Social Media monitoring is a way to understand what’s being said about your brand on the internet. Monitoring of social media platforms encompasses the direct communication between you and your customer, such as messages from customers on your Facebook page, as well as indirect interactions such as customer reviews of your product in online forums.
Monitoring your customers’ view of your brand can yield great market intelligence, as well as help you manage any customer negativity and help mitigate damage by dealing efficiently with complaints or bad experiences with your brand before these escalate. Social monitoring can also yield other benefits such as driving innovation and provide a source of market research.
A number of tools are available to help companies with their multilanguage social media monitoring across the various social media platforms. These tools crawl the internet continuously to index all brand mentions across forums, blogs, news sites, review sites, and social networks such as Facebook, Instagram and Sina Weibo.
This allows you to monitor the conversation at both the macro level, identifying trends and topics of conversation, as well as at the micro level by indexing comments from across the web into a digest or list format for easier monitoring via one account.
You can also employ sentiment analysis, which uses artificial intelligence to process language and determine whether your brand is being talked about in a positive or negative way.
There are some barriers to monitoring: sites such as LinkedIn don’t allow monitoring tools to cover all its content, and sentiment analysis is difficult to perfect. If you’re working across multiple languages, there will inevitably be linguistic barriers to social monitoring.
If you’re working across local social platforms in unique geographies, such as Japan’s Mixi or South Korea’s Cyworld, you may find your monitoring tools don’t extend their reach across these platforms or their languages. It may be the case that you need to adapt your monitoring activities and handle each territory uniquely rather than take a global approach.
Some tools have multilingual monitoring capabilities. HootSuite is able to detect 14 languages including Arabic, Catalan, French, German, Japanese, Korean, Russian, Spanish, and Turkish. They also cover a range of platforms including Mixi in Japan and China’s Sina Weibo, Renren and Tencent Weibo.
Whilst this could enable global monitoring you’ll almost certainly have to face a unique competitive environment and brand perception situation in each market you operate in, so you may be best advised to conduct monitoring specifically for each territory in which you operate in addition to tracking the bigger global picture.
If you’re operating on local social platforms, some are better supplied with analysis tools than others. Japan’s Mixi offers owners of Mixi pages the ability to track the number page views, unique users, followers, comments, and “likes”. Sina Weibo offers its own monitoring tool similar to Klout, and allows you to identify your followers by their gender and location amongst other characteristics.
International social monitoring
Don’t compare apples and oranges! If your brand is engaged across multiple international social platforms, you may find it difficult to monitor consistently when norms of behaviour aren’t the same across all platforms and geographies. There’s some research which suggests British and American audiences are more likely to advocate passionately on social media, using terms such as ‘love’, compared to other nationalities.
On the other hand, Chinese consumers are thought to take information delivered via social media more credibly than from other sources. So whilst a passionate expression of advocacy may be rarer in some markets, you might consider it to be more sincere or more valuable when it does occur.
You’ll need to take cultural aspects such as these on board when you’re trying to conduct sentiment analysis internationally.
To some extent, different social media platforms offer broadly similar kinds of tools. Many platforms offer something comparable to Facebook’s ‘like’ button, for example.
But there are some subtle differences which mean it isn’t advisable to compare measurements across different social platforms because they don’t quite assess the same thing.
For example, on Russia’s popular VK social networking site, ‘likes’ for content such as comments and posts don’t get automatically pushed to the user’s wall, but are saved in a private ‘favourites’ section instead. The user then has to press a second ‘share with friends’ button to share an item on their wall.
This two-stage sharing process means visible advocacy is just that bit harder to achieve from a VK audiences. You might give this more weight in any analysis you perform in this channel than in a site like Facebook where advocacy is a one-click process.
Sina Weibo, China’s preferred microblogging site, is often talked of in the same breath as Twitter however the platforms are actually quite different and so is the kind of content shared on the two platforms. Chinese users of Sina Weibo tend to use the platform to share jokes, images and videos and a significantly large percentage of posts are reposts – or retweets if you wish to compare to Twitter.
This ‘repetition’ of content affects the trends formed and there may be a snowball effect which means fewer topics trend but trending topics often achieve extremely high visibility. By contrast on Twitter trending topics have more to do with current events and the effect of retweets is not as significant, with users less likely to include links and content such as images or video in what they retweet.
Retweeting of trending topics is many times more popular on Sina Weibo than on Twitter, but participants tend to contribute to fewer topics. It’s important to adapt your analysis of the two platforms to take into account these differences: a retweet on Twitter may carry more weight than on Sina Weibo for example, and the achievement of becoming a trending topic is perhaps less significant.
In China, you’ll also see advocacy affected by astroturfing behaviours, with companies employing writers to seed positive content about their brand online and to attack competitors with negative reviews. One notorious case saw a major dairy company spreading online rumours about their competitors with a fabricated story about them selling contaminated milk products. Take trends such as these into account as you approach any measurement of advocacy strength.
In an environment such as this, you’ll probably want to invest some of your monitoring efforts into carefully identifying the source of reviews for both yourself and any competitors you’re monitoring. Don’t make the mistake of drawing the wrong conclusions about consumer behaviour and brand preferences by basing your understanding on the values and behaviours of your home market.
The fragmented nature of the social media landscape in China also means that you’ll see different demographic groups gravitating towards their own platforms much more than in a country such as the US or UK where mega platforms such as Facebook tend to cater for a wide range of participants.
China’s social landscape often sees more competition, and historically it’s not been an uncommon phenomenon to see at least two major local players for each type of social media platform: Sina Weibo and Weixin are currently rubbing shoulders; Tudou, Youku, Sougua, and Qiyi compete in the video market; and until fairly recently Renren and Kaixin001 were in stiff competition.
If you’re dealing with a unique social media environment such as China it’s probably wise to conduct analysis separately from other territories in which you operate to make sure you really understand what’s taking place and what it means. Local market knowledge and native-speaker language support will be invaluable.
In any case, it’s unlikely your brand will face the same competitive environment in any two markets. With different business goals for each market, your strategic approach to social media is best adapted, if not completely re-written, for each territory in which you operate.
Metrics & tools
How to measure
When measuring any aspect of social media it’s important to be able to distinguish what’s really important to your business. A wide range of metrics can be gathered across a huge number of platforms and it’s easy to become bogged down in data.
Your metrics should be determined by what you want to achieve through social and your business’s strategic goals. If you’re operating across multiple geographies, it’s probably the case that your company has different objectives in each market. You’ll need to decide how social contributes to these efforts in every market in which you operate.
The temptation is usually to track too many unnecessary metrics and lose focus or to concentrate on easily measurable metrics such as followers or likes whilst overlooking to question what value lies behind such metrics. It’s easier to track followers, fans, and likes but less easy to determine sentiment or intent.
The key is to first determine why your business is participating in social media in the first place, then to decide what data you actually care about. The analysis of your social media presence should be determined by what you want to achieve in that channel. This allows you to measure the impact and value of social media by creating metrics and key performance indicators in social channels that align with your key business goals.
It’s a very good idea to determine what success looks like by setting targets for metrics such as rates of conversation participation, both across social media generally but for each channel in particular. This makes it easier both to measure success and to identify key areas where you can improve what you are doing.
This approach helps you identify which social platforms are working best for you. The sites that send the traffic with the best conversion rates should become the priority of your social media strategy.
Which tools to choose
Tools to measure social media can be limited in scope. As an example, Twitter tools often offer the ability to measure a very limited range of metrics despite Twitter having its own analytics platform providing users with a broad overview of their tweets including, engagement, impressions and follower demographics.
For example single-purpose single-platform tools such as who.unfollowed.me which offers insight into who is following your account or has unfollowed you historically, or single-purpose multi-platform tools such as Klout which specifically gauges social influence. Many of these tools are low cost, but their scope is extremely limited in terms of what platforms they cover and which performance aspects they monitor.
Better insight can be gained from more comprehensive tools which have the capability to analyse a range of metrics across a variety of platforms. These include Socialbakers or HootSuite, both of which monitor the bigger social platforms such as Facebook, Twitter, Google+ and Youtube. The disadvantage to the global business is that these tools have often tended to have a regional or language focus. This means that they might not cover all of the markets or platforms you operate in.
There are a number of factors to consider when selecting which monitoring tools you’re going to use, not least functionality and coverage. You may need to balance having all the features you want against coverage of the languages and platforms you want to monitor.
Depending on how you handle reporting, it’s a good idea to also consider what report building or charting features are offered. Having great monitoring tools is one thing, but you need to be able to feed your observations back into your business quickly and easily in order to respond appropriately.
Usability and a well-designed interface are important and you’ll need to consider the cost; some tools are available for free, others offer their pricing based on the number of users you want to have access to the interface, or on the amount of data or features you need and what level of support.
Depending on how your global business is structured, you may want to consider location-based tools to let you drill down to a local level. Some monitoring platforms offer the ability to filter mentions based on country and assess brand reputation on a locational level. Some also offer the ability to do so on a language level.
If you get the opportunity, try and take advantage of trial periods before you select the platform that’s right for you. It’ll probably be the case that one platform can’t fulfil all your requirements and you’ll need to experiment to get the range of reporting that’s right for your needs. You may find yourself operating across several platforms to get access to the data you require.
What to measure
Fans, followers, likes
Increasing your fan and follower numbers is important, but it’s a bit of an unreliable indicator of success. The majority of social media are relatively stagnant, are spread between multiple social networks and consume more information without engaging or contributing much.
It’s good to set reasonable targets to increase these numbers but remember engagement metrics, such as conversation participation and sharing, are more important than the number of followers.
Be wary of agencies promising a fast increase in follower numbers as these are easy to fake and may be of low value. One high-end fashion brand sells almost exclusively to high net worth women over 35 years old, but because this segment is difficult to engage with on social media it has actively engaged more twenty-year-olds among its online followers. This makes the brand look popular but isn’t reflecting true engagement with the audience it really wants to access.
Conversation participation metrics should give you an indication of how people are interacting with your brand. These metrics monitor things like comments on a Facebook post or responding to a tweet. It’s difficult to judge what constitutes a successful level of conversation so you’ll probably want to determine your own internal benchmarking for the kinds of social campaigns you’re doing by which to measure your progress.
It’s then a good idea to set targets to increase your posts’ average number of comments, likes, or retweets as part of a process to refine and improve what you do in all your social channels. Track which sorts of content elicit the best sharing responses and use this to inform your social content strategy to ultimately increase engagement moving forward.
Social reach and influence
This looks at whether or not people are sharing your content by retweeting or writing a blog post about your social media campaign. Not all sharers are equal: you’ll want to monitor who is the most influential in your sphere both in terms of their number of followers but also how influential they are for your specific industry or subject.
It’s great if your content is shared by a follower with 1000 people in their network but it might be even better if that follower is a trade journalist in your industry even if they have only half as many followers.
It’s now common for retailers to enlist the help of social media influencers to bridge the gap between customers and their brand – especially in the beauty and fashion industry. These influencers usually have far fewer followers than celebrities, plus higher engagement rates with their audience allowing brands to make genuine authentic connections with new followers.
Sentiment analysis (SA) is an attempt to gauge attitude towards a brand or product by analysing comments or reviews about it across the internet, encompassing blogs and social networking sites. Sentiment analysis uses natural language processing, employing artificial intelligence to assess the strength of opinion expressed in written form. Inevitably the limits of machine intelligence lead to some inaccuracies.
For example, SA processors may not be able to identify the irony, sarcasm or even take ‘neutral’ sentiment of social media posts or whether they’ll be counted as positive or negative sentiment. It all depends on the sophistication of the sentiment analysis tool.
It’s a good idea to take measures of sentiment as a guide rather than a completely spot-on measure of brand feeling. It’s likely to be the case that the general sentiment about your brand changes more quickly than the way people express themselves, so sentiment analysis measures are a useful indicator of how sentiment about your brand changes over time.
It seems to be the case that most sentiment analysis programmes that are commercially available tend to have been developed exclusively for the English language. Very few companies are offering tools which cater to multiple languages. Part of the challenge with this is that it’s often the case that people from different cultural backgrounds express things like passion, sincerity and sarcasm to different degrees.
There’s been only limited research into sentiment analysis across multiple languages, but it’s been shown that where languages are assigned scores based on the positivity of sentiment expressed, some languages such as Chinese and Korean have a tendency to produce higher sentiment scores than others like Italian. This suggests a linguistic tendency towards positivity in some languages.
It’s also the case that automated language translation programs on which sentiment analysis are often highly inaccurate, sometimes skipping amounts of text or making serious errors on what they do translate, and reduce longer form text into sentence fragments in an unhelpful way.
Whilst it’s worth exploring sentiment analysis tools, it’s important to bear in mind the inevitable inaccuracies of such tools. You may find it more helpful to use a dedicated SA tool for each language market in which you operate, if available, rather than try a multilinguistic one, and never abandon examining individual brand mentions to at least some extent. This is time-consuming but is important to help you understand what’s really going on.
McDonald’s has a massive international social media presence, with 2.5-3 million mentions of the brand through social media each month. This activity is tracked using tools including Hootsuite, Radian6 and Sprout Social, with speed of response and visibility of emerging trends being given top priority by the company.
McDonald’s main goal is to gain a better understanding of trends as they bubble up and the company is constantly seeking to improve its performance in this area.
The company has learnt from its 2012 social media mistake. The company started using the hashtag #McDStories which was rapidly taken up by critics of the brand, hijacking the trend to comment on bad experiences or make points about business practices. Unfortunately for the fast-food brand, the hashtag is still currently being used by Twitter users.
Hashtag hijack social media fail by McDonald’s
McDonalds’ insights team constantly monitors trend lines to see what’s happening at the macro level, but the team also monitor as many individual postings as they can. This is resource-heavy but means the brand isn’t reliant on sentiment analytics and is more likely to be closer to the action this way. Nipping problems in the bud before they develop and become bigger concerns seems to be a key priority, and the social team works closely with the legal team to determine the best response.
McDonald’s Director of Social Media describes them as being in constant test-and-learn mode. In a situation such as releasing a new product, they will prepare many different messages to put out across various social channels. As these are released, the team will evaluate how people react to them.
This includes using A/B testing, where messages are phrased different ways to see which performs best. For example, when releasing a new hot sandwich, McDonald’s found that emphasising that it contained bacon seemed to drive more engagement. This learning was applied next time a product containing bacon was released, to positive effect.
McDonald’s is also experienced in coordinating teams across time zones – something they did during the London Olympic games in order to manage the HootSuite dashboard in different time zones and across staff shifts. McDonald’s has an innovative approach to social media which also applies to analytics, with a special emphasis on speed of response, identifying trends as they emerge and responding in real time.
Coca-Cola keeps a very tight concentration on how content for its social channels relates to the brand strategy. This means monitoring very carefully how the content they produce is shared and how it flows across networks and cultures. One approach to raise engagement is their habit of posting questions that require short 1 or 2-word answers, making it undemanding to engage with the brand.
With 33 mentions of Coke per minute in English globally alone, Coca-Cola set out to become the world’s most socially responsive and engaged marketing organisation. To do this, the iconic soft drink brand created The Coca-Cola Hub Network in 2013, consisting of a sophisticated high-tech control centre – separate from digital marketing teams – that listens, analyses and engages in real-time.
The hub monitors 600 or so accounts on Facebook, Twitter, Instagram, Snapchat and other social networks looking at social mentions and online conversations not only to gain insights but also to tie together insights its markets are reporting.
For example, social media strategist, Ben Kealy, noted that when small towns in America were going through processes of regeneration – renovations of old buildings in particular – large numbers of old Coca-Cola signs were also being repurposed. This allowed the team at the hub to identify stories about the brand that would have otherwise been overlooked.
Social successes for Coke included repurposing video content that had been popular on YouTube and making this into a TV advert which was then extended to other markets. Coke have referred to this approach as ‘Liquid and Linked’.
This refers to ‘Liquid content’ which is creative work the audience finds so compelling, authentic and culturally relevant that it can flow through any medium via a viral effect. ’Linked’ content is linked to the brand strategy and business objectives. Essentially Coke is creating a brand story that is consistent across all channels and markets by gaining a very clear understanding of what works.
What do you want to achieve
Make sure you know what you are trying to achieve in social channels and use this as a focus when you approach measuring and monitoring.
Don’t compare apples and oranges
Metrics mean different things on different platforms. Subtleties of language and culture should also influence your analysis of social media activity around your brand.
Don’t rely on sentiment analysis tools too much
It’s not easy for artificial intelligence to get sentiment analysis right; make sure you keep checking the data behind the analysis to keep abreast of real conversations.
Consider your approach
It’s wise to conduct social media monitoring at local market level as well as globally. Your specific goals and competitive environment should inform your approach to social media in each market you operate in.
Choose your tools carefully
You may need to rely on more than one to get the data you require.