Social Listening in Practice: Market Research
This guide looks at how to use social data for market research in a number of different scenarios, including:
- Creating a series of regular reports of a new advertising campaign
- Carrying out more detailed research among a defined audience
- Joining up social analytics and CRM systems
Get the download
Below is an excerpt of "Social Listening in Practice: Market Research". To get your free download, and unlimited access to the whole of bizibl.com, simply log in or join free.
What market research do we need?
Before embarking on any social media research the first question to ask is: why? The more tightly defined the question, the more beneficial the research is likely to be. What is the brief? What are we trying to achieve?
- Do people like our competitors’ products or services more than ours?
- What do existing customers think of our new advertisements?
- How should we redesign a new product for this particular market?
- How many people are thinking of buying X product, compared to Y product?
All of these questions can be addressed using social listening tools. Once the reason for the research is clearly defined, researchers typically have two interconnected challenges facing them:
- How to gather and aggregate vast amounts of data
- How to sort out the ‘signal’ from the ‘noise’
When set up with a clear purpose (the ‘why?’), social media analytics can support both quantitative (numerical, scientific and controlled) and qualitative (more ‘real world’ and interpretative) research approaches. Of course, we have to recognize that social media analytics will only ever bring back research from people who regularly use social media channels; the audience we’re sampling from is initially self-selecting and may not represent the population as a whole.
However, even with this limitation, the volume and accuracy of data available can help generate tangible insights for all sorts of organizations and brands. For the purpose of this guide, we’re going to look at how to use social data for market research in a number of different scenarios.
- To create a ‘snapshot report’ or a series of regular reports – e.g., a brand tracking report or sentiment analysis of a new advertising campaign
- To carry out more detailed research amongst a defined audience – e.g., what does our current audience think of recent activity we’ve carried out
- Co-creation of new products
- Joining up social analytics and CRM systems
But before we look at some scenarios, let’s consider the basics.
Case Study 1: Kmart - Advertising effectiveness research
When Kmart, the third largest discount store chain in the world, produced a new ad in Spring 2013, they wanted to research what shoppers thought of it before they spent significant sums of money buying media space.
They released their ‘Ship my Pants’ ad on YouTube alone and monitored reactions. Within a month, the ad had been viewed 17 million times on YouTube and been discussed 13,000 times on social media. Given the success on social channels, the ad was launched with a sizeable media spend on TV and did extremely well.
But just as importantly, Kmart’s initial social media research helped them to plan and create a follow-up ad, ‘Big Gas Saving’, which was launched a few weeks later and truly went ‘viral’ - generating 4000 social media mentions a day at its peak.
Using Social Media Research To Create A Snapshot Or A Series Of Regular Reports
One of the many benefits of a social media monitoring and analytics tool is that, once set up, it can be left to run to create regular reports based on the same data searches. It can take time and resources to establish the template, but once it is set up, it’s a remarkably cost-effective way to create regular comparison reports.
This differs from traditional market research panels, which might revisit the opinions of the same 1000+ people on a monthly or quarterly basis – although this process can also be replicated within social media analytics tools. It just depends how the ‘who’ is defined during the set up.
Ongoing or one-off report
When setting up an ongoing tracking report, the ‘W questions’ need to be even more tightly defined.
Who to track
Consider whether you value everyone’s opinion equally, or whether some opinions are more valuable than others. You’ll need to think about whether you want to research by geography or age, and whether you want to target people who have purchased or shown intent to purchase our product or a competitors’ product.
For one-off reports, such as for product or advertising launches, it is common to track absolutely every mention of the relevant keywords initially (e.g., new product or advertisement) and then analyze sentiment and intent to purchase within specific audience demographics at once.
You may wish to discover if more people are mentioning you online than previously, and understand why. You could also uncover the people who are aware of a particular product or advertising launch.
Alternatively, you could learn whether people are feeling more positive about your brand(s) or keywords this month compared to last month/quarter/year. Or, you could just track/monitor the extremes – to understand what’s causing particularly positive or negative sentiment so it can be dealt with elsewhere in the business. All of these types of market research dashboards can be set up for a one-off, or on an ongoing basis.
In terms of frequency, tracking reports and dashboards will have data sourced and presented dependent on how much data is required to make the research useful and how frequently decisions will need to be made from the insights generated – these could be weekly, monthly, quarterly, or bi-annually.
Discover which channels are most important. Is there a link between increased comments/brand mentions/sentiment on particular social channels and traffic to your website and/or sales?
Case Study 2: London 2012 – Making useful, real-time decisions
During the 2012 Olympics, LOCOG - the London Organising Committee of The Olympic and Paralympic Games - created hourly and daily tracking reports. These were to help them understand how visitors to Olympic venues felt about their experiences and how they expressed these feelings on social media. These tracking reports enabled the organizers to take swift decisions based on how people were interacting with their environment – including changes to announcements, music, feedback on food and drink provision and even real-time crowd-management.
Competitors/share Of Voice Report
One of the most frequently produced market research reports is designed to generate oneoff or regular analytics of a competitor’s behavior. This type of report can be broad, e.g., comparing like-with-like using a series of competitor brand names or keywords. It can also be a deep analysis of what people are saying about a specific competitor’s product or brand.
Some businesses use this type of report to track keywords associated with any number of brands, to understand ‘brand attributes’ as defined by the people who use the products day-in, day-out, rather than the brand managers. The ‘who’ here could be anyone who mentions any of the brands or keywords to be tracked, or it could be more tightly defined to actual customers. The ‘what’, ‘when’ and ‘where’ would clearly vary according to the main purpose of the research.
When Brandwatch analyzed the different types of food mentioned alongside CocaCola and Pepsi online, we found some interesting associations.
When people mention Coke online, they’re much more likely to mention burgers and French fries than those who mention Pepsi. But when people talk about Pepsi and then mention a type of food, 75% of the mentions are for pizza.
The research also tracked which brand of rum was most associated with Coke or Pepsi, and found that people talking about Appleton rum tended to talk about Pepsi, but Captain Morgan was more associated with Coke
Of course, we can’t draw any conclusions from this research alone, but such information can help brands understand which other products are most associated with their own, and can therefore influence everything from sales and marketing strategy through to commercial partnerships.
If the initial ‘who’ of any research is left quite broad – perhaps anyone who has bought or intends to buy one of our products - the social data can demonstrate significant insights about both the product and the audience’s relationship with it. Social media analytics can relatively easily identify if the audience purchasing a product is indeed the target audience in an advertising campaign, and consequently if the campaign should be amended in any way.
Case Study 3: ASDA (Walmart) – Dealing with criticism of advertising campaigns
In December 2012, ASDA in the UK - a Walmart company - launched their Christmas ad, which showed busy moms (their target audience) doing all the preparations for their family Christmas celebrations. Following a number of complaints to the UK’s advertising regulator, much of the British media chose to criticize the advertisement, suggesting that it demeaned men, was sexist and reinforced gender stereotypes.
The advertisement was widely discussed on social media – often critically. Asda carried out social media and traditional analysis to better understand the profiles of the people who were complaining about it and those who were supportive of the advertisement. Their analysis shows that their target audience – busy moms – overwhelmingly liked the ad, and understood the supermarket’s sentiments. They related to it and thought more positively about the supermarket after watching it.
Despite what appeared at first glance to be a social media backlash against the advertisment, their target audience liked the campaign. Asda apologized, but continued to very proudly run the campaign – due in large part to the research it carried out across social media.
Research Among A Defined Audience
It’s entirely possible to track the social outputs of a core group of individuals or accounts to mimic the way that some market research panels operate. The same 1000 or so people are tracked on an ongoing basis. But one of the benefits of a social media monitoring/analytics tool is that those panels can also be created on an ‘ad-hoc’ basis, and be self-selected, without them even knowing about it.
For this type of research, the key variable is the ‘who’ - and that audience can be defined very tightly indeed. A good social media analytics tool will make it relatively easy to define a target audience group in any number of ways. But as well as defining an audience by age, the best social tools can also help identify an audience to research based on factors including:
- Influence (defined by social media footprint)
- Intent to purchase
- Previous purchasers
- Competitors’ customers
- Geography – by country, city, and within a certain distance of a particular location
- Previous visitors to a particular location (e.g., a specific store)
- Propensity to purchase for someone else
- Loyalty to a particular brand/product
- Propensity to switch to a competitor product
- Relevance to a specific audience (e.g, with high lifetime value)
Insights gained from analyzing the social media activity of these tightly defined audience cohorts can prove very valuable when it comes to helping to shape or create a marketing strategy.
Case Study 4: H&M - The effect of celebrity partnerships
When H&M ran a celebrity-backed campaign in 2013, Brandwatch tracked the impact that working with celebrities including Beyonce, David Beckham and Lana Del Rey was having on the brand itself. The research looked in particular at the number of mentions of the celebrity and the numbers of people who expressed any ‘intent to purchase’ alongside the campaign. The research was illuminating.
Although David Beckham generated more than 2000 more mentions than the nextmentioned celebrity, Beyonce drove 2000 more ‘intent to purchase’ mentions than he did - for example, mentions such as ‘I’m definitely going to buy that bag!’ or ‘Ordering that top when I get paid’. Based on this research alone, the indicators were that a longer-term relationship with Beckham would probably lead to greater brand visibility, but a Beyoncé partnership could lead to more sales.
Case Study 5: Who really buys music concert tickets?
Brandwatch wanted to research the UK ticket market to understand who really buys concert tickets. Is it the people talking about the concerts? Looking at the UK concert market it was clear that more than 70% of all the conversations about events and concerts were for Pop/R&B concerts, with just under 20% of conversation around Indie/Rock gigs, and less than 10% for comedy. Amongst those R&B concerts, most people were talking about tickets for One Direction, followed by Katy Perry.
Researchers split out those conversations still further to look at ‘purchase mentions’ and found that for different groups, the audience who were the decision-makers for purchase were often very different from the audience talking most about the artist. For One Direction, 43% of the ‘purchase intent’ conversations were by parents, but for Katy Perry, the number was only 4%. This would suggest two very different sales/marketing campaigns would be required for these two artists.
Co-creation of new products/R&D
Social channels are increasingly being used by brands and organizations to generate feedback on product innovations. Organizations like Starbucks paved the way for developing consumer-facing, innovationsourcing platforms with www.mystarbucksidea.com, but many others have followed. In the UK, potato chip manufacturers Walkers Crisps turned to their communities, both on and offline, to help develop a new flavor to launch into the UK market. Their ‘Do Us A Flavour’ campaign attracted more than 1 million votes.
Brands around the world have found similar ways to tap into consumer interest, often by tracking and measuring feedback across social channels. Where people are encouraged to use hashtags to contribute to a discussion, it is relatively easy to identify suggestions and group them, but social media listening tools can also pick up and group other, more subtle suggestions – both encouraged and spontaneous – to help develop and refine products. Increasingly, brands are trying to encourage fans to join ‘closed’ community groups, such as forums, because this type of insight can be incredibly valuable. Social media analytics tools are just as useful in such ‘closed’ groups.
The future for social media analytics and market research
All of the above use-cases can be identified and implemented now. By asking and then answering what, who, when, where and why early in the process, social media analytics can play a significant part in market research and brand planning. But it’s also worth considering some of the more innovative approaches to social media analytics, and looking at one or two things yet to cross over to the market research mainstream yet.
Joining Up Social Analytics And CRM Systems
A good social media analytics tool can offer incredible insights into what people say publicly about products and brands, and how they interact with those brands on public platforms. But even with the best analytics tools, they might tell only part of the story. They can measure attributes like ‘intent to purchase’ or sentiment, but on their own generally can’t track actual behavior.
However, when social analytics data is studied in conjunction with internal CRM systems, researchers and analysts can draw even stronger conclusions and make even more credible predictions. The marriage of social analytics data with private CRM data provides a treasure trove of opportunities for researchers and analysts.
Of course, there may well be ethical questions here: should organizations tell their customers that they are tracking their behavior across their social channels, as well as their on-site behavior? Some businesses are already adopting this degree of transparency to help build out a more detailed picture of exactly who their customers are.