Research claims that 25% of tweets are not worth reading. So what?


English: Microphone

Image via Wikipedia

According to research from a team at Carnegie Mellon University, the Massachusetts Institute of Technology and the Georgia Institute of Technology, we think that 25% of tweets are not worth reading. The study found that, when asked to rate tweets by people they follow, only 36% of tweets were marked favourably, 25% were marked less favourably and the balance (39%) received no strong feeling either way. Press coverage of this study has invariably interpreted this to mean that up to a quarter of what we say on Twitter is a waste of time (see the Daily Mail and the Daily Telegraph coverage of the research).

The research itself asked users to self-nominate themselves to take part and, in exchange for having their own tweets rated, we asked to rate samples of tweets from people they follow. As with much academic research, this does take them out of their normal context when using Twitter but the results are interesting and informative. Maybe not for the interpretation that is taken by some of those reporting on it, but for what it tells us about how we use Twitter. Or perhaps how it reinforces what we should already know.

People are not interested in everything that people say on Twitter. They are not even interested in everything that the people they choose to follow say. There should be nothing surprising or controversial about this. It is fairly normal in all our social interactions that we are more interested in some things and less interested in others. I’m mainly surprised that we are uninterested in only 25% of things that people we follow say on Twitter.

Twitter is a classic social network. People who use it by following people (rather than by following hashtags or search strings) make a choice about who to follow based on who they are, what they say in their biography and perhaps some of their tweets at the time that we choose to follow them. I am unlikely to share everything in common with them - I may be interested in their tweets about BBC Question Time on a Thursday, for example, but less interested in their Tweets about the Super Bowl. I am unlikely to find you interesting all the time. And that’s nothing personal. And nothing unusual.

So as a reader I am unlikely to find everything that anybody says on Twitter interesting - I mentally filter out what I want to read and what I don’t want to read. If I really don’t want to read things on a certain topic, I can always filter it out with Tweetdeck or the like.

Just as readers are not necessarily interested in reading everything, those who write tweets are not necessarily writing them to be read. There is a clear disconnect between the person writing the tweets and the people reading them. The writer is not (in most cases) thinking about who will be reading it and why. They are just saying something. Saying it because they want to. That in itself is motivation and on the rare occasion that a tweet will be retweeted or responded to they will get further gratification.

So we are not interested in everything even our closest friends say (probably true in real life and on Twitter). And people are often writing on Twitter for the act of writing something and not necessrily composing it for specific audience or a specific reaction. Given that most people follow a collection of people with many different interests, some of whom they know and some of whom they don’t, it surprises me that only 1 in 4 tweets that we see are not of interest to us. This study certainly doesn’t show that those Tweets are a waste of time.

The basics of social media monitoring


social-media-monitoring-toolsThis is our first post from the Social Media Monitoring - 2010 review series. In it we’ll cover the basics of social media monitoring.


Social media tools make it possible for people to have conversations online. The uptake in conversations, comments and reviews has been explosive and the importance of these conversations is growing by the day. Among many other things, people are discussing brands, describing their purchase intentions and asking for assistance in making buying decisions or product support.

The opportunity for organisations is clear. They can now listen-in on the conversations of their customers, potential customers and other stakeholders in a way that was previously impossible. Through social media monitoring it is possible to gain insights from the conversations people are having online every day and to make improvements to products, customer service and marketing as a result.

Real-world ethnography has been around for a while – the process of analysing the context in which people act, usually researched by observing subjects in their natural habitat. It can teach us a lot about behaviour and influencing factors, however it is expensive and subject to The Observer’s Paradox (see also Schrödinger’s Cat).

Social media monitoring brings observational research to a mass audience. By tracking what is said in forums, on Twitter and in other social networks, brands can gain customer insight. But beyond getting geeky researcher’s excited, it can also offer very practical benefits to organisations. Customer service teams can listen out for customer issues online and then and resolve them. Competitor Intelligence departments can find out what customers are saying about competitors’ products. PR Managers can get early warning of pending PR disasters before they hit main-stream media and most of all, by listening first, companies can be better prepared to join online conversations and become social.

Social media monitoring clearly has tangible business benefits and as a result it’s a hot topic. Furthermore, the power and importance of what people are writing online is increasing. The reach an influential blogger can have is extraordinary. And according to Neilsen, consumer recommendations have now become the most powerful form of advertising (78% of people trust customer reviews). As a result companies need to monitor word-of-mouth more than ever.

Using search engines to monitor online conversations

Many online conversations can be accessed with ease and for free using Google or other search engines. Simply using your brand name as a search term, or using keywords that are associated with your brand (eg, for Starbucks you could search for “Starbucks” or “whole bean coffee”) you can find conversations that are related to your brand. Taking that process a step further you can set up Google Alerts so that you get an email when someone mentions your keywords.

However, if you search in this way you’ll probably end up with hundreds of thousands of returned results and a limited number of ways to analyse the data further. You will also get a mixture of professional and user generated content. It is possible to use some free buzz tracking tools to focus on certain areas. For example Omgili and Board Tracker are great ways to search forums. But until Google enters the social media monitoring market, the best way for enterprises to track social media is by using a paid-for tool.

The benefits of using social media monitoring tools

Social media monitoring tools deal with the two problems of searching and analysing the online conversation. The tools use similar web crawling technology to search engines in the way that they read online conversations. However, unlike search engines, the tools clean, de-duplicate and categorise the conversations and then store them in a database.  As our report and future posts will show, some tools do these things better than others.

Social media monitoring tools also allow you to enter search terms into the database so that you can customise the way you view the results. The tools count the conversations that contain your search terms and provide you with the ability to display this information in graphs and charts.  Most tools also allow you to divide by location or media type (eg, Twitter or blogs) and at the cutting edge, some social media monitoring tools provide workflow management process that can help you disseminate conversations within your organisation, others are starting to combine buzz tracking with CRM in a bid to create single-customer-view Social CRM. And there are some tools that allow you to respond to conversations across the web from a single dashboard.

One key feature that marketeers have been most keen on is sentiment analysis.

What is sentiment?

Sentiment is a thought, view, or attitude that is often based more on emotion than reason. In the context of social media monitoring, it is the concept of deciding whether a specific online conversation is positive or negative. This is really useful in helping you determine the themes and topics that are driving both good and bad conversations about your brand.

Sentiment can also allow you to track the overall impact of marketing campaigns or news about your brand. We suspect the main reason people have latched onto sentiment is because it gives the impression that the plethora of web conversations can be summarised in a single number. Businesses love to track numbers and sentiment is often the KPI of choice for social media.

This is dangerous. Sentiment is more nuanced than a single number and using an automated tool to assess how people feel puts too much faith in the today’s software. We don’t believe that the tools on the market have nailed sentiment analysis yet. The tools can be extremely valuable, but it is important to understand their limitations as it is to understand their capabilities.

One piece of advice - it’s not about the bike

The most important thing to bear in mind when choosing a social media monitoring tool is that ongoing human interaction and interpretation are essential to get real value. If there is one mistake that companies are making it’s that they buy into a dashboard expecting insights on a plate. Months later they look back and wonder why the dashboard hasn’t changed their business.

Buzz tracking opens up opportunities for insight, but it is worthless without sufficient people resource and internal processes to act on information.

I am biased. My background is research (FreshMinds, our sister firm, has been twice named UK Research Agency of the Year by the Market Research Society) and we’re not selling a tool. Rather we help companies select the right tool and help them get value out of it on an ongoing basis. But I think you’ll find most of the software vendors will concur that their happiest clients are the ones who have properly resourced the listening effort and invested sufficient time in interpretation, dissemination and action. After all, what’s the point in listening if you never act on what you hear?

The next blog post in our series will be about setting up each of the leading tools to get the most out of them.

Social media monitoring breakfast seminar

We’re also holding a free social media monitoring breakfast seminar on 15th April in London, where we’ll be presenting the findings of our report, as well as giving practical tips and advice about social media monitoring and the best way to analyse results. You can register for the event by clicking on the button below:

Register for Social media monitoring in London, United Kingdom  on Eventbrite

Read the other posts from our social media monitoring review 2010.