Lessons Learned from a Twitter Robot

A robot didn’t write this post.

But if technology continues at the current pace, a future with bot-authored posts may not be too far off on the horizon.

Depressing? Definitely. Crazy? Maybe not. Automation is alive and well in the world of social media and the debate rages on whether social media automation defeats the purpose of social media altogether.

An oversimplified summary of the argument: (Some) people advocate for the humanization of brands while (some) brands automate their advocacy toward people.

In the battle of automation versus humanization there’s no clear hero and no clear villain. (It can be as confusing and frustrating as that sentence above was to read).

So who better to turn to for answers on the automation debate than a Twitter robot?

For the past month, I ran a Twitter account where a robot produced every bit of content it spewed out into the social media ether. This robot relied solely on other people’s thoughts and added no knowledge of its own to the social media sphere because it was completely, 100% automated.

Here’s what I did:

• Followed 1,744 accounts suggested in this post by HubSpot

• Used RSS feeds to automatically Tweet content produced by 50 digital media blogs (mostly selected from the AdAge Power 150 and Junta42)

• Picked (what I thought was) a cool name (@contentasaurus) and accompanying avatar (pictured)

• Sat back and waited to see what happened

And here’s what the Twitter robot taught me:

Automation is not quite automatic

It takes a considerable amount of time and effort to set up an automated Twitter account which can cause you to wonder if your productivity shortcut was actually worth it.

You’ve got to do the dull work of setting up the feeds (I used Twitterfeed for this) and go through hundreds or thousands of Twitter accounts and mindlessly click “Follow.”

Then you’ve got to determine your sources and vet them for frequency and content (so you don’t flood people’s streams or find your auto-bot Tweeting bizarre personal posts from a blogger that are entirely out of context to your audience).

And even if you’re careful:

Sometimes, you look pretty dumb

Even when you carefully select sources, automatically posting from RSS feeds will sometimes lead to embarrassing gems like this one, which @contentasaurus dutifully Tweeted:

People will follow bots, but engagement is low.

After I set up the account, I prepared to be mostly ignored. But that didn’t happen (which probably says a lot about the quality of content produced by the sources I selected).

Some statistics from the 43-day life of @contentasaurus:

• The account posted more than 2,500 Tweets (about 60 Tweets per day)

• Which generated 443 clicks

• Which resulted in about 80 @mentions and 19 Retweets

• And a Klout score of 43 (for whatever that’s worth)

Ignoring people is painful. And it turns them off.

The hardest part of the experiment was ignoring the nice messages people sent to @contentasaurus; thanking it for sharing their post or wishing it a good weekend. Despite my impulse to act like a human and respond to them, I let the conversation languish, because that’s what a robot would do.

It turns out that when you don’t respond to people in conversations, they eventually stop trying to talk to you.

So, what are the lessons I learned from my Twitter robot?

• Automation can be useful, but it’s got to be carefully set up, monitored, and curated (In other words, not truly automated)

• Automation will never replace conversation

• Automation, when mixed with curation and conversation, can be very powerful

Related articles

• How to Automate a Twitter Feed with Hootsuite and Bufferapp (wolf-howl.com)
• When (And When Not) to Automate Your Social Media Updates (hubspot.com)
• How To: Use Social Media Automation The Right Way (oneforty.com)