5 Epic Social Media Automation Fails: Bots Just Aren’t People

So you’ve set up your social media automation bot and it’s plugging along automatically updating your favorite social media outlets and even responding to tweets and posts—now you can just sit back and let your social bot do its thing, right? Not quite. While social media automation can be a godsend for any busy marketer, it can also hurt your brand if you aren’t paying attention.

Here are 5 examples illustrating why relying too heavily on social media automation can be disastrous.

1. Life doesn’t always allow for scheduled posts; there’s simply no way to foresee everything and anything…and frankly, life just gets weird sometimes. Let’s say you’re a B2B company located in a city with an MLB team and you regularly send out enthusiastic GO TEAM posts. It works out great because sometimes the team and players respond or retweet and it’s really helped you build your social reach.

But let’s say you schedule all your GO TEAM posts in advance, and one of your messages on game day is, “John Jones is gonna hit one out of the park today!” with a photo of the player in mid-swing. However, a few hours before your post goes live, something bad happens. Maybe John Jones is in a horrific car accident or arrested for beating his child. You forgot to take down your scheduled post and it goes live. Now your innocuous message is suddenly stunningly insensitive.

2. Trolls are everywhere and bots are fodder, as Coca-Cola recently learned when they had to suspend their Super Bowl automated #makeithappy campaign after Gawker pulled a scandalous prank. The intent of the campaign was to turn negative messages to positive by turning tweets into ASCII art. Unfortunately, the bot was unable to filter intent, so Gawker made its own bot called @MeinCoke which tweeted lines from Mein Kampf with the #makeithappy hashtag...and the Cokebot responded in the only way it knew how, by turning Hitler’s words into happy cartoons.
3. Bots don’t get sarcasm…or any other human emotion. Some social auto-responder programs are pretty great; they can run out a whole string of automated responses based on simple keywords and commands. Welcome to the future of social. BUT, be aware those bots can’t interpret emotions or create an emotional connection with your target audience. In fact, their attempts to do so can get pretty ugly, and fast.

For example, a bot response to something like, “Oceanic is such a great airline. You never know where you’ll end up.” might be, “Hey, thanks! We’re glad you enjoyed your flight.” …and it could go on from there. It’s not hard to imagine your Twitter feed flooded with sarcastic trolls baiting the bot, followed immediately by industry blog posts pointing out your gaffe. Ouch. Credibility blown.

4. Bots can go completely nuts. AT&T hired an agency to publicize a Ticket Chasers’ campaign. The intent was to identify and tweet to people who met the following criteria:

• Bloggers (who would help spread the news about Ticket Chasers)
• People who live in the cities where the promotion is running
• People who mention basketball or March Madness

The ideal recipient might be a popular Boston blogger who writes about March Madness. The bot, unfortunately, was set up improperly, and the result was spamtastic. It sent a flood of tweets to anyone who met any one of the conditions. AT&T wound up apologizing for social media gone wild. Repeating the same mistake can get your business banned from Twitter search results.

5. Autoreplies can get you in trouble and draw unwanted attention. Without a human to screen posts, your bot may misinterpret, reply to, or repost an inappropriate or brand-damaging (or worse, obscene) message. Take, for example, when Bank of America’s help bot (@BofA_Help) tried too hard to “help” members of the Occupy movement by responding to their tweets. Yikes.

The lesson here is that automation can save you time and money, but careful monitoring is a must. Schedule posts that simply can’t backfire, watch ALL responses to ensure your replies are in line with their purpose, and check in often to respond to anything unexpected. You might also want to set aside time to go over replies in person. A bot might miss a warm lead by misinterpreting a reply—and that’s just plain bad for business.