Handling social media crises

When problems happen and customers take to social media to vent their frustration, it may lead to a social media crisis like the one faced by United or Comcast. So, it is important to detect problems quickly, and act effectively.

I recently published a paper, co-authored with Finola Kerrigan, Dirk vom Lehn, Cagri Yalkin, Marc Braun and Nicola Steinmetz, presenting a model to help organisations manage a social media crises. This blog post offers a brief summary of the paper, and you can access the full version here, for free. The paper presents the framework and shows its application to the well-known social media crisis faced by Domino’s Pizza, when some employees of this chain of restaurants posted a video on YouTube showing them engaging in unhygienic food preparation practices.

Monitoring Social Media Conversations

The first thing to bear in mind is that one simple tweet or YouTube video can damage the brand’s reputation. Because content can be uploaded and shared so easily, it is important to detect problems early, and act very quickly. This means that businesses need to implement regular scanning of social media platforms, to monitor their brands, and to be alerted by changes in the mood towards their brand.

The large number and variety of social media platforms, the volume of content available, and the spreadability of social media content, make this a difficult task. However, it is also true that consumers do not post on every single social media platform. Rather, they tend to favour some platforms over others, and this may vary by industry and country. Likewise, consumers use different social media platforms for specific activities, for instance talking about positive experiences on Facebook, but negative ones on Twitter. This means that firms need to focus on those platforms that are particularly relevant for their customers, and which reflect their behaviours and preferences.

We also advise that firms need to go beyond monitoring quantitative information such as sentiment polarity towards the brand, and how it changes. Rather, brands need to monitor qualitative aspects such as whether the expressed sentiment relates to core reputation elements, or secondary ones. Messages about the core reputation elements of a brand are more damaging than messages about secondary elements. For instance, messages about cleanliness are more damaging for Domino’s Pizza than the United Nations (UN), because product contamination is the biggest threat to a food and beverage company’s reputation. Conversely, messages about unfair treatment of employees, like this story, is likely to be more damaging for the United Nations than for Domino’s pizza, because the former champions human rights.

Reacting to Negative Social Media Conversations

Once a crisis is detected, it is crucial to acknowledge it, as failing to communicate during a crisis may suggest that the company is struggling, or that it does not care about its customers. A recent example here in England is how Talk Talk handled communication regarding the hacking of its systems. While much criticism is being directed at Talk Talk’s management of their technical infrastructure, the company’s openness and immediate engagement with the media, including social media, might have helped contain the damage to the brand’s reputation.

During a crisis, companies need to communicate regularly through a variety of media channels, and making sure to include those platforms where the social media crisis is panning out. So, if there is a lot of negative content on YouTube, the brand needs to post their responses there, too, so that users searching for the original video can also see the company’s response. It is also a good idea to use the hashtags, keywords or titles trending the most (in relation to that social media crisis), to help spread the company’s response.

In terms of the message itself, it is important for the response to be congruent with the brand’s reputation, the users’ motivations and the functionalities of the social media platform. This is because social media users have specific expectations about how and when firms should interact with them on a particular platform.

Finally, it is important to engage credible social media influencers in the firm’s recovery efforts. This is because the opinion of trusted parties (e.g., industry experts or advisors) is more credible than any statements put forward by the brand. Social media influencers are not necessarily the same as those with influence in the offline environment, and so the company needs to identify the opinion leaders of the online communities trusted by the brand’s customers.

My colleagues and I hope that this framework supports managers in monitoring and reacting to ongoing online conversations about their brands. Do let us know how it helps you.

October 2015 round up

October was busy. Very busy. This included a lot of firefighting, which I tend to find very draining. My #5pm pictures show that I spent a lot of time at my desk. But, unfortunately, only a minority of that time was spent on writing or, for that matter, on anything remotely related to research. Instead, my month was dominated by teaching, and teaching preparation, plus lots of data crunching and report writing related to my role as programme director for the MSc Marketing.


As usual, here are my highlights for the month that has just ended.


10 06I worked really hard on a bid for a research grant for my work on the social media profiles of children, and managed to partner with three interesting researchers in other fields and institutions, as well as get a number of companies interested in the project. But, in the end, we did not submit the proposal for the call that we had in mind and, instead, decided to continue working on the proposal and target another funding opportunity.

I also continued my investigation of the use of social media in the business to business context, and compiled a number of really interesting (and varied) case studies.


Apart from the work on the research bid, I did not do much writing. Academic writing, that is. I did seem to do a lot of writing, but it was on programme-related, and this on administrative matters. As far as academic writing is concerned, I worked on some paper revisions. That was it.

But I did have a paper published. It is a piece looking at the handling of social media crisis, co-authored with Finola Kerrigan, Dirk von Lehn, Cagri Yalkin, and two of Dirk’s former students, Marc Braun and Nicola Steinmetz. It is available here, and it is open access.

November is AcWriMo and I am determined to go back to writing regularly and purposefully. I will write everyday: sometimes this will be a couple of hours. Other times, this will be just 15 minutes. But write, I shall!


I ran a session on sentiment analysis for our final year undergraduate Marketing students. I have been running a module on Customer Relationship Management for our MSc Marketing students, and a module on Services Marketing for our MBA students. And while this may represent only 20 hours of delivery, it represents around 5 times that in preparation. And, then, there is all the admin, the marking (dissertations, at the moment), and so on. No wonder I did not manage to do much writing this month.

10 18


Unfortunately, I had to miss a couple of events where I was hoping to up my learning, like a colleague’s talk on ‘bounded systematic reviews’, or a meet up on optimisation.

But I learned quite a lot about privacy enhancing technology, while working on the research grant about children’s social media profiles. That was really interesting, and reminded me of why I love to work on interdisciplinary projects.

And I have been learning, or updating my knowledge, about various marketing topics, from second marking MSc dissertations – for instance, last week I read a dissertation about neuromarketing, which is an area that I am very curious about, but not very knowledgeable.

What were October’s highlights for you?

Please do not cut and paste this article

Image source: Wikimedia

I was preparing a short article, and found a quote on the Financial Times that I really wanted to cite. So, I selected the paragraph in question, and copied it. Then, I pasted the excerpt into a word document. When I re-read it, I found the following sentences, at the end of the section that I had just copied into the word document:

High quality global journalism requires investment. Please share this article with others using the link below, do not cut & paste the article. See our Ts&Cs and Copyright Policy for more detail. Email ftsales.support@ft.com to buy additional rights. http://www.ft.com/cms/s/0/683f8cc0-6da7-11e5-aca9-d87542bf8673.html#ixzz3oMATcyq8

It seems that the FT is trying to tap into our innate need to reciprocate good deeds: we did high quality work for you; in return, you use the link below to share this article. As I was after a particular quote, only, these instructions weren’t entirely relevant for me. Though, the request did make me stop on my tracks, and I checked my article twice, to ensure that I was acknowledging the source fully and properly.

I am making a note of this example, to use in my classes when we talk about influencing consumer behaviour. You can read more about the principle of reciprocation, and how it influences behaviour, here.

I wonder how many people actually went back and used that particular link to share the article, though. Would you?

Spotted elsewhere: interview with co-founder of Instagram


A short interview with Kevin Systrom, co-founder of Instagram, by BBC’s Steph McGovern.



Spotted on the BBC website.


So what?

It is a short interview, but Steph McGovern and Kevin Systrom cover quite a bit of ground such as longevity on social media, Instagram’s rivals, kids spending too much time on social media, and so on. For me, however, the key segment is at 1m44s when the interviewer mentions selfies. Kevin Systrom says:

In college and high school I studied Art History. What are like the main subjects? Basically, it’s food and portraits. So, I actually don’t think that we are seeing the emergence of the selfie. I think, actually, this has been a trend throughout history. Pictures of your self for posterity sake. Being able to remember and be nostalgic of a time when either you were younger or were at a place. This is just something that has continued throughout time and I think, now, it is leveraging the fact that you can share it with your friends.


It seems to me like an attempt to gentrify the selfie.


What caught your eye on the web, this week?

Presentation on the challenges of using Twitter for sentiment analysis

As usual, I am sharing you with you my latest presentation. This one was on the topic of using Twitter for sentiment analysis, building on the work done with Yuvraj Padmanabhan (open access paper  available here).


I really don’t want to put you off using Twitter for customer insight. On the contrary – it has lots of benefits (slide 6), and various advantages over other methods of studying the impact of emotions on behaviour (slides 4 and 5). My goal is only to emphasise the issues to watch out for when doing so, and to invest some time in going through your data corpus and tweaking your analysis tool to make sure you really tap into Twitter’s great potential.


Also, really question the suitability of the medium to study what you are interested in. For instance, Pratik Thakar, Head of creative content for Coca-Cola Asia-Pacific, describes social media as a ‘big focus group’, which offers ‘a good way to identify trends and what people are talking about—but not why they are talking’.


Enjoy the presentation, and share your ideas and suggestions with me.

September 2015 round up

A few days ago, my line manager asked me if the start of the semester had felt like a great shock, following the sabbatical. That got me thinking about how different things are now and six months ago, and how I feel about it.

Six months ago I was doing things (reading, writing, meeting, discussing…) in depth, and I was doing things purposefully and with a plan of action. But now, my reading is superficial, my writing is scattered, my meetings are rushed, and my discussions are fragmented. Now, I am jumping from one task to another, and reacting more than acting.

It feels like six months ago I was in this very quite pool, doing lengths back and forth, and now I am in different one with a waves machine. Wave pools are exciting and fun. But someone has turned on the power of the wave machine so high that all I can feel, right now, is waves going up and down, and back and forth. I know that I will enjoy the pool again when someone turns the wave machine down a bit. But, right now, it is not fun at all.

My collection of #5pm pictures is a bit sparse this month, because I was without a phone for several days: first, because my phone had to be repaired; later, because of a poorly timed upgrade. I borrowed an iPad and took some pictures, initially. But the iPad is big and heavy, and I did not carry it with me all the time. So, on some occasions I took a picture well past 5 pm, while on others I forgot to take a picture altogether. Hence the gaps in the slide. But, hey, those gaps are part of my story for this month, too. The absence of pictures on some days tells something.


And now for my usual monthly highlights.


I did some research for a project on the use of social media in the business to business context, and for another one on the sharenting phenomenon and the implications it has for children as consumers. On three occasions, when I talked about the sharenting work with other people, I was asked whether I thought that parents, schools, etc that post information about children on social media are the good guys or the bad guys. This question (and, specially, the fact that it was asked three times) really puzzled me. I don’t see parents, or indeed, sharenting as good or bad per se. Just like parents that drive above the speed limit are not good people or bad people. I am ‘only’ interested in the behaviour, because behaviours are never neutral. They have  consequences. And, so, it is important that we study them, so that we can anticipate (and prevent) big problems. Does that make sense?

09 10Anyway, I am working on a research bid for this sharenting work, which I am deeply excited about… but also deeply frustrated because I really struggle to find the time and, above all, the mind space to work on it, in this ‘pool with a very powerful wave machine’. Can someone turned it down, already? Please.

On a different note, my applications for two small grants were rejected because… the funder has since changed their priorities. Really frustrating.


Writing has been sparse and really fragmented. It isn’t just poor quantity, but also of poor quality, which, I suppose, is to be expected when there are big gaps between each writing session and when the thinking process between each section is so superficial and fragmented.

My priority for October is to go back to routine and purposeful writing (as opposed to binge writing), even when there is too much going on.


IMG_0017This month I developed materials for the modules that I am teaching this semester, including a few podcasts.

09 14

I also have a new PhD student, who is looking at the role of social media in developing relationships between B2B organisations. Though, I am not sure that it is right to put ‘PhD supervision’ in the ‘Teaching’ header because I feel like I learn at least as much as I ‘guide’ (definitely not teach!) the candidate… and this is certainly the case with this new, very bright PhD student.


I went to a meet-up and learned about various Internet of Things applications – for instance, open sensors and their role in flood monitoring and management.

IMG_0415Also, my enforced smartphone detox taught me that my phone is really important for my productivity. I can answer many e-mails and note down things while I am waiting for someone, or on the train. My phone helps me to get out of bed in the morning. I use it to take pictures of things that I want to remember or check later, and it is faster than if I had to write it down somewhere. The podcasts that I listen to on my phone make my grocery shopping more bearable, and I learn something new between the vegetables aisle and the dairy one. And having my calendar at hand makes things a lot more easy and less embarrassing (I avoid to double book myself!).

What were September’s highlights for you?

Coca-Cola: When it comes to customer feedback, negative is more valuable than positive

As I am putting together some notes on the role of Twitter to study sentiment, I come across an interview with Coca-Cola’s head of creative content for Asia-Pacific, Pratik Thakar, here.

Image source: Mr Pratik Thakar’s LinkedIn profile

At one point, Thakar says the following about social media as a source of consumer insight:

I believe that social media is a big focus group. It’s a good way to identify trends and what people are talking about—but not why they are talking.

However, he goes on to say that, in his view, negative statements are more valuable than positive ones:

When people say good things, you don’t just take it as it is. Someone might be asking them to say it; there might be some design mechanism working. But when people are unhappy, they go super-loud, and they are genuine at that time. If someone’s not happy with our packaging or communications, we immediately hear it, and that’s good because [the reaction] can’t be designed or manufactured. If someone says ‘I love Coke, it’s so nice’, I won’t take that as input. So we have to figure it out based on context. I think negativity is more genuine on social media than positivity.

I wrote before about the value of complaints and customer criticism (for instance, here). But I never considered this angle that customers might be more genuine when criticising a brand than when praising it. The value of sentiment analysis of Twitter conversations is quite limited, then. Not just for all the reasons that Yuvraj Padmanabhan and I discuss in this paper (open access), but also because some expressions of sentiment may be more genuine than others.

What do you think of Thakar’s comment?