This article is part of the On Tech newsletter. You can sign up here to receive it weekdays.
Here’s a conundrum of the digital age: It’s now possible and valuable to evaluate the feedback of others before buying a product, trying a restaurant or booking a hotel. But the growing ability to manipulate that feedback makes it hard to have faith.
We have more information than ever but may be more poorly informed.
A recent academic research paper examined the prevalence of paid customer reviews on Amazon, and how the company and shoppers responded to them. The researchers found that Amazon is deleting a large share of ratings for which merchants paid, as these violate the company’s rules. But in most instances Amazon didn’t act fast enough, so people were still influenced by the bogus reviews.
Customer ratings heavily influence what people buy online. This research suggests that Amazon could be doing more to ensure the credibility of reviews, and that we need to be even more skeptical shoppers.
Two of the paper’s authors, Brett Hollenbeck and Sherry He, talked me through how they combed groups on Facebook where merchants solicit glowing reviews on their Amazon merchandise, typically in exchange for a free product, cash or other incentives. Over nine months, their team tracked about 1,500 products with solicited reviews.
The research found these paid-for reviews worked, to a point. The average rating and sales of the products increased, but only for a week or two before ratings fell as soon as merchants stopped buying reviews. It was often still financially worthwhile for the merchants, they said.
An Amazon representative told me that the company devotes significant resources to rooting out and preventing inauthentic reviews, and that it catches many before they ever appear on its site.
The researchers found that Amazon eventually deleted roughly one-third of the bogus reviews, but typically only after an average lag of more than 100 days. Long before then, unhappy customers left a significant number of one-star reviews, a sign that they didn’t like what they bought and possibly even felt deceived by it.
Previous investigations and analyses have examined the cottage industry of bogus Amazon product reviews. This research is different in spotlighting Amazon’s response.
It’s impossible to catch all bad actors. But the fact that Amazon eventually deletes a significant portion of bought-off reviews shows that the company is able to spot inauthentic ones but doesn’t have the resources or doesn’t care enough to catch them before the damage is done.
“They have almost unlimited resources and this seems to pose a threat to people’s confidence in the company,” said Dr. Hollenbeck, an assistant professor of marketing at the Anderson School of Management at the University of California, Los Angeles.
The researchers said their findings had made them more cautious online shoppers and suggested tips for the rest of us. People should be particularly wary of reviews for products that are expensive and for items bought during the holiday shopping period and in categories where many merchants are offering nearly identical products. Those cases have higher instances of bought-off reviews.
They also said it’s safer but not foolproof to buy from merchants whose names you recognize. In their analysis, the majority of solicited reviews came from relatively unknown merchants, mostly in China. Here are more online shopping tips from Wirecutter, The New York Times’s product recommendation site.
When companies fuel people’s suspicions
It’s bad when companies aren’t upfront with their customers. It’s much worse when companies aren’t upfront in ways that fuel conspiracy theories.
Apple agreed on Wednesday to pay $113 million to settle an investigation by more than 30 states into its past practice of secretly slowing down older iPhones to preserve their battery life, my colleague Jack Nicas reported. In 2017, Apple acknowledged that it had reprogrammed its software to slow down phones with older batteries in some circumstances to prevent them from shutting off unexpectedly.
What Apple had been doing was not necessarily wrong, but the way the company communicated with customers was clueless.
Apple knew that people had suspected for years that the company intentionally made people’s existing iPhones slower when new models were coming out so that people would buy new phones. There has never been evidence of this, and Apple has gotten angry about these rumors over the years.
The problem was that when Apple trained its software to slow down iPhones — for perhaps a sensible reason — it didn’t sufficiently explain what it was doing. And that fueled the conspiracy theories that people already had about their iPhones. Apple created unnecessary controversy for itself.
Likewise, Facebook made a similar error when it acknowledged having human reviewers listen to audio clips from people using its services but didn’t properly explain why. There may have been legitimate reasons for Facebook to review people’s audio recordings from its Messenger app and other products, but the company wasn’t transparent about what it was doing — either to customers or its workers.
Again, this activity played into long-held suspicions that Facebook was listening to people’s private conversations. Facebook executives have rebutted these suspicions. It’s harder to trust Facebook saying it’s not secretly listening to people, when its workers do actually listen to people without their true knowledge or consent.
My free advice for rich companies: Don’t do anything that undermines your own attempts to bat down conspiracy theories.
Before we go …
Hugs to this
A worker found a tiny owl tucked in the branches of the Rockefeller Center Christmas tree. The owl is now safe at a wildlife center getting “all the mice he will eat.”
We want to hear from you. Tell us what you think of this newsletter and what else you’d like us to explore. You can reach us at ontech@nytimes.com.
If you don’t already get this newsletter in your inbox, please sign up here.
