House of Fraser: How to Use Social Media Data for Investment Risk

What can investors learn from how companies are discussed in social media? How can social media data be used in risk management? As House of Fraser becomes the latest casualty on the high street, we analyse social media reactions in the twelve months before the company's move into administration.

I was once asked by an investment fund if social media can offer any clues to the outlook for retailers and consumer brands. It’s an interesting line of enquiry, and one which raises several more questions about how we interpret what people say to each other in social media and what signals we should be looking for.

Recently the UK high street suffered another major casualty with the collapse of House of Fraser, the department stores group now subject to a rescue offer from Mike Ashley’s Sports Direct.

Judging by the steady decline in the value of House of Fraser debt over the past twelve months, investors in the company’s bonds had a fairly good idea that all was not well; senior management changes, expensive rents and the chain’s failure to fully embrace online all contributed to a loss of confidence.

But what about customer confidence? In amongst the many articles written about House of Fraser in the year before administration, very few commentators thought to ask what customers were thinking.

So let’s track back over the past 12 months. Were there signals in social media that customers were becoming dissatisfied?

They say ‘retail is detail’. Customers are good at picking up on the small details of their shopping experience. In isolation, each of these comments might not seem significant, but if we collect up the views of House of Fraser customers, a picture starts to emerge of items out of stock, delayed orders and fewer staff available to help on the shop floor.

Using our emotiQ segmentation model to analyse the feelings expressed in social media, Fig 1 shows how negative customer reactions grew over the course of 2017-18. More customers became fearful and anxious. Fear is the classification we use when people express feelings of apprehension or uncertainty. Much of this ‘fear’ came from not knowing if a product would be available or if a delivery would arrive on time. The consequence was more Disappointment and less Trust.  Customer goodwill was ebbing away as the company  failed to trade its way out of trouble.


I’m not suggesting that social media explains everything, but in the case of House of Fraser there was a clear signal of intent from customers. When taken into consideration alongside financial and industry data, the results of the social media analysis gave a more rounded understanding of the likely outcome.

Looking ahead, I think investors can make more use of social media research as part of their risk tool kit. House of Fraser is a case in point, but the approach works for investments on an upward trajectory too. We can measure customer surprise and delight just as effectively.

Speed and scale can make social media a useful leading indicator and a source of new thinking. Often the signals we find in social media confirm a hypothesis, but sometimes the results are unexpected. When this happens, our data creates new opportunities for investors to challenge accepted wisdom and to act differently. 


We welcome opportunities to talk with investors and industry commentators who want to investigate how social media data can be used to boost the speed and effectiveness of research. If you’re thinking ‘we should at least explore this’, please contact us at


Have your say