Needles, haystacks: 5 tips for working with social data

Needles, haystacks: 5 tips for working with social data

Let’s start with a challenge.

Most Market Research is obsessed with demographics because we needed it to find and understand our target market. I get this and understand the pressure to apply the same framework to social insight.

But you really don’t need to.

In a world of cross-generational, cross-geographic influence I wonder if we should be this committed to traditional ideas of demography.

Increasingly, it’s aspects of our lifestyle, values, and interests that define the ‘groups’ we form – more than the demographic boxes we are put in. What Seth Godin and others refer to as psychographics

Social data already tells you ‘where’ your audiences are and enables you to find them. It also tells you a bit about ‘who’ they are, albeit from a more lifestyle/interest-led means of classification.

You don’t need demographics if you already know where the party is and who’s going. You just need to think about the research process slightly differently. Here are some tips to get you started.

 

1. Keep an open mind and be prepared to travel to some strange places

Doing a project on banking? Don’t be surprised to see people talking in pet forums. Looking at how a certain retailer is perceived online? Well, people also talk about this stuff in TV/audio forums. It can be quite eye-opening to see the breadth of places people are talking about you or what you’re interested.

So, don’t limit your search to the social channels you think might be relevant – keep it open and explore what’s out there. People make friends on specific forums then go on to talk about all sorts of things. Keep your sources open but fine tune your search, by…

 

2. Focusing: match your business objectives to how people really talk

It’s important to invest time with your agency at the beginning of a social insight project so that you can align your research objectives to the way that people actually talk online.

It’s not like traditional brand trackers where you set the questions and give clear options to respond to. The beauty of it is that it’s naturally occurring conversation, about what people chose to talk about without being asked. You might want to know if people think your brand is ‘on the rise’, but that’s just not how people talk about brands in real life. Invest time upfront thinking broadly and learning their language. This will help you focus on the right words and terms – which will then help you search for the right content and potentially mirror later when you chose to talk to them.

 

3. Targeting: location, location, location

Decent demographic data is hard to find in social, which can make focusing on specific audience types challenging. You can’t simply sample your way to the audience.

One way of tackling this is to look for the social spaces which attract specific audiences. These tend to emerge around particular life-stage or areas of interest – rather than age or location. The demographic profile of many of these will skew based on the subject – which acts a short-cut to your audience. Think Mumsnet and Netmums, Babycentre, Pistonhead (for cars, male skew), StudentRooms, Badger and Blade (about shaving), or on a more sombre note the excellent cancer community from Macmillan.

 

4. Targeting: the secret door

You get social data by using keywords which tell your listening tool what you’re interested in. But what if the things you’re interested in are too conceptual or too mundane to come up in keyword searches? When your keywords aren’t working you need to find a less obvious ‘secret door’ in.

One way of doing this is to use proxy terms – things that only they would talk about.

There’s the obvious stuff like, if you want to find grandparents, add “my granddaughter” or “my grandson” to your search terms.

Then the more conceptual, for example, how do women over 45 think about beauty. Here you need to be a bit creative and don’t rely on keywords alone. You need to think about what products only they might use (to help you find relevant review data) or what forum threads they might be engaging with (here’s a starter for 10 from Mumsnet “Season of Mists and Mellow Crepiness….

It helps if you’ve spent some time immersing yourself in how people naturally talk about this stuff (we’ll share some tips on this next time).

 

5. Ignore the nonsense: get the chisel out

Working meaningfully with social data can take a bit of effort up-front. Filtering out ‘the noise’ for example. The amount of giveaways, competitions, adverts adds meaningless clutter to the content you get and makes it harder to get to the core of what people think and feel – the things you care about.

We’ve learnt that no matter how clever you think your search is, people will always use your terms in unexpected and irrelevant ways.

It’s there, you just need to keep digging, testing your data collection query, refining it, testing it again and getting creative in what you look for.

 

Thanks for reading. If any of this strikes a cord, do get in touch, we’d love to chat.

 

About L+LR

We are the Social Insights agency.

We make social data simple. Simple for you to know what’s meaningful, to find new insights, to see opportunities and to spark inspiration.

We’ll help you use social to find and understand the people that matter to you. So you can create the right experiences, engage new audiences and appeal to people in the best way.

People are busy, we’ll help you become something they want to be part of.

 

This Social life

Our newsletter shares a little of what we've learnt, sound good? Sign up.

Thank you for subscribing.

Something went wrong.