Big Data? What about Small Data?


By Matthew Robinson

Everyone’s falling over themselves to show an interest in Big Data these days. Mind you, that’s despite Big Data dropping, very rapidly, from the “peak of inflated expectations” (on Gartner’s Hype Cycle for Emerging Technologies) and descending speedily now toward the “trough of disillusionment”.

Gartner says this swift decline is simply due to the existing “consistency” in the way we approach Big Data, and because most new advances are incremental rather than revolutionary.

However, it occurs to me this narrative overlooks a further explanation for Big Data’s super fast passage along the Hype Cycle. Which is that no-one has ever really understood what Big Data is, and what it promises. And further to that, it seems Big Data may only have succeeded in distracting the marketing industry from paying proper attention to the small data that supports straightforward, yet highly valuable, actionable analytics. Which is a pity.

It’s also significantly counter-intuitive, being as research carried out by Econsultancy and Adobe found that the three digital-related areas considered top priority for organisations in 2015 were 1) Targeting and personalisation (30%), 2) Content optimisation (29%) and 3) Social media engagement (27%) – all of which can be advanced by small data analytics.

So what then is big data?

Well, Big Data is a broad term that serves both to describe the sheer scale of business data available to capture now, as well as the analytics opportunity to interpret and extract value from said data. In short, Big Data is analytics at very significant scale. Which, in fairness, is something you can probably infer quite easily from the term “Big Data”.

Just to be very clear though, what Big Data isn’t is a meaningful label to affix to this or that vendor solution. Those that brand tools on the basis of their supposed Big Data capability are really only bigging up their sales pitch, not your analytics potential.

Still, where I really think the confusion and lack of understanding creeps in is on account of Big Data being the enabler of such a vast array of business plans and opportunities. In particular, of improvements to process, of innovations to product design, and of driving through company / staff efficiencies. In essence, of lending support to the very established practice of business analysis.

And that’s confusing if and when the possibilities and benefits of Big Data get surfaced in the far more confined context of actionable (web) analytics – to which the wider practice of Big Data is much more selectively relevant.

All of which results in a deep seated confusion as to what Big Data is, who it will benefit and how. Certainly amongst those who work in digital marketing, and who engage with the analytics community.

The good news is though, Big Data need not darken your door if *all* you’re wanting to achieve is good, actionable analytics, scaled to improve *just* digital content and comms. For this you don’t need Big Data, you need only small data, a clear set of objectives, effectiveness KPIs and a commitment to optimisation through testing.

Of course, it could be argued that part of the problem with Big Data is that the very name invites you to give immediate consideration to… data – in the interests of gaining some valuable insights. And yet, it’s widely understood in analytics that you almost never get to useful insight when you start with data. For the same reason that you won’t get the most out of FAQs if you go immediately to the answers without first looking at the questions.

By all means get stuck into Big Data projects if it’s understood from the outset they are the key to unlocking product development or customer relationships or something else clearly defined… But don’t reach for it as an insights lever simply because you’re now aware that it’s out there, that it’s a thing, and that it therefore needs adopting some how.

Instead, take care to ensure you’re exploiting all of the opportunities to analyse online performance and to improve afforded by small data. Like, for instance, in respect of:

Channel mix

Are you over reliant on paid media interrupting people to drive traffic? Is it easy at all times for people to find your brand if they themselves are actually looking for that which you have to offer? Study your channel drivers, but also pay attention to the (organic) keyword searches that push the most users to which particular top pages. This identification of search engine keywords that drive traffic (and on-site conversion) is then a key enabler of SEO and improved content visibility. 

Landing pages

Are all of the entry points to your online properties fostering a positive first impression of your brand? Which pages are the highest bouncing? Contemplate and hypothesise why that might be and A/B test alternative variants.

Content pillars

Is your content relevant to the audience you’re wanting to reach? Does it work hard in support of wider online objectives? Have your page engagement metrics help refine your content mix and/or their content format, and surface more prominently the pages, tools etc that work the hardest.

In short, show commitment to small data insights - because it’s proven they can demonstrate the effectiveness of digital and/or reveal what, where and how to improve online marketing content. And show interest in Big Data possibilities – on condition that precise thought has been given to what it is they’re expected to deliver.

If we adhere to that, we won’t go far wrong! And I won’t have to needlessly start calling myself a Data Scientist.

Photo: Fox Photos/Getty Images