Yesterday, I discussed the official launch of Sponsored Tweets and voiced my opinion about the service. I wasn't impressed. But there's a saying about opinions that starts with "Opinions are like...".
So I thought it would be helpful to look at Sponsored Tweets from a different perspective: performance. The proof is always in the pudding and when it comes to marketing, that means that the proof is in the performance.
So I decided to see what I could dig up to answer a simple question: does Sponsored Tweets make sense for marketers. And thanks to a comment, I stumbled upon the perfect case study.
On July 25, Jeremy Schoemaker (@shoemoney), one of Sponsored Tweet's "web celebrities" and an internet entrepreneur perhaps most famous for a photo in which he shows off a $130,000 AdSense check, posted a sponsored tweet for American retailer Kmart. Based on a screenshot he took of his Sponsored Tweets account that he shared on his blog, we know that Schoemaker was paid $250.90 to issue the following tweet:
The tweet was available to the more than 60,000 people who follow @shoemoney. So what were the results?
According to bit.ly, the URL shortening service which was used for the link, as of this moment, the sponsored tweet has generated 226 clicks. In other words, approximately 0.35% of @shoemoney's followers clicked through. Of the clicks, bit.ly indicates that only 118 came from users in the United States, where Kmart is located.
So let's do some simple math:
- All told, Kmart paid about $1.10 for each clicks (excluding whatever fees it paid to Sponsored Tweets).
- It paid approximately $2.12 for each meaningful click (e.g. a click from a user in the United States, where Kmart stores are located).
- If we're generous and assume that all 60,000+ of @shoemoney's followers saw the tweet once, Kmart's CPM was over $4. That's quite expensive given the market today. Of course, in reality, given the nature of the Twitter stream, it's likely that the tweet was seen by far fewer. Which means an even higher CPM. If we assume, for instance, that 25% of @shoemoney's followers saw the tweet once, generating 15,000 impressions ('tweet views'), the CPM for Kmart's text ad would be over $12. Ouch.
Of course, we have no way of knowing whether @shoemoney's sponsored tweet generated any sales for Kmart. With a couple of good sales, it's possible that the sponsored tweet generated a profit. Even though we're not able to look at actual ROI, we can still evaluate the sponsored tweet comparatively, which I think is even more important:
- With AdWords, KMart almost certainly could have promoted its sale using relevant keywords, many at a cost lower than the cost-per-click of the sponsored tweet. And since Kmart would have been able to restrict the display of its AdWords ads to users in locations where Kmart actually does business, it would have avoided paying for geographically useless clicks. In addition, it's worth pointing out that 'branding' with AdWords is free, which would have potentially been beneficial for Kmart's one-day promo.
- On an ad network or ad exchange, Kmart would almost certainly have been able to purchase remnant display inventory at a CPM less than the equivalent CPM here. And if we assume that the CPM of the @shoemoney tweet was closer to or exceeded say $8-10, it would have been able to access premium inventory at lower cost. Targeting is of importance (again); even with bottom of the barrel display ads, Kmart would have been apply to apply more targeting options than it had here.
Bottom line: Kmart's sponsored tweet was comparatively very expensive and dollar for dollar, it could have obtained more meaningful reach for less money. Therefore, whether or not Kmart turned a profit on a 'campaign' that is a rounding error in its marketing budget is a moot point in the context of opportunity cost.
The @shoemoney/Kmart sponsored tweets case study highlights several problems with the pay-per-tweet concept:
- There's no real targeting. Schoemaker may be a noted internet entrepreneur but what do we really know about the characteristics of his Twitter followers? Where do they live? How old are they? Are they primarily male or female? Are his followers frequently active on Twitter? How many followers are fake spam accounts? Unless you have this type of data, you're buying blind when you purchase a sponsored tweet. To put it another way, when you buy based on number of followers or some arbitrary assumptions about follower characteristics, you're buying the wrong data.
- There's no engagement. How engaged are @shoemoney's followers with him on Twitter? That's a tough question to answer but based on the fact that the Kmart tweet generated a lackluster 226 clicks, most on the first day, one thing is clear: few people on Twitter have a truly 'captive audience'. A big part of this is the nature of the stream; there's so much going on every second of the day and users can only focus their attention on so many things. My usage of Twitter is modest and I'd guess I probably miss at least 75% of the tweets posted by the ~280 accounts I follow.
- Sponsored tweets don't make for compelling content. I'd venture a guess that when @shoemoney posts an interesting tweet, he gets a decent response from his 60,000+ followers. But trying to get people interested in a sponsored tweet for a Kmart sale is a lot like putting lipstick on a pig (no offense to pigs). Content is king and on Twitter, it's content that drives results. It's why a post like the 10 Twitter Commandments has generated over 7,000 clicks through bit.ly links while other posts I've written generate far, far less.
Of course, I'm sure that IZEA and other proponents of this type of marketing will try to dismiss the numbers. Or, more likely, ignore them altogether through the creation of new myths. A presentation for Sponsored Tweets tells advertisers they can "reach the masses" and that Sponsored Tweets has the "ability to drive massive traffic from influencers in short time frames". Sounds like a swtweet deal right? But if you claim that you can drive traffic and a sponsored tweet to a 60,000+ follower account only generates a couple hundred visits, that's what you should be measured against. Unless somebody is going to claim that bit.ly lost, say, a few thousand clicks, it'd be interesting to see how IZEA responds to the fact that one of its "web celebrities" delivered very little traffic.
The proof is indeed in the pudding and for a marketer to truly get something out of Twitter, it's obviously going to take a lot more than sponsored tweets. Given that Twitter is still a relatively small and unproven marketing medium, any company that is going to make an effort would probably be wise to invest in a real Twitter presence because even though there's greater risk, the results seen here indicate that media buyers are wasting time and money.
Photo credit: thepinkpeppercorn via Flickr.