• Research reveals that online customer service tools such as chatbots can be harnessed to falsely promote poor quality goods
• Relying on chatbots may lead to misleading ratings that do not accurately reflect product quality
• Whilst chatbot interactions, especially with those displaying humanlike characteristics, may enhance user enjoyment, there is a danger of compromising the depth of reviews left

As much as it pains me to say it, ‘tis (nearly!) the season. According a poll conducted by UK retailer Marks & Spencer’s, half of shoppers intend to have their seasonal purchases made before the end of this month. Retailers are all too aware of shoppers trying to get the jump on things, with festive adverts hitting television screens as soon as Halloween has wrapped up and even stocking seasonal treats such as mince pies alongside the Back To School supplies in supermarkets.

The six-week period between mid-November and December 25^th is the busiest of the year for retailers. Of course, in this period we’ll also see Black Friday come to the fore, tempting shoppers to part with even more of their money to obtain one-off deals. According to Statista, UK shoppers spent £84.90 billion on food, decorations and gifts in 2023. The figure for 2024 is predicted to be even higher. 

That same report also revealed that 74% of shoppers last year chose to make their purchases online, and that almost half of those held a preference for doing all of their shopping in one place – using online retail giants such as Amazon, rather than visiting multiple, individual sites.

Finding the perfect gift for a loved one has, perhaps, never been easier. Last year, more than 600 million products were listed on Amazon, the majority sold via third party vendors, giving small businesses and unknown names an opportunity to access a much bigger market.

But with so much choice also comes the challenge of deciding which product is the best to buy.

Whilst price is undeniably a factor in consumer spending habits, when faced with multiple similar products with similar price tags, shoppers often need a helping hand in identifying the one that offers the best value for money, the highest quality and that most effectively meets their needs.

And here is where online reviews come in.

The power of online reviews

When we want to assess quality – whether we’re buying a gift, picking a restaurant or a holiday destination or even selecting what to watch at the cinema, we look to reviews to help steer our decisions.

Companies like Trustpilot, TripAdvisor and Which? make good business out of providing consumers with a trustworthy source of review-based information. Typically, the items, restaurants, locations and films that receive the most or the highest ratings are the ones seen as being most worth our investment.

Reviews, as a result, are big business. If you’re a retailer looking to increase sales, and especially where advertising is not possible, reviews and ratings are the best way to elevate your visibility and show your worth.

Sites like Amazon make reviewing a purchase easy, prompting customers to choose a star rating and leave a short comment, without needing to log back in to do so.

But what happens when those reviews are not as reliable as customers believe them to be? Fake ratings and misleading reviews are a growing problem for online retailers, and the advancements in technology such as AI are making the fakes both more populous and harder to spot.

For online retail giant Amazon, tackling fake reviews has become an ongoing battle as they seek to maintain standards and protect customers. According to the BBC, in 2022, Amazon reported over 23,000 social media groups, which contained a collective 46 million members and followers, that facilitated fake reviews for their products.

Sorting fact from fiction

But reviews remain a vital part of the consumer experience. “Online reviews are immensely important for consumers to make good purchase decisions,” says Dr Dimitrios Tsekouras, a Senior Lecturer in Business Information Management at Rotterdam School of Management Erasmus University (RSM). “For this to happen online reviews need to be helpful, and ratings need to accurately reflect product quality.”

So how can retailers act to ensure that customers can easily leave reviews when they shop, but also that the reviews they read before making a purchase are an authentic reflection of the goods or services they’ll be receiving?

Harnessing AI to make such online interactions easier is one way to go. Chatbots taking on feedback for a company have become incredibly popular.

Working with colleagues Dr Dominik Gutt, also of RSM, and Dr Irina Heimbach of WHU, Otto Beisheim School of Management, Dr Tsekouras’ research explores how customers engage with human-like chatbots when leaving online reviews.

And in doing so, the researchers discovered that chatbots, whilst handy for enhancing customer engagement, might be unwittingly facilitating inaccurate or misleading reviews. When consumers use chatbots to submit online reviews, the study found that they tend to give higher ratings but provide less detailed feedback.

The researchers conducted a series of online experiments and a field experiment to better understand how consumers approached the review process. In the online experiments, participants were asked to watch a short movie for which they later provided an online review, either through a chatbot with either moderate or highly humanlike characteristics, or through a conventional review form.

The field experiment took a similar approach. The researchers collected course evaluations for a university class either through a chatbot or through a conventional form.

They found that, for those that gave reviews via a chatbot, the overall ratings tended to be higher than those assigned via conventional forms. So star-based, or one-to-ten styled systems tended to provide a more favourable view of the product. The reason why? The researchers say that consumer enjoyment of the interaction plays a part.

“When consumers interact with a moderately humanlike chatbot, ratings – usually on a scale from 1 to 10 – increase because consumers enjoy the experience of interacting with the chatbot,” says Dr Tsekouras. “And if it’s very humanlike then ratings increase again because it is perceived to have a human social presence.”

Whilst this is great news for retailers looking to secure high ratings, the findings were not all so positive. Talking with a chatbot, it was found, also decreased the length of the review left, providing a reduced level of depth and detail.

And this lack of information resulted in the reviews being far less helpful than they initially appeared. “If ratings are better than they should be, and not very detailed, this can mislead consumers and harm the reputation of online selling platforms,” explains Dr Tsekouras.

Retailers’ responsibilities

The study, they say, further highlights the potential risks for online retail platforms like Amazon, when relying on reviews to guide consumer purchasing decisions.

Whilst chatbots might make the prospect of leaving a review easier and more enjoyable, the researchers say, it could provide a convenient avenue for low-quality sellers to inauthentically boost their products’ value, abusing chatbots to boost their ratings and disguise the low quality of their products.

Going further, the researchers suggest the findings call for policymakers to consider regulations regarding chatbot usage for review solicitation, to best ensure transparency and maintain the integrity of online marketplaces.

So, in seeking online bargains this year, it might pay to look beyond the stars and instead pay closer attention to the write-ups customers give – and perhaps compare across a few other reputable sites.

And, of course, as the saying goes; if something is too good to be true, it probably isn’t.

By, Kerry Ruffle

Interested in this topic? You might also like this…