Nextdoor’s New Reminder Aims to Cool Down Heated Conversations

Nextdoor, a social app that connects neighbors, wants its users to think twice before posting a mean comment during a heated conversation.

On Tuesday, Nextdoor said it’s releasing a new Constructive Conversations Reminder in the US that will encourage neighbors to respond to comments with empathy if the conversation is likely to provoke an argument. The reminder tells users to find common ground, challenge the idea not the person and identify clear actions neighbors can take. Nextdoor is hoping the reminder will decrease the amount of negativity on the social media where people discuss hot-button issues such as politics and crime. The company is planning to roll out the reminder internationally over the next quarter.

The reminder builds on other tools Nextdoor has released to encourage neighbors to be nice to one another online. In 2019, the company released a kindness reminder that alerted users if they were about to post a comment that was similar to previously reported content, giving users the ability to edit their reply. Nextdoor will also notify users if they’re about to post racist remarks.

Like other social networks, Nextdoor has grappled with complaints about racism and harassment on its platform. Social networks are trying to improve content moderation as civil rights advocates, politicians and others pressure them to do more to crack down on online content that can lead to real-world harm. At the same time, conservatives say their content is being censored, allegations social networks deny. 

Billionaire , who runs and SpaceX, struck a deal last week to buy Twitter for $44 billion because he thinks the company isn’t doing enough to safeguard free speech. The First Amendment, though, doesn’t bar companies from restricting speech. The deal has sparked a debate about what social media companies should do to improve content moderation. 

Content moderation on Nextdoor

Nextdoor has rules about what users can’t post on the social media app such as scams, child pornography and hateful language. Unlike Twitter, Nextdoor also verifies a user’s address and requires neighbors to provide their real names. The social media site relies on more than 230,000 volunteer community moderators who vote in groups about whether a post violated Nextdoor’s rules. That means content removal could vary depending on the neighborhood. Swearing, for example, might be more acceptable in New York than other cities, Kiran Prasad, head of Product at Nextdoor, said in an interview

«We really do believe that community moderation is the way to go…because it allows for the nuance of the different environments. Saying the whole US has to work one way I think is just flawed,» Prasad said. 

Nextdoor uses data from voting to train its machine learning models where computers learn and adapt to new information without being explicitly programmed. That helps Nextdoor predict whether a user is about to post a hurtful remark or if a conversation is likely to become contentious, prompting the platform to show users reminders.

Using signals such as keywords, the rate at which people are commenting and a user’s past behavior, Nextdoor can predict if a comment thread is likely about to get heated. 

Prasad said the company has already seen success with kindness reminders. In 2021, Nextdoor said in a that neighbors who received a kindness reminder edited or withheld their post or comment on average more than 34% of the time.

Prasad said the company will also show content lower in its feed if they are contentious. 

«We are not engagement at all costs so we will downrank this content even if it’s contentious and…getting people to engage more. We want to make sure that we’re a kind platform first and foremost,» he said.

Nextdoor said it’s also trying to be more transparent about how it moderates content. Users worldwide will receive a notification within the app and via email if their post or comment was taken down along with an explanation and a chance to appeal. Previously, Nextdoor users could already appeal a decision, DigitalMiddle but they weren’t shown a notification about removed content.