TripAdvisor weeded out 1.4 million fake reviews in 2018

Posted by Edith MacLeod on 23 Sep, 2019
View comments Content
Inaugural Transparency Report outlines measures taken to combat fake reviews in 2018.

TripAdvisor reviews.

TripAdvisor has outlined the measures it takes to combat fake reviews on its site as well as details of how many were caught by their automatic filters and human moderators last year.

It follows a critical investigation earlier this month by the Which? consumer rights magazine which found that some of the highest ranking hotels on the site were being boosted to their top positions by fake reviews.

In its first report of this kind, TripAdvisor said of the 66m reviews submitted in 2018, 2.1% were identified as fraudulent and 73% of these were blocked by automatic fraud detection systems before they were posted. Less than 0.6% of review submissions were identified as fraudulent and removed after being posted.

Fake reviews proportion.

4.7% of reviews in total were removed in 2018, for reasons including fraud and other guideline violations.

Most business or property based reviews fall into 3 categories: biased positive, biased negative and paid. The overwhelming majority fall into the first category, with someone connected to the business trying to manipulate an improvement to rankings or reputation.

Property-based fraud.

How content is moderated

TripAdvisor uses both machine and human moderation. In 2018, 100% of reviews went through automatic screening before being posted, with 2.7m being further screened by content moderators.  

Under 1% of reviews posted were reported to TripAdvisor by the community, of which 43% were removed after assessment.

In their report, TripAdvisor said they employed sophisticated review fraud detection technology, using best practice taken from the financial sector to spot abnormal electronic behaviour patterns.

Behaviour patterns.

Fake reviews are detected from all over the world, but TripAdvisor said fraudelent activity in the form of paid reviews was more prevalent in certain markets. 2018, when the World Cup was held in Russia,saw a higher proportion of fake reviews originating there.

Penalties

TripAdvisor said that under 1% of businesses try to flout the rules. Where it does find evidence of fraud various penalties can be applied including a content ban, a ranking penalty or - in the worst instances - a red penalty badge.

Reviews matter

Reviews have become a fundamental part of online activity with many customers using them to guide their purchase decisions.

For a site like TripAdvisor where users go to research hotels and trips, it forms a critical part of their content and needs to be seen as reliable and trustworthy.

In a press release TripAdvisor’s senior director of trust and safety, Becky Foley, said this issue was a top priority for them and took a swipe at other tech giants for not doing enough.

“As long as other review platforms aren't taking aggressive action, then fraudsters will continue to exploit and extort small businesses for cash. It is time other platforms like Google and Facebook stepped up to the plate to join us in tackling this problem head on."

Google took steps against fake reviews with a tightening up of the rules around review rich results last week.

Amazon faces its own problems with fake reviews. Which? reported in April that thousands of fake reviews were taken down over the past 3 years, but more needed to be done. It found a massive rise in the number of unverified and potentially fake reviews in March this year.

Any erosion of trust among consumers is bad for the sector. With online reviews influencing the purchase decisions of 93% of consumers, it's a problem that can’t be ignored.

You can download the TripAdvisor's full Transparency Report here.

Recent articles

Google says you don’t need to wait for a new core update for site recovery
Posted by Edith MacLeod on 14 September 2020
Why speed matters in 2020 [Infographic]
Posted by Wordtracker on 9 September 2020
Understanding Google Site Performance and Web Vitals
Posted by Owen Powis on 9 September 2020
Google makes it easier to find licensable images in Search
Posted by Edith MacLeod on 8 September 2020
Mobile marketing campaigns that worked (and what you can learn from them)
Posted by Matt Diggity on 2 September 2020