There are lots of great posts and articles on Big Data and and the impact of social media. Many discuss the value to be had from social media whilst others provide commentary on how difficult it can be to get true social business intelligence. But how do you actually wade through masses and masses of conversations to produce insight that is actionable and moves the social strategy further towards business goal attainment?

Dion Hinchcliffe provides much of the thought leadership on Big Data. As he observed recently, social data is like a tidal wave which is only set to get bigger in coming years. But as he also posed – is all observable information valuable? Clay Shirky says that ‘The problem is Not Information Overload. It’s Filter Failure‘. And nowhere is this more of a problem than in social media monitoring

So we agree. Therefore,  how do you filter what’s relevant and where should you focus?

How to filter observable data – 5 tips

It’s tempting to try to analyse all observable data but you would never get to the point of making any conclusions. We thought we would share five tips on how to filter observable data to get to the ‘nuggets of gold in the hill full of dirt’ that are the social media data streams.

We monitor the social data streams for a number of clients and these tips come from our practical experience of monitoring masses of social media data. For one client, which is a B2B business, our monitoring for Asia Pacific returns about 50,000 conversations a month. This includes monitoring social media for a number of competitors, though we only do that for certain keywords and phrases.

Here are 5 ways we turn those 50,000 conversations into actionable insights:

1. Only focus on key terms

Obviously, you want to monitor social media for brand mentions, product and service mentions and key competitors. We also monitor for some key words and phrases which are important to a client’s positioning and messaging, and also from time to time important to their special events. And we monitor for specific words or phrases that might signal a product or service issue. What is key is to focus on relevant terms and (extra tip) update them periodically. Branding never stands still – nor should social media monitoring.

2. Text Analytics are your friend

Word Clouds and other text analytics are your friends in social media monitoring.

They provide clues on the terms and words that are used in conjunction with discussions on your brand, products and services as well as those of your competitors. They will help you understand if you should think about keyword and SEO changes. And don’t forget the use of text analytics around participants in the social channels.

If your social media monitoring tool provides word clouds based on the bio’s of a set of Twitter handles in conversations (like Sysomos does for example) it can provide a lot of insight into the kinds of people who are influential in the conversations in which you are interested. You can see that there is a lot of “soft” work to be done here in analyzing, sifting, filtering and relating information to the objectives of the search. This is much of the real work of monitoring, which is not automated, and where experienced eyes pay business dividends.

3. Influencers are important

We understand that you shouldn’t focus on influencers alone but lets face it – when it comes to engagement you want to get the biggest amplification for your effort. In fact Michael Wu‘s new book makes this point over and over again in relation to the power of SuperFans.

Starting with key influencers in each primary social network that is relevant to your business is critical to leveraging your amplification. Your social media monitoring efforts should allow you to easily identify key influencers in order that you can can construct and monitor an effective outreach program. You might also want to do some “background checking” on those key influencers beyond your monitoring software e.g. Klout, Kred to further check their reach and influence and topics.

4. Share of Voice

We don’t focus a lot on just quantitative metrics in our social media monitoring. The pure number of followers or mentions or tweets or whatever is interesting but not terribly germane to your social business strategy. But Share of Voice can be.

Comparing levels of mentions and conversations by key term can tell a business a lot about its messaging, its competitive position, its effectiveness in cutting through the mass of activity. And particularly important is to look to see how Share of Voice is different by network or channel to see if you  have a well balanced engagement. For example we’ve found some cloud vendors dominate Twitter and are almost non-existent in other channels, and vice versa.

5. Sentiment

We see the discussion threads on the topic of not running a social business strategy on the basis of ‘they love me, they love me not’. But that’s a shallow view of sentiment.

The vast majority of social data is banal and has little impact on how the brand is perceived. In our experience 80% (or more) of the mentions or conversations will be neutral. And interestingly, they will generally be tagged correctly by the natural language processor. But the negative and positive mentions are rarely 100% correct and need to be carefully scrutinised. For example, I saw a post today on ‘Reliable Disaster Recovery’  – most engines would classify that as negative but it’s clearly not.

Positive mentions are important to understand in terms of what features, behaviours and experiences delight people who interact with your brand – so that you can repeat them; emphasize them and include them in your positioning. Though repeating them is most important of all because it will help you turn customers into advocates – and if they are influencers as well then you have double the result (and even more if SuperFans!).

But negative mentions are also key, although they can be harder to face. Obviously they can alert you to immediate issues that need response and resolution. But remember that the person behind the negative mention cares enough in the first instance to comment. It’s an opportunity to engage; to discover improvement opportunities and to turn a detractor into a fan. Don’t miss it. And don’t miss the opportunity to seriously think about how you may be able to get your positive influencers, or community, to take up the issues on your behalf and resolve them. That has to done done delicately but the rewards are there, and often the person making negative comments can become one of your greatest fans.

Whilst there is lot more than the above to being effective in social media monitoring, these 5 tips are a great place to start any social business intelligence program.

Where do you focus?

What tips do you have for effective social business intelligence ? Please comment below.

Will B (follow on Twitter at @wbosma and @iGo2)