The rise of 'Punk Analytics'

By the mid-1970s rock music was dominated by prog-rock and long, complex, concept-laden albums. The music was created using multi-track recording and very difficult to replicate live without trucks full of equipment and lots of highly-skilled session musicians.

But things changed after 1976; the ‘anyone can play guitar’ do-it-yourself (DIY) ethic of punk altered everything in rock, stripping it back to its essence and making it simple again.

Furthermore, the punk attitude then extended to fashion, to art, to design. This empowered people. We all thought “I can do that.” Punk made us fearless. Sure, it was stupid sometimes. But it was joyful, and inclusive.

This transition, from exclusive domination by specialists to inclusive accessibility is a trend repeated in many fields. Take this blog, through the medium of the internet I can write and publish without needing the help of editors, typesetters, printers, distributors etc. Anyone with an opinion can share it. It’s punk publishing. We’re all in control of the presses.

So, what have we seen in BI until very recently? A field dominated by mavens, a small number of technical specialists whose role was predicated on arcane skills, and a large number of business people in their thrall. People who, just like rock fans in the early 70s waiting for the next double album to be released, waited months for a data model to be designed and a report coded that would deliver what they needed.

These truly were data priests. Like Rick Wakeman, behind a stack of expensive keyboards, this approach stacked costly technology on technology. Even the nomenclature was defiantly and deliberately obscure, “yeah, we need an EDW fed via ETL from an ODS, and then a fringed MOLAP hypercube to enable drilling with a hyperbolic tree UI…”. And the business people went “wow, it’s really complicated” whist feeling vaguely shut out of the process of creation and remote from the data.

The mavens sought virtuosity and aspired to deliver to a high concept - a set of clear user requirements - that they could deliver the whole of in one ‘great’ work, no matter how long it took. But business decision makers got bored of this, bored of waiting, bored of complexity – it wasn’t helping them – and looked for an alternative way, a do-it-yourself way.

So, by now you’re likely anticipating where I’m going with this train of thought. In the last few years we’ve entered the era of punk-style analytics. With the rise of new technologies that circumvent much of the need for mavens anyone can play data nowadays. This new approach displays characteristics shared with punk:

  • No barriers: You can download data discovery products for free, and get started with nothing to stop you except access to the data you want to explore. There’s no need to wait for someone else to provide you with the means to get started.
  • Mistakes are part of the process: Jamming with data is very often a trigger to finding insights. We get better through trying stuff out. Both in terms of our use of an analytic software product and our familiarity with the data.
  • Fast is good: Think of a Ramones song. Fast and to the point. Fact is that business questions come thick and fast, and being able to riff through data at speed often works best. Many of the questions we want to analyse and answer are transient, and the visualisations and apps are throwaway. Use and discard.
  • Perfection and polish are not the aim: If it’s perfect it has likely been manipulated to adhere to an agenda or to push a conclusion. The idea should not be to create flawless visualisations (think infographics) but a more transparent, less processed route to data that can flex.
  • Engagement with issues of the moment: Punk songs are about the world as it is right now. Data discovery prompts engaged debate too. Questioning orthodoxies about how we measure and evaluate the subject being analysed. It does that because the framework used is a starting point for active exploration, not an endpoint for passive consumption.
  • The collective experience is valuable in itself: No solos thanks! While self-expression and creativity are important they’re secondary to the collaborative act to working and playing together with data to achieve a common result, as this in turn prompts action.

In addition, and despite the marketing messages, not all new analytics has a punk ethos. Some approaches are just building a new wave of mavens – the new visualisation gurus, often yesterday’s Excel gurus, still revelling in their virtuosity.

Sitting alone in their cubes, these specialists create beautifully crafted visuals – which are just so – and then distribute these as perfect dashboards to people with Reader software to look at and be impressed by. Not punk: perfectly polished, self-publicising, one-to-many maven created artefacts.

The approach is exclusive and not collective in its approach, it’s not about engaging as many people in playing with data as possible. What’s the aim of creating ‘just so’ visualisations? Who benefits?

The real work happens when more people can explore data and learn through play together.

When they pick up their data and play. Fast and loud and loose.

James Richardson is a Business Analytics Strategist at Qlik and a former Gartner analyst.