Updated 9 min read
Spotting fake reviews: toward an international standard?
“On the Internet, nobody knows you’re a dog.” This quote, from a famous cartoon by Peter Steiner, shows two dogs sitting in front of a computer. It captures the dilemma faced by the managers of sites whose appeal rests on consumer reviews (such as Trip Advisor, Yelp, Amazon…): how can you be sure that the reviews were written by users who really experienced the service (a hotel room, a restaurant…) or the product (a book, a toy…) under review — in short**, how do you tell fake reviews** from real ones**?**
The stakes for these managers are high, because nearly 9 out of 10 French people consult consumer reviews on the Internet and 89% of them consider them “useful” or “very useful.” The weight of user reviews keeps growing, especially in the products and services sector, such as hospitality and restaurants. And yet, the vast majority believe that some consumer reviews are fake(1).
Inconsistent and inaccurate consumer reviews
According to a recent investigation by the Directorate-General for Competition (DGCCRF), Consumer Affairs and Fraud Control, the anomaly rate is rising: it stands at around 45% across all sectors combined (automotive, home appliances, furniture, clothing, services) in 2013, whereas this rate was 28.8% in 2010.
According to the DGCCRF, there are still just as many ways to skew a review, from moderation choices designed to remove negative reviews, to the more subtle promotion of positive comments, and including the writing of fake reviews or reviews obtained in exchange for the chance to win a prize draw. New practices are also emerging, such as misleading claims of compliance with the NF standard for online reviews.
Fake consumer reviews penalized at the international level
At the international level, significant abuses have recently been penalized. The Competition and Market Authority (AGCM), the Italian antitrust regulator, fined Trip Advisor 500,000 euros last December. The Italian authority found that, despite the messages about the “authentic and genuine” nature of the site’s reviews, some dubious content still slipped through (2).
Restoring trust in consumer reviews
The methods for collecting and processing reviews are indeed very inconsistent. The importance of reviews is in fact such that they have spawned a genuine business. Those involved include specialized portals (collecting, processing and presenting reviews), technical providers (in particular the developers of review-filtering software), polling institutes, e-reputation agencies, price-comparison sites, retailers and e-commerce sites.
So how can users’ trust in the truthfulness of reviews be restored? In order to promote reliable reviews and win back users’ trust, new methods are emerging online, with varying degrees of success.
Spotting fake reviews with “contextual algorithms”
Among them are “contextual algorithms.” Building on linguists’ work on lying, American researchers at Cornell University have identified textual characteristics specific to fake reviews. Using indicators such as writing style or register, text length or the use of certain words, the researchers claim they can tell fake reviews from real ones with 90% certainty. To test its accuracy, users are even invited to post a review in English in a field on the site www.reviewskeptic.com. The problem is that e-reputation agencies specializing in creating fake reviews adapt their writing style to the latest advances in detection systems. The race between research and forgers thus makes the detection filter difficult and fragile to maintain over time(3).
Spotting fake reviews with an invoice for the service
Some sites also offer “guaranteed” reviews. To do this — an important step — these sites’ teams ask for proof that the consumer actually made a transaction, by requesting an invoice for the service being rated. For example, on the matchmaking site truster.fr, the accuracy of the information submitted by registered professionals is verified before publication. Without registration, anonymity is impossible and the customer agrees to take a genuine approach. For now, Truster.fr only offers matchmaking in the Île-de-France region(4).
Spotting fake reviews through standardization
Beyond algorithms or sites that encourage people to provide invoices in order to better identify themselves, there are also codes of good conduct put in place by industry players.
In France, AFNOR (the French Standardization Association) is behind the international voluntary standardization project. This project builds on the momentum of 2013, with the publication of the NF Z74-501 standard, which sets out a reliable method for processing consumer reviews. Published last July 4, it is based on three stages: collecting, moderating and presenting the comments left by users. Among the many criteria to follow — insofar as the site manager decides to, since the standard is voluntary — the verification of the reviewer’s identity becomes more thorough, along with proof of purchase or use (5).
The first committee of the International Organization for Standardization (ISO) on the e-reputation of companies and public administrations was recently created. Its goals: to be aligned with other countries and to “make best practices accessible,” according to Olivier Gibert, spokesperson for the AFNOR group.
Driven by France, via AFNOR, this international voluntary standardization project is tackling ambitious topics that go beyond the mere processing of online consumer reviews. Currently, 9 countries (including Germany, the United Kingdom, Canada and China) are taking part in this work, while 18 others are content with observer status, the United States having left the working group. The next meeting is scheduled for September or October 2015 in China.
Laurent Petit, in charge of customer dialogue at Decathlon and chair of the ISO voluntary standardization committee, tells us more.
Is corporate reputation now tied to the digitalization of consumer behavior?
Corporate reputation is more than ever tied to the digitalization of our environment: there are currently more mobile phones than humans on earth. In general, we have an intense and loyal relationship with our screens. Today, the Internet is an integral part of the pre-purchase process for a product or service. 80 to 90% of customers check the product page online before buying it on a site or in a store. Of course, a company’s reputation is synonymous with trust for the future buyer.
This reputation is directly linked to the recommendations made by consumers. They are the ones who “make and break” a company’s image. Reputation is, after all, what people say about you when you’re not there. This is what companies have to accept. We now know that the comments customers make about their experience with a product or service have a greater impact on the purchase decision than the company’s official messaging.
Communication today differs from that of the 1980s or 1990s, when companies and brands delivered a top-down message to their consumers. Now, “customers” have taken over and contribute very significantly to their reputation.
The work on standardizing corporate reputation will help build this relationship of trust. A first piece of high-quality work was carried out in France with the creation of the AFNOR standard on customer reviews — work that we want to internationalize and broaden to reputation with this ISO Technical Committee.
In this context, what can the international voluntary standardization project bring?
Customer reviews are one of the key components of our work. In the end, we concluded that customer reviews raise the same issue in every country in the world. The French reviews standard is not fully replicable as it stands, but it is a very good working basis that we will be able to internationalize. Because we increasingly realize that customers around the world behave in relatively similar ways. The importance of reviews and their impact on the purchase decision no longer need to be demonstrated, whether you are in China, England, France or Italy…
Today, we benefit from tremendous involvement and close attention from the countries present.
Countries have indeed realized that a company’s reputation depends on communication by the customer, for the customer. When you walk past a restaurant, you feel reassured if it’s full; you can draw a parallel with customer reviews, while keeping in mind that quantity must not come at the expense of the quality and truthfulness of the reviews.
All the participating countries — notably China, Canada and Japan, to name just a few — enrich our sharing of experiences, and that is one of the goals of this standardization committee. We have the basic ingredients; all that remains is to perfect the recipe so as to satisfy as many people as possible.
And that is also why we would be delighted to welcome new countries. The more there are, the more international and internationalizable this standard will be. The strength of our committee is being able to work in each country with standardization bodies such as AFNOR and with companies from various sectors (hospitality, sporting goods, …)
Will this standard include the definition of a glossary on online reputation?
At our first meeting, we did indeed decide to set up a working group focused on a glossary of online reputation.
It is important to agree on the terms we are going to use. A very simple example around reviews: what is a customer review — a consumer review, a user review, a buyer review?
It is up to us, all together, to find definitions that satisfy the participating countries but also, and above all, companies and customers around the world. On this point, I am very optimistic, because every country present is genuinely committed to sharing experience and to mutual enrichment of knowledge, without any territorial ulterior motive.
Sources:
3. https://lejournal.cnrs.fr/billets/demasquer-les-faux-avis-du-web
5. http://www.marketingattitude.net/2014/05/nf-z74-501-une-norme-pour-les-avis-de-consommateurs/
See it on your own brand
Find out where AI engines mention you
Semji tracks which prompts surface your brand across ChatGPT, Google AI Overviews, Perplexity, Gemini, Claude and DeepSeek — then shows you the content to write to close the gaps.
ChatGPTGoogle AI OverviewsPerplexityGeminiClaudeDeepSeek
By
Equipe Semji
![9 essentials you can't overlook in your international SEO strategy [+examples] - Semji](/media/2022/07/MEA_STRAT_internationalisation.png)