Big data – little impact?

Posted by: in Analytics, Communication Strategy, Corporate and Strategy, Integrated Communications, Marketing Insights, Pre-Commerce, Public Relations Practice, Social Media Insights & Trends, Thought Leadership on November 4, 2012

Debating the future of digital marketing at the I-COM Summit in Rome, a who’s who in online media measurement: both challenges and opportunities abound

 

The big themes: impact, influence, insight
If all the contributions at this year’s I-COM, the International Conference on Online Media Measurement held in Rome from 14th to 18th October – all the presentations, all the panel discussions, all the face-to-face conversations in the breakout sessions, over coffee or lunch or during the rich social program in the evenings – were condensed into one big word cloud: what would it look like? Perhaps a little bit like a word frequency visualization of I-COM’s own newsfeed:


Listening to all the experts and thought leaders in their respective fields, it is evident that the overarching themes are all about impact, influence and insight, mapping multi-channel, multi-screen, multi-session customer journeys through a complex and increasingly fragmented media landscape, measured in real time with big data collection and processing tools; data scientists standardizing and normalizing data and developing metrics and multiple data views to power and populate client dashboards, displaying actionable results to ensure the right message reaches the right targets at the right time through the right channels, to improve and optimize business results.
As part of an online landscape that is constantly evolving, companies are operating in a highly dynamic and disruptive environment – and organizations that are nimble and flexible enough to constantly adapt to these changes will have a critical competitive advantage.

Make clients’ lives easier, not harder
Big data and dynamic dashboards were among the most discussed themes, and as a panel and follow-up breakout session on Applying Consumer & Audience Analytics to Business Decisions made very clear – whilst everybody in the industry acknowledges the huge opportunities, there is also a good deal of caution: dashboards are not the answer to all questions, especially when they are powered by (big data) analytics that are not (yet) connected (enough) to business decisions.
The message was loud and clear: too often, clients are too busy to engage with the data on offer; there is a tendency to provide too much data and dashboards with too many dials and charts. The challenge is to avoid information overload and to cut right to the chase of actionable insight, to help clients reach the right people at the right time with the right message through the right channels.
Also, if it doesn’t relate to business results, it has little purpose: for social media to be commercially relevant, clients need data that proves the revenue impact of social media – and that means tying the measurement of social media activity to business outcomes. At the same time, stakeholders realize that the business impact of social media does not yet lie in significant revenue generation, but rather in better understanding customer experiences and customer journeys, and in learning from richer data and better insight to generate more conducive environments for successful engagement.

The ‘value of social’ – start with sound measurement
Benchmarking: What is the Value & Currency for Social – that was the title of the panel discussionin which I participated, together with an illustrious mix of client and agency representatives: Amanda Richards from Unilever, Brad Smallwood from Facebook, John Lovett from the Digital Analytics Association, Marshall Manson from Edelman, Bryan Tookey from Brandwatch and last but not least, Layla Pavone from Isobar Communication as a moderator.

Talking about Social Media Measurement Standards

Nobody expected a final, exhaustive answer, and none was given. However, there was agreement that for measurement and evaluation to be worth our paying clients’ while, measurement needed to get better: more consistent, more rigorous, more tangible, more transparent, more comparable, more accurate, more replicable – in short, more professional. Only then will it be seen as a mission-critical business activity by clients. Only then will measurement become an integral part of any marketing communications program, any PR campaign, with the measurement and evaluation of social media a logical extension of best practice, based on agreed and accepted industry standards.

Some challenges, many opportunities
Yes there was excitement over the boundless opportunities to track and measure online behavior, to generate ever-expanding data sets, to slice and dice and analyze the data with ever more complex and sophisticated models. However, healthy doses of skepticism and the commitment to keep one’s feet on the ground weren’t in short supply either. Sound advice for clients included:

  • Don’t fret over real time, just make sure you hit the right time, and use the insight to reach the right audience in the right place with the right message
  • Accompany consumers on their journey long before they reach their destination and hit the ‘buy’ button; the decision-making process starts – and often ends – long before someone sets foot in your store or lands on your home page; get better at e-commerce by engaging pre-commerce
  • Content is king, source is secondary
  • Speak the ten languages that cover more than 80% of the (online) world’s population
  • Manage social media in four steps:
    • Map out the social media landscape: where do people go? What do they search for? How do they search? Which sites do they visit? How do they share information? What content is available?
    • Identify key influencers (contextually)
    • Identify influencers’ influencers
    • Engage, provide the right information through the right channels at the right time to maximize impact
  • Start listening; be curious; find and engage your top influencers; be visual; be nimble, adapt; create good content; distribute it where and when it will make an impact

Hybrid solutions optimize the return from machines and humans

Barriers to data- and analytics-driven models remain, and they stem from unreliable data, from inconsistent approaches to measurement, and from impenetrable methodologies when clients want transparency, not black boxes. Bad data anecdotes stick in the mind: Bettina Sherick, SVP Digital Strategic Marketing at 20th Century Fox International, amused the audience with her measurement experience around the release of the movie The Watch, where social media listening not only unearthed fan opinions on Ben Stiller and Vince Vaughan, but also plenty of comment on various Swiss luxury timepieces. And that was just finding themes and groups of words, not the notoriously difficult rating of sentiment.

The solution is to have as much automation as possible and as much human intervention as necessary for robust, reliable results. The best research designs will balance the input from software engineering and data sciences both with a human, social science-led approach, as well as with business practice: hybrid solutions where algorithms are constantly refined by machine-human learning based on actual business outcomes, to generate intelligent and intelligible results: actionable findings that lead to better business decisions, and ultimately to better business performance.

Digital is more than just ones and zeros
The social science angle is critical: you can’t get the right answer if you don’t know the right question to ask, and social media data is first and foremost linguistic data: documents, sentences, words. Scott Thomson, the Global Head of Analytics at UK agency Naked Communication argued succinctly that data should be approached with a hypothesis, with scientific rigor: collecting data is simple, investigative curious thinking is not. Therefore we must not build data warehousing solutions with complex business intelligence tools without linking them to actual business processes and/or having the skills to draw the right insights.

Less hype, more realism: a data cube is not a crystal ball
Perhaps the most important take-away from my two days in Rome was the reassurance that the industry as a whole understands both the opportunities, and the limitations: a data cube is not a crystal ball, and human behavior cannot be predicted from empirical data. That works for physics, which is governed by precise natural laws. Human psychology is impulsive, inconsistent and often random, and big data won’t change that. However, access to more data on human (online) behavior will help us understand better how people behave online, who and what influences them, and how that leads to decision-making. By sharing that understanding with our clients, we help them make better, more informed decisions – a critical competitive advantage now and even more so in future.
So no – I don’t think big data has little impact. Quite the opposite: it will transform the way we understand and do business on all levels. And it will be most successful where it is approached with humility, rather than hype. At WCG, we fully embrace the huge opportunities that big data offers for smart analytics, to better understand and predict the behavior of target customers, communities and stakeholders, to optimize content delivery: to the right people, at the right time, through the right channels.

By: Thomas Stoeckle

Global analytics lead at WCG in London and passionate about doing good research.

Find me on: Twitter
Pre-Commerce Check out W2O Group President Bob Pearson's new book, Pre-Commerce, in which he shares ideas for leaders to engage directly with customers to shape their brand and marketplace success. Now available for order on Amazon.com! http://amzn.to/bAmvFN. Join the conversation #precommerce.

One Response

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Sam Knowles said

    Great piece, Thomas. Beautifully written and thought-provoking. I’d almost say you saved me a two-day trip to Rome ….

Some HTML is OK

(required)

(required, but never shared)

or, reply to this post via trackback.