A brief experiment with Twitter ads

By Erin Griffith , written on December 19, 2012

From The News Desk

Earlier this week I tried out Twitter's self-serve ad platform.

Announced last year and slowly rolled out to users, the program holds a lot of promise for Twitter's revenue growth. Twitter had previously sold its Promoted Tweets, Promoted Accounts, and Promoted Trends manually to large brand advertisers; a self-serve ad product means mom-and-pop retailers can buy ads cheaply and with no additional labor on Twitter's part. Ad agency GroupM estimated it could bring in $400 million for the company this year.

Twitter has touted the power its ad format holds for small, local businesses. The tools are simple to use but can be targeted by geography. Restaurants, for example, can make sure their Tweets are seen by followers in their city. Not everyone is convinced, though. The company struggles with tying sales back to Tweets, which makes it difficult for small businesses to justify the spend. With search marketing on Google, the click-through rates are clear. Not so with this.

I tried it out on a $20 budget, which I ran through in two days, accumulating 1,028 impressions of my promoted Tweets. For $11, I got 26 people to click on my links at a 2.53 percent clickthrough rate. That clickthrough rate is a hell of a lot better than banner ads, which tend to perform below a tenth of a percent.

I dug around for what's normal and found that, on average, Tweeters with less than 5,000 followers see clickthrough rates of 7.16 percent. That is incredibly high. I have around 6,400 followers and a measly average clickthrough rate of 1.83 percent, according to Tweetthrough.

Surprisingly, my most popular Promoted Tweet wasn't for smart content or a funny video, but a solicitation. My Tweet for our upcoming PandoMonthly in New York garnered 10 clicks out of 619 impressions. According to, the original Tweet was Retweeted twice but only generated five organic clicks.

So promoted Tweets beat what I could do on my own. Still, they felt expensive. Not that Twitter provided any sort of benchmarking tools for comparison.

Meanwhile my promoted account got 1,870 impressions that translated to a mere eight follows, or a .43 percent follow rate. That cost me $9. My conversion rate was higher than others who experimented with it, but it still felt pretty lame, especially since Twitter said to expect 21 to 32 new followers a day. When I'm Tweeting regularly I add more than eight followers than that each day organically.

I ended the experiment having no good sense of whether it "worked." Any comp stats I found through my own research. Twitter told me what to bid and what to allocate, but I had no idea whom I was bidding on, and why those bid prices were recommended. In fact, I wasn't really told how the bidding even worked. I was allowed to pick a region to focus on and chose the entire US, since I write for a national publication. But my most popular Promoted Tweet was about an event here in NYC and most of the impressions were probably irrelevant.

Beyond that, I mistakenly missed the part where you designate which Tweets you wanted to promote. Twitter automatically surfaced what it deemed as my most engaging Tweets and started promoting them. At one point the machine started promoting this lame Tweet, where I whined about how badly my Promoted account was performing.


I had to continually jump in and filter out at-replies and other weird tweets that made no sense to promote.

Lastly, I couldn't control what time of day my Promotions happened. I woke up at 8am on the second day and saw that I'd already filled my daily budget limit between midnight and 2am.

The tools were oversimplified to a fault. I know bigger buyers have more controls and levers on their buying experience. There's no way someone would deploy large sums of money on something this basic. But money is just as precious to small businesses – they want to know what the hell they are getting for it. I know that $20 is a stupidly low budget. But I set my days for $5 a day, which is almost $2,000 a year. That's not a small sum for a local restaurant or shop.

One year ago I proclaimed that 2012 would be the year of ROI for social media marketing. For Twitter's self-serve platform, I'm not so sure that's true.

[Illustration by Hallie Bateman]