Online brand sentiment analysis

An introduction to measuring and managing online sentiment

Love. Hate. Frustration. Happiness. Security. Excitement. Satisfaction. Desire. Apprehension. Impatience.

Just a small selection of common emotions that we feel pretty much every day, as we go about our lives. Many of these emotions are felt and experienced as we interact with hundreds of brands throughout the day. Perhaps it’s the steaming hot coffee and service of Starbucks in the morning, making you feel satisfied, or the feeling of desire of the latest 7 series BMW, as you take the bus home from work, through to the painfully complicated payment system for something you wish to buy online, leaving you frustrated and angry.

The power of social media channels has created a new world of venting and consumer voice. For example with a Product recall—you can sure there’s more than one blog post about it. And then there is disappointing customer service? A quick mention of the brand shared with a few thousand followers won’t go unnoticed.

This last point is what I aim to bring a bit of focus to in this article:

How do brand managers and online marketeers really get to understand the diverse and highly fluctuating landscape of ‘real – time’ customer sentiment with regards to their brand online. What tools do they use to uncover these feelings from their online consumers, but perhaps most importantly, how can this insight and tracking ability be used effectively to drive strategic thinking to tackle the issues or build on successes?

Quality vs Quantity

Playing the numbers game on visits, shares, re-tweets and mentions online is what Katie Delahaye Paine, author of Measure What Matters: Online Tools For Understanding Customers, Social Media, Engagement, and Key Relationships, says are to quote 'quantity' or 'vanity metrics.'

She says that if we look at numbers only, it can give us a false sense of hope that our content is generating leads for our brand or business. With sentiment analysis, we dig deeper and look at 'quality metrics'.

'Quality metrics include opinions, feelings, satisfaction ratings, the quality of shares, comments, re-tweets, replies, ratings or conversations, as well as the overall quality of engagement over time. The benefit of this measurement and analysis is that it can help uncover and drive learnings from the key aspects of your brand; awareness, appeal, service, and content and allow you to uncover the positive, negative or indifferent aspects of your brand being shared online and importantly react to it.'

Of course, the overall quantitative analytics packages are hugely important, but used in isolation can be highly misleading in terms of online brand engagement.

So… How does sentiment analysis work?

Sentiment analysis platforms are like all other online data mining systems, they are based on bespoke algorithms. In this case it’s algorithms which recognise certain words as ‘positive’ or ‘negative’, letting you know if your brand is being adored or floored. If someone tweets using the word 'awful', 'disappointing' etc. then the sentiment analysis will assign that post as ‘negative’. You can get problems though if someone uses sarcasm or irony as the tone… most tools, takes everything at face value.

It is commonly known for the term 'opinion mining.' Opinion mining aims to determine how a certain person or group reacts to a topic they are being referred to. They react because they are either interested or involved.

Social media monitoring tools, some of which will be discussed below, generally have a high level of customisation in their search terms, and provide flexibility in the number of different things you can search for at the same time. This means that often you’ll be able to monitor your competitors too if you want to, and compare how their sentiment measures up.

In this video, Claudio Pinhanez , Senior Manager at IBM Research Brazil, talks about sentiment analysis in social media and explains why we should analyze data and how we can explore it. He uses the World Cup in Brazil that has just passed as an example citing interesting examples of a team, player or the country of Brazil’s brand image, amongst other interesting scenarios.

Sentiment Analysis Tools

Many tools are out there to be utilised by brands. Using the right tools in a dedicated fashion, with adequate time and budget assigned to investment in regular 'online listening and monitoring' will help you gather, analyze, and manage conversations about your brand.

This can then provide real insights and learnings on the levels of engagement your content marketing efforts are generating and the general online reputation you are creating online via content that your audiences are creating. So, below are just 4 0f the dozens of tools (ranging from basic, to comprehensive) that I have looked into to see how they may help provide value with regards to online sentiment analysis and their respective ability to provide valuable insight.

1. Social Mention

This is basically the social media equivalent to Google Alerts; a useful tool that allows you to track mentions for keywords that you will have pre-defined in video, blogs, Q&A, hash tags and even podcasts. It also indicates if mentions are positive, negative, or neutral.

When Mention spots that somebody talks about you or your brand you can instantly see this through one of their apps (Web, Desktop, iPhone or Android). Alternatively you can get daily summary alerts which shows you the mentions for a particular day.

It is good at tracking what people say about you or your business online to allow for instant or well timed responses, to avoid escalation procedures. Visit socialmention site.

2. Trackur

Trackur is another relatively comprehensive online monitoring tool that enables users to track a wide range of keywords, issues across mainstream and social monitoring. The ability to see the influence of the content creators around your brand too is a good advantage with Trackur, to enable you to gauge the importance of a response (if any).

Whilst being quite a clunky and unrefined user experience, it has a powerful analytics engine and it’s customization features allows it to be white labelled for end user, agency and reseller contexts. Visit Trackur site.

3. Google Alerts

This ‘old school’ alert mechanism is the ultimate hands off experience when it comes to online reputation monitoring. It is often overlooked in many online discussions around sentiment or online reputation monitoring and analysis. As a big fan of its simplicity and the nature of its email (or online reader) alerts I see it has having some great advantages.

• It’s free
• It’s fast
• You can monitor things you wouldn’t always think of as obvious beyond the brand mentions such as buyer behaviour, site security / spam and content distribution.

Whilst alerts clogging up your inbox may not be the ideal scenario for most people, a good tip is to set the ‘Deliver to’option as feeds that way you get all these alerts in Google Reader and can setup different folders for different alert types. One not to be ignored and usually it is used alongside a dedicated tool, rather than stand alone. Visit Google Alerts site.

4. Meltwater

Meltwater’s stand out feature is that it uses use low-level crowdsourcing to enable it to catch irony and sarcasm. Melt water actually employ people to read sample data and assign it a sentiment rating. These ratings are then used to update the algorithm in the hope that it will identify the sentiment more accurately. As a result, Melt water claims that its automated analysis is up to 80% – well ahead of the standard 50 – 60% accuracy that others claim to have.

Another great advantage of this platform is that it does’t limit the amount of results you have access to. Other tools such as social mention offer a certain number of results per month for a certain price, but this has limitations at times where the brand or product is in the spotlight and suddenly the posts increase by a factor of 50 – leaving the company having to pay extra to see those results.

Meltwater by contrast lets you have all the data that’s out there and it really powerful, but is one of the more expensive tools on the market at over £7k per year. Visit Meltwater Site

From Insight to Action

These small selection of tools, from the dozens available in the marketplace, all provide great power in the form of data mining and providing insight on audiences emotions towards your brand, product or service. As can be seen from this very small selection, each tool can provide numerical evidence of sentiment and semantic web data and others also have the advantage of rating the importance.

Like all big data activities, the real challenge is defining the why, what, who and when. Why are we measuring this in the first place, what tool(s) are we going to do use and what will we do with the data, when are we going to take this action and who is going to do it?

Tips to plan your sentiment analysis

Below are some tips on helping to answer these questions and enabling sentiment analysis to form part of your digital strategy.

• 1. Define its importance

Don’t invest based on the tools being there and needing used, invest on the need based on the online following you have and the outbound content. If you have a very low online audience and your online footprint relatively small for example, the timing of an investment in sentiment analysis, amongst a tiny audience, could be questioned.

• 2. Set KPIs

Without KPIs set you’ll never really know if you are meeting your goals. Of course you can be loose in saying you want to increase overall positive sentiment over time, but this then doesn’t then feed back adequately enough into the investment in time, people, software and other resources

KPIs should be set that relate to data returned on specific content posts, campaign sentiment, brand sentiment, mentions etc , individually and in combination and importantly linked to other analytics to include traffic sources, conversions and product / content views based on the sentiment around the specific content.

Importantly like all KPIs these need to be adjusted based on learnings, resource availability, your competitors activities and your own content trends.

• 3. Share the Sentiment

It may well be the task of an online marketing or brand manager to run the reports and setup the configuration, but it must be the responsibility of the organisation as a whole to have visibility of and know what online and offline activities may be contributing to the emotions being presented by the online audiences.

For example,  an online retailer could be getting negative comments on online customer service off the back of a team profile feature online. Written responses or addressing these concerns online can only do so much; the root of the problem may lie deeper in people and processes.

• 4. Beware of Comparisons

In today’s world consumers love to compare online and this extends to their and other consumers’ opinion, not only on price or product features, When comparing one brand with another,people will often express something along the lines of  'I love company X but hate company Y' or 'company X’s blue widget is so much better than company Y’s'.

Under analysis, this could produce a neutral result as the negative sentiment cancels out the positive, when in fact there was a clear expression of positive sentiment about company X and negative sentiment about the other. Being able to spot and disentangle these kinds of issues at the data processing stage is important to establish a robust foundation for the sentiment analysis.

• 5. Analytical Minds

Sentiment analysis, like any data or semantic web analysis for that matter, is not straightforward and without proper analytical minds, is open to substantial mis-interpretations and subsequent poor decision-making.

Online content producers are the ones taking action from sentiment and in fact their content has likely been the catalyst for sentiment ranges, along with other behaviours, but are online marketers the best people, or person equipped to analyse sentiment, amongst other things?

If you have the luxury in your business, think carefully about the distinction of analysis and marketing. Even a marketeer with fantastic analytical skills may be too close to, or be unwilling to accept change to his or her content activities, to truly drive change the changes required.

It is also vital that this analysis does not site in isolation and is monitored alongside other data collection and analysis tools including Google Analytics, E-CRM packages and Facebook Insights etc.

• 6. Test and Refine

Content Marketers around the globe will all likely agree that the same pieces of content read or shared by similar audiences do not always produce similar results. A/B split testing has shown this in contexts of conversion and online behaviour and audience diversity and a range of factors, many of which we’ll never truly be able to be understood,  due to the granularity involved in reaching the behaviour of each individual will always show trends different from what we imagine.

With a strategy that has sentiment improvement at its core, you should invest time in testing, analyzing and refining your messaging to over time understand what really drives lasting change. Reverting back to point 3, this can also drive change elsewhere in the organisation if the sentiment is directed towards an offline context.

So, in conclusion, sentiment analysis can help you gain insight into so much about how consumers feel about your brand, your campaign, your product and your service. There are lots of tools that can help and the choice is really important depending on your business, your time, your budget and resources. Most importantly though it’s essential to think about the right steps to taking action to this sentiment and thinking strategically about how you can actually improve, maintain or change this sentiment over time.

Image credit / Copyright: stuartphoto / 123RF Stock Photo