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, 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, 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 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 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.

Patricio Robles

Published 6 August, 2009 by Patricio Robles

Patricio Robles is a tech reporter at Econsultancy. Follow him on Twitter.

2641 more posts from this author

You might be interested in

Comments (11)

Save or Cancel
dan barker

dan barker, E-Business Consultant at Dan Barker


I mentioned this on twitter but thought I'd mention it here: Most of Shoemoney's followers are affiliates. $250 is expensive if the aim was direct sales. $250 is far less expensive to get the world's most famous affiliate to tell a few thousand affiliate fans that you're running a sale.

Bearing that in mind, not really "fail" but quite a clever idea?


almost 9 years ago


Ted Murphy

You are calling the click results of a single tweet for a single advertiser a "case study"? This is such a riciulously small sample that it has no statistical relevance what so ever.

What if this single tweet was a huge success? Would you tout the soulton as the most effective medium ever and encourage people to use it? Calling this a case study and trying to make a conclusion based on one tweet is professionally irresponsible.

This campaign involves hundreds of people, some with CPC's as low a few cents, some with CPCs of a few dollars. The blended result is very favorable for the advertiser.

Some things everyone should keep in mind.

1. Different messages have different click through rates for different followers.

2. I agree that followers aren't the answer, which is why we created FAR (

3. Prices for a tweet vary greatly in the marketplace. Celebrities carry a premium. The best campaigns utilize a mix of different tweeter types. We find that some of the smaller tweeters actually deliver the best value for advertisers. The trick is aggregating enough of them to provide scale.

4. Advertisers are provided with a wealth of information to make informed decisions uprfont. Our system is focused around transparency of indentity. You can hand pick the people you want to work with. If that isn't targeting I don't know what is.

Some Tweeters are going to create more value than others for specific advertisers. Advertisers have the tools to get rid of low performers and re-engage those that work best for their brand.

almost 9 years ago

Patricio Robles

Patricio Robles, Tech Reporter at Econsultancy


Clever how? Clearly given the number of clicks, it's pretty clear that very few people had an interest in this. I believe that Kmart has an affiliate program so if it wanted to get its affiliates excited about the sale, it could have sent them all an email gratis.


I had a hunch somebody from IZEA would respond with spin and sure enough, here you are.

The bottom line, unfortunately, is that results speak for themselves.

Let me ask a few simple questions:

1. You tout that you have the "ability to drive massive traffic from influencers in short time frames". Is @shoemoney not such an 'influencer'? You do list him as a "web celebrity" after all.

2. When exactly is a client supposed to expect the "massive traffic" part kick in? How many $250 tweets is a client supposed to buy?

3. If you state that "We find that some of the smaller tweeters actually deliver the best value for advertisers" then why do you so heavily promote your high-priced "celebrities"? Again, this statement contradicts your claim of having the "ability to drive massive traffic from influencers in short time frames". It seems to me that if your smaller Twitterers actually provide the most value, you're admitting that the 'influencer' bit is pure garbage.

4. Following up on #3, if "the trick is aggregating enough of them to provide scale" why does your buying model focus attention on individual Twitter users, especially high-priced "celebrities"? If aggregation is key, wouldn't you be better off with a model like Magpie?

5. How is data like the number of followers, followeds, ratio of followers to followeds, age of account constitute a "wealth of information"? Do you have any evidence whatsoever that this data is correlated with results?

6. Do you really believe that targeting is choosing who you want to work with? In other words, you believe that if I choose to advertise on Sites A, B and C, that constitutes real targeting?

Like I said, the proof is in the pudding. You can argue semantics ("this isn't a case study") but call it what you want: if you're going to sell a $250 tweet that essentially delivers nothing for the buyer, you shouldn't be surprised when someone points it out.

almost 9 years ago


Ted Murphy

1. If you don't think Jeremy is a web celebrity or influencer perhaps you should do some research. As I stated different campaigns will have different results based on a wide variety of factors. When the right advertiser message is put in the hands of the right infleuncer it can explode.

2. Can you buy "massive traffic" on Google for $250? No. For the keywords I buy the average is about $4 a click. That would get me about 62 clicks. The more you spend the more traffic you can generate. That's not rocket science.

3. We have generated 40,000 clicks from a single celebrity tweet. I am reasonably sure that influence matters.

4. The most effective campaigns should include a mix of celebrities and smaller, lesser known influencers. In both cases transparency of indentiy is key. Advertisers should try to pick the best people possible for their campaign.

5. Yes, and we provide that to the advertisers.

6. Do you have a better suggestion than hand picking individuals? I would love to hear it.

How many campaigns have you run through the system for how many clients? How much multi-variant testing have you done to enhance your results and test your hypothethis? None.

You freely admit your bias and predisposition to sponsored tweets.

To label this a "case study" is a joke. You should just call it a rant.

almost 9 years ago

Patricio Robles

Patricio Robles, Tech Reporter at Econsultancy


1. Frankly, I don't know what an 'influencer' is here. Please tell me. All I can gather from your comment that "when the right advertiser message is put in the hands of the right infleuncer it can explode" as it relates to the Kmart example I've given is that Kmart put the wrong message in the hands of the wrong influencer. Which leads to the question: does Sponsored Tweets do anything at all to deliver for advertisers what your marketing materials promote or do advertisers have to pray that their $250 tweet is being posted by "the right influencer"? Frankly, there's little value in your service if getting the right message in the hands of the right people comes down to guesswork and luck.

2. Please don't be disingenuous. Google doesn't promise massive traffic -- you do (your presentation uses those exact words). Google offers a pay-per-click service; you get a valid click from your targeted ads, you pay. Of course there's generally a correlation between budget and reach, but we're not talking about that. We're talking about comparative costs. Here, I pointed out that paying @shoemoney $250 was comparatively expensive on both a CPC and CPM basis for Kmart and a lot of the budget was wasted due to a lack of targeting (on Google, for instance, Kmart wouldn't have had to pay for clicks from outside its geographic market).

3. That's great. Since you talk a lot about transparency and all of your campaigns are apparently out in the open, why don't you enlighten us and share a 'case study' of your own? I'm sure a lot of us would be interested to see who tweeted and what was tweeted. After all, if we can figure out what's responsible for the difference between 200 clicks and 40,000 clicks, it would be valuable stuff, right?

4. Again, how do advertisers know who the best people are and when should they expect to start seeing results? Do they have to buy 10, 20, 50 100 tweets before they see a hint of progress? On a side note, you may have invented a new phrase: the lesser-known influencer. They're like influencers without all the influence, I take it? Which brings up another question: is everybody on Sponsored Tweets an influencer in your eyes?

5. I logged in as an advertiser and didn't see any statistical data that correlates the metrics you provide to actual or expected results. All I see is how many followers a person has, how many people they follow, a bunch of meaningless ratios, etc. They're just numbers. Not very useful.

6. You've stated that "the trick is aggregating enough [Twitterers] to provide scale". Yet your buying model is predicated on advertisers going through a database of hundreds (or thousands) of Twitter users, hand-selecting the ones that they'd like to tweet. Not only is this inefficient for a media buyer but it seems counter to what you say works best (aggregation). If you can't take it from there, you might want to hire someone to help you figure it out.

Finally, it's quite presumptive of you to assume that I'm going to spend time and money on Sponsored Tweets. Nothing in your response even hints that it'd be worth an investment.

I do have an opinion about your service but to call it bias and this a rant is a cop-out. I've pointed out logical flaws in your model and demonstrated how a campaign with one of your top "web celebrities" was a flop for a major brand. Your response to this has revealed contradictions between what you promote (massive traffic from targeted influencers) and what you say is the key to success (aggregating lots of Twitter users of all kinds).

The bottom line is that I've analyzed actual numbers from a campaign readers can look at and evaluate for themselves. The door is wide open for you to come back with a case study or numbers of your own that would show that the example I've provided here is an outlier.

You haven't done that. Why?

almost 9 years ago



No one is talking about how many of those clicks are from bots! does not break them down and this is, IMHO, the biggest problem new analytics yet have to face on a serious level.

We're doing a series of accuracy tests on all these clicks and in seconds after a tiny URL had been tweeted all these came and were registered as clicks! How many of them do you think are legit, human, clicks? Not a lot.

He who solves this puzzle will be rich.

almost 9 years ago



What is clear here is that there is a very poor match between Jeremy and KMart. If the promo had been for an Affiliate Marketing tool or ebook then the results would have likely been much better.

Perhaps someone will want to pick up the gauntlet here and provide us with a case study where the match between content author and offer are better? Maybe run another ad using Jeremy again and offer something appropriate to his audience.

almost 9 years ago


Dale M

I live the passionate debate on the topic of Good job Roman on your point about the fact that bots might account for some clicks that are not being accounted for.

almost 9 years ago


Terese Newman

I'm just curious here, but as a marketer trying to get the word out, say, about an upcoming regional concert, what are the options to using a company (software) such as Sponsored Tweets? So far, it's been just rolling up my sleeves and contacting bloggers, twitterers, article writers, etc. to get the word out. Any help in this area would be appreciated.

almost 9 years ago


Karl Foxley

This has been one of my best reads online for quite a while. I'm not a fan of the paid to tweet or sponsored tweet type services and your case study further illustrates why I will not use or promote them.

I would love to read any further case studies that validate Ted's claims.

Great post Patricio.



almost 9 years ago



I'm baffled by some of these comments. A case study on ONE sponsored tweet? KMart was sponsoring a sale so they chose a web entrepreneur to do so? THAT sounds more like a fail to me. Not to mention: a few shoppers at KMart interested in the sale could easily spend well over $250. Followers looking for adsense tips clearly wouldn't be interested in a sale at KMart. How irrelevant.

over 7 years ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.