Ask 100 marketers what to do with social media and you’ll get 100 different answers. Ask them about how effective it is as a marketing tool and you’ll get a spectrum of responses ranging from boundless enthusiasm to outright derision.
The one thing that most will probably have in common, however, is that they think of social media primarily as an outbound marketing tool – that is, a medium for sending out information, whether to one person or to many. Those who use it for inbound communications are likely to say they use it for customer service – responding to questions or problems.
Relatively few are likely to say they use it in a way that’s quantitative or analytical. Yet, with so much information being wilfully placed in the public domain by millions of consumers, brands that aren’t doing this are surely missing out on vital knowledge.
One example is provided by telecoms company O2, which has set out to build a ‘brand graph’ based on its Twitter users. The project, run in partnership with agency Face and social data company DataSift, has sought to assess O2’s social media audience according to demographics, behaviours, interests and the ways different users influence each other.
Among the things O2 has discovered is a significant and unexpected community of social media users interested in fashion. The brand has also found that three-quarters of those who tweet O2 don’t follow its Twitter account, and that its own activity on Twitter has relatively little impact on the conversations people are having about it.
Brands are also beginning to use social media to get a quantitative idea of how positive or negative social media sentiment is, using tools such as SparkScore, or by combining analysis of YouGov’s SoMA social media data with its BrandIndex scores. Marketing Week recently gave its own demonstration in a cover feature on the Olympic sponsors’ use of social media.
Technology that analyses positive and negative sentiment from social media posts automatically has been notoriously unable to demonstrate reliability that’s much better than flipping a coin. The companies behind SparkScore – Satmetrix and Metavana – claim to attain 90 per cent accuracy, though the details of how remain a secret.
As with other marketing uses, social media remains unproven as a research tool. But what is undeniable is that there is a huge volume of consumer data surging onto the internet every day, and there’s nothing to stop brands from sifting it for gold.
Have your say now on who should be the Data Professional of the Year.
The ballot closes on Tuesday 9 October.