"Do you think we should go to this new place for dinner tonight?"

"Hmm, not sure let's check what the online customer reviews say…"

If this sounds familiar, you, like about 80 per cent of all customers, rely on online reviews to guide your purchase decisions – or what in the marketing world is called electronic Word of Mouth (eWOM).

Super-charged by COVID-19 restrictions – when for many digital media has become the main window to the outside world – the influence of customer reviews on our purchases has intensified, and the implications for businesses are profound.

One bad restaurant review, for example, can result in up to 30 customers deciding to dine elsewhere. Though businesses are strongly advised to respond to negative online reviews, most of them don't know how.

Analysing almost half a million online complaints, my co-researchers and I have devised a framework to help businesses detect and mitigate the effects of online complaints to their reputation and bottom line.

Reviews are so powerful that when customers see large numbers of positive reviews for a venue, they can be convinced to visit regardless of what is actually on offer. In 2017, for example, the "Shed at Dulwich" became the top rated London restaurant on Trip Advisor after journalist Oobah Butler had friends write glowing reviews of what was actually just his garden shed and not a restaurant at all.

Gaining an extra star on Yelp leads to anywhere between 19 and 27 per cent more seated guests in a restaurant, increasing revenue by five to nine per cent.

But, what happens when these powerful customer reviews are negative?

A negative review can do just as much damage, if not more, as a positive review can do good. More than half of potential patrons, for example, won't visit if they read reviews where previous customers have complained about cleanliness or bad value for money.

At least 30 customers are lost for every one complaint about a hotel's cleanliness. Half of those considering booking will only do so if the venue has a 4-star rating or above.

Using an empirical and machine based text mining approach, we analysed 472,995 social media posts in which customers had complained about one of 89 listed US firms in the S&P 500 stock index.

We found that customer complaints can quickly accumulate but not all will have the same impact on businesses. Of the almost half a million complaints we analysed, less than five per cent went viral and truly influenced prospective customers.

The process by which prospective customers let negative complaints drive their own purchase decisions is best described by contagion theory – they essentially 'catch' the negative emotions the reviewer is displaying in their complaint simply by reading the review or complaint.

Our findings show that this contagion primarily depends on the emotions expressed in the review. The expression of intense, high-arousal emotions – like anger, fear, anxiety and disgust—are most likely to influence how prospective customers then feel and respond to the review.

A negative complaint also becomes more relevant to prospective customers if there is a sense of connection between reviewer and reader. For example, complaints by a regular reviewer compared to a one-time reviewer will carry more credence with a potential customer reading them.

But, while more emotive complaints appear to carry more weight, they lose their influence if they descend into abuse.

It means more appropriate and detailed complaints, such as "I was angry because I had to wait for an hour to be served" tend to be more influential than complaints using irrational or abusive language, such as "these were all idiots".

We further found that understanding how the review is written, beyond simply what is written, is key if businesses are to respond effectively.

For businesses facing a tremendous amount of reviews, automatic text-mining software may help in identifying those reviews with the most potential for influencing others, so that they can be prioritised for a response.

Responding is vital if businesses are to prevent the fallout from one bad review spreading, causing greater damage to reputation and the bottom line. But we found that about one third of the negative posts we analysed simply didn't receive any response from the business.

It is also important that when responding, businesses not only need to focus on reassuring the complainant but also reassuring any prospective customers reading the complaint.

When businesses did respond in our analysis, the majority used a combination of either attempting to move the conversation to a private channel like email or phone (61 per cent) and/or providing an apology (53 per cent).

Apologies and requests to switch to a private channel are generally effective in reducing the negative impact of the complaint on other customers, but the response needs to be provided swiftly.

If the complaint or negative review has been seen and responded to by many other customers before the business responds, then these approaches tend to be ineffective. Such late responses risk being viewed more as damage control rather than authentic attempt to respond to a customer's concerns.

We found that about three per cent of businesses responded with offers of compensation, but the wisdom of offering compensation is debatable. Yes, it might assuage frustration, but we found that the strategy tended to backfire by encouraging more customers to complain.

Depending on the circumstances, we found that a more effective approach was for businesses to respond either with empathy or a reasoned explanation, but only 6 per cent and 8 per cent respectively adopted either of these strategies.

In general, quick expressions of empathy were more effective than offering explanations or reasoning with the complaining customer. But, there is an important exception.

If a review contains an extraordinary amount of what we call high-arousal emotions, such as anger, an empathetic reply risks simply feeding the agitation. In these circumstances, a rational, fact-based explanation tends to better help calm the situation.

The ideal of course, is to have satisfied customers who then don't complain, but nobody is perfect. If an unhappy customer posts a negative online review, businesses can limit the damage by first prioritizing the most potentially influential complaints/reviews, and then providing well-considered responses based on our framework.

Our research suggests the appropriate management of complaints can reduce the influence of negative posts on prospective customers by up to 11 per cent.

That translates into less sharing and liking of negative posts, and more customers visiting your business.

Professor Stephan Ludwig is Director of the new Master of Digital Marketing at the University of Melbourne.

Banner: Getty Images