are being pressured to understand data in these specific terms. But is the data extracted truly ‘smart’? Woodcock believes there needs to be more of a drive towards three key ideals:

“What we tend to work on is thinking about smart data in terms of being the following: Accurate, actionable and agile. Data accuracy is the first fundamental. The data must have enough reliability and precision to drive business value – there is no point basing insight on inaccurate data. ‘Enough reliability’ implies it doesn’t have to be totally accurate, but analysts have to understand the parameters of accuracy. The second fundamental is that the data must be actionable and be able to drive practical, scalable actions that can be applied within the business – and this reinforces the ‘purpose’ point above. Many organisations have access to a mass of data but aren’t able to focus down to the datasets they need to make wise business decisions. Thirdly, data has to be agile – it has to be available in an acceptable timeframe (not necessarily ‘real-time’). There is little point in getting data on late deliveries two days after the event!"

A single-view of your customer

Nick Broomfield adds, “As an example, we do a lot of work in the consumer packaged goods industry. These businesses don’t have traditional access to transactional sales data, and so they come to us and say ‘we’re looking at the role of consumer personal data (normally personally identifiable information or ‘PII’), and we want to augment that with social, attitudinal and behavioural-data'.

“They need to think about making this smart – putting together a  single view of the consumer stored and managed in a flexible database solution; getting to grips with the idea of developing standardised core data sets and understanding how to drive creativity from data and analytics.

Our research with the WFA found that starting work with small data sets can enable marketers to more easily meet with success in identifying insights that can be applied across the business. This helps to demonstrate that it’s worth investing more in the right people and tools to ensure the delivery of relevant, personalised multi-channel experiences that drive incremental business performance and sustainable competitive advantage. ”

Learning to listen

Those multichannel experiences offer businesses an opportunity to build data sets, but also to ‘listen’ more acutely to what consumers are saying and doing, through different channels. While this should be the holy grail for marketers, the act of actually making this information ‘smart’ is embedded in what a business is listening for.

“Buzz monitoring or online listening is an issue,” says Woodcock. “For brands to really understand their customers and how they are feeling, the fundamental of listening is not just to listen for brand mentions - because that is very inside-out - but to listen to consumers’ feelings and thoughts around the events where they are consuming or using your product or brand, so you can really understand how you can position your brand differently and the type of content and messaging you should put out around it.

“Again, the key is to develop a purpose for what you want to listen to – for instance, understanding the ‘big brand idea’ and how effective and relevant it is to your target market. But if you are trying to gauge the impact of, say, omnichannel marketing over the value of solus channels, you’ll set ‘listening analytics’ up differently and you’ll listen differently. There is validity in ‘random’ listening, but in our experience listening with a purpose is gives more specific business actions. Marketers are definitely starting to now understand this. Social media data for analysis has only been around for five or so  years, and the technologies have evolved quickly to carry out analysis above and beyond key words into deeper cognitive and sentiment analysis. This enables marketers to get to the real meaning of what people are saying - despite the slang, bad grammar and colloquialisms - and it is this evolution which is driving social data to become ‘smart’.”

The key is to develop a purpose for what you want to listen to – for instance, understanding the ‘big brand idea’ and how effective and relevant it is to your target market.

Creating a culture of data management

Ultimately, smart data is about being a smart organisation. Creating a culture by which all people place data at the heart of everything they do. While this may sound fairly whimsical, it’s a huge change process that requires new thinking about who should be in charge of making business decisions:

“Smart data, and deriving insight from it, needs to be embedded in driving the organisation forward, not a backroom function,” adds Woodcock. “We talk about data scientists now rather than data analysts, because the role has moved on in its sophistication. It’s more sophisticated because traditional data analysts used to have access mainly to transactional, structured data. Data scientists have to have combined skills; looking at structured data together with unstructured data - particularly from textual and voice inputs.

"They also need to be able to combine transactional (usually structured), research (semi-structured) and social (normally unstructured) data –using data triangulation methods. Transactional data – what have customers done - research data – what are your customers’ more considered views and social or contact-centre data – what are they thinking and saying right now – can be powerful in combination, reinforcing insights and enabling a more powerful argument to business leaders about any change that is required. You will obtain greater insights if you merge those types of data together and are able to make sense out of it!

“But organisations that simply hire data scientists are finding they also need a commercial ‘middle-man’ in order to truly develop this smart data ethos. This is someone who is comfortable with numbers and analysis techniques, but not necessarily a statistician per se, but, vitally, someone who is also very comfortable with dealing with senior managers in the business.

"That person is the interface, sitting in senior-level meetings and understanding the business issues and what the senior guys are worrying about. They can go back to their teams and say ‘we need an insight on x-y-z’. They also need to be persuasive - having the skills and conviction to take insights to the board, even when the insight is not one that business managers want to hear! . So they have to be confident enough to fight their corner and get insight recognised as a powerful contributor to business success.

“These roles are really important in larger-scale organisations. Without the right data and processes, the human understanding and the ability to persuade others in the business to apply the insight, it’s impossible to truly build smart data beyond being just an ideal.”