co-working space


How to analyse social media data

Essencient Co-Founder and CEO Rob Lancashire tells us how to analyse social media data

Social media data holds a treasure trove of insights into a brand’s image, but they’re often drowned out by “noisy” social mentions. To separate the two, we talked to Rob Lancashire of Essencient – a tech start-up that helps companies identify their meaningful brand mentions. Here he shares how to extract every drop of value from social media.

1. Identify meaningful mentions

Essencient defines meaningful social mentions – which account for just 5 per cent of all brand mentions – as ones that contain “intent to do something, sentiment for or against the brand, or a request for help or advice”. In other words, something of use for gauging the brand’s image. “Let’s say you’re BMW – you want to hear when someone says: ‘Just saw the new BMW on the road. I want one.’” You don’t want to hear a heap of irrelevant hashtags: “Car show! #mercedes #bmw #audi.”

“There’s no substitute for reading them all,” says Rob, “but it’s an impossible task without making a huge investment in ‘readers’.” Instead, Essencient’s software analyses the unstructured text in social conversations to establish the essence of each post. It then tags meaningful mentions to create a more manageable feed of high-quality content for social media teams.

“We calculate that each employee can evaluate about 1,400 tweets a day, which is roughly the number of meaningful mentions found in 28,000 raw mentions.” With Essencient’s technology, 140,000 mentions a day can be handled by a team of five.

2. Decide how to respond

“Engagement is a tricky matter – part art, part science,” Rob explains. Users simultaneously expect posts to be taken seriously and object to intrusions into their privacy, so the brand must be sensitive when deciding when to intervene. “When a mention implies loyalty, a direct approach is often well received. But when you’re selling something, it’s far less welcome, and should be integrated into a campaign targeted at a more broadly defined audience.”


The appliance maker Miele responded to a complaint posted by a widely-read blogger by showing up with a lorry full of new appliances. As a result, thousands of critics rapidly became thousands of fans. Miele understood the importance of “knowing” the customer before responding – in this case, the blogger’s reach meant a direct response had a huge impact on its brand image.

3. Target competitors

In areas where a firm has the edge over its peers, competitors’ critics could become future customers. Rob recommends applying the same policy to competitors’ social media as you do to your own. He also suggests taking an “arms-length” approach to engaging competitors’ detractors: “A direct response to a social mention saying ‘We hear you’re unhappy with your current supplier. How about talking to us?’ can be very negatively received. But an awareness campaign can reap rich rewards.”

The insights to be gained from non-noisy social conversations are invaluable, and should inform any operational areas that reflect “the voice of the customer”. That voice, when expressed in a social mention, is honest and informative, since the comments are unsolicited (especially the 97% of comments that aren’t tagged with a label like @adidas or #JustDoIt). Figuring out how to analyse social media data can unlock a whole new approach to brand image – so start by switching off the noise.

Rob Lancashire co-founded the Essencient group, and is CEO of the UK business. Essencient operates from a Regus co-working space based in London Wall, Liverpool Street, London