What is 'Advanced Analytics' and how does it help you avoid basement analytics syndrome?

The Channel 4 comedy 'The IT Crowd' gave us a window into the world of the 'standard nerds' at Reynholm Industries. Roy and Moss worked from a dingy, untidy and unkempt basement – a far cry from the slick London architecture enjoyed by the rest of the business. On day one of the inaugural Chief Analytics Officer Europe conference, the 'Basement Analytics' metaphor was one of the strongest.

I started my keynote speech on day one of the conference with another metaphor, one taken from a previous Pulse article. Serve a courgette to a Brit and an American and you’d likely see a debate about where the zucchini starts and the courgette ends. The need to eat gives way to an acceptance that they are one and the same thing and it’s the Atlantic Ocean that prevents us from using a standard terminology. Let's not get started on the different continental versions of a chip: do you want chips or crisps? Do you want potato crisps or vegetable crisps?

We use the same language to describe different things and different language to describe the same thing.

The vegetable metaphor was intended to probe the sometimes confusing use of language by Chief Analytics Officers. Was it surprising to learn how different we are in our choice of language about our role, our impact and our methodologies?

Are we actually a group of horses and unicorns mixed together – with only our magic horns to distinguish the real from the abstract – or are we all workhorses all focused on the same goal of value for our organisations?

At this emergent stage, our industry is beginning to start to cluster the subtle away from the blindingly obvious and the confusing metaphors. Like Fresher’s in the first few days of University, we are carefully positioning ourselves and assessing those in our group who will lead, those who will teach and those who will crash and burn within the first term. We are beginning to understand how sub groups will form around common interests, passions and frustrations. 

A common language is possible, especially if we understand that 'value' means different things to different groups.

For government, the focus is on transparency, accountability and enabling better customer journeys through the linkage of previously separate datasets. For commerce, our aim is to help our business understand what is changing and how to manage that change to bring value for our users and our shareholders. For academia, who may be a little late to the 'big data Fresher’s ball' but who can still make a memorable entrance, they are focused on understanding the questions we face in industry, the ethics of what we do and the provenance of it. Academia should be looking for global truths rather than commercially focused turning points.

Academia’s biggest barrier to innovation is the availability of useful data, particularly because most of the innovation in commercial analytics is happening within corporations who use closed, proprietary datasets that involve people. The legalities and practicalities of sharing that data with academia must be an important priority for the next few years. It’s not computing hardware, desks or people that academics need to uncover – it’s the data itself.

The line between commerce and academia will blur when the boundaries between data analysis and data science become more widely accepted across our discipline. Data Analytics provides observational analysis that feed frameworks to understand what is changing. Data Science searches for the 'unknown unknown's' to evidence why things are changing. Whilst commerce can help academia understand the context and the questions we need to ask of our data, academia can help commerce tread the line between science and analytics.

If I notice it’s raining today in Leeds, am I doing science?

The British, more so than their American counterparts, love to talk about the weather. Why? It’s probably because our weather is so changeable and unpredictable. If I notice it’s raining (again) today in Leeds, am I doing science? Science is the 'systematic study of the structure and behaviour of the physical and natural world through observation and experiment.' Do a few, probably quite similar, observations make a systematic study of the structure and behaviour of the weather?

If I invest in one of the excellent IoT weather monitoring devices, and use that device to publish a graph on my website of the rainfall levels in Leeds in 2016, am I doing science? I have some data, limited to a specific geographic location, and I am communicating that data, but does that data really allow me to properly understand, systematically, the behaviour of the weather?

At Bloom, we take various data sources to help us understand what we don’t know about how consumers are behaving and why they exhibit that behaviour. We start with clear hypotheses built from our understanding of how brands should and could engage with their audience. We collect data, transform it and subject it to a barrage of algorithmic investigation to understand what causes changes in the behaviour of our consumer. This gives us lots of 'whats' – the facts about our data – that we then use to build our conclusions: the 'so whats?'.

It is this systematic study of facts generated from the data that moves us away from simply counting and recording things to systematically studying cause and effect.

How important is this pursuit of the unknown unknown's? How important is it that we follow scientific methodology to achieve our goals? 'Advanced Analytics' means different things to different people but we arrived at a consensus during a fascinating discussion group on the subject during the CAO conference. We must always 'shoot for the stars' – to provide us with a 'North Star' which helps guide us on our innovation journey. Advanced Analytics is that which is yet impossible in your organisation, wherever you are on your journey. For some, that is firmly in the art of rigorous scientific prediction around cause and effect. For others it is simply connecting the business to a better way of KPI reporting and dash-boarding.

Basement Analytics happens when the analytics, regardless of how advanced or analytical it is, is not embedded within the organisations DNA.

By developing project teams, focused on solving business problems rather than myopically questioning technology or the next set of buzzwords, you begin a process to take the analyst out of the basement and send them on their way to bringing value to the organisation.

The analytical role is much more than simply crunching numbers – what is essential is that the analytical function teaches the rest of the organisation how to handle data and how data can be used to answer the business questions of today and tomorrow. This 'teacher' role, to help enable the wider business to understand data, is critical. For many businesses, where the core focus is more creative than analytical, the analytical function must skilfully help communicate the benefits of data in a language that can be universally understood within the organisation.

If LinkedIn had existed in 1900, we would have seen a rise in the number of 'Electricity Officers' and their value and their importance to the economy. Over time, the benefits of electricity have become clear and, even though I don’t understand it fully, I trust that there is a team of trained professionals who will help me get the value I need from it and take me on the journey I need to turn oil into light.

Next time I blow a fuse, I’ll make sure I call for the Chief Electricity Officer to come and install a new circuit board. An American Oil Executive would have a hard time selling me a barrel of oil but, once they have turned it into Diesel, I’ll happily pay a premium for it. The Chief Analytics Officers, once the language is cleared up and we understand what our 'Advanced Analytics' is, will be the ones turning the 'data' oil into the diesel of the 21st century economy.

Image Credit: Frank Schulenburg