My first foray into sponsored tweets. FAIL.

Featured

Ian Lurie Sep 20 2010

Note that I am using ad.ly as part of this test. I am not saying ad.ly is bad. They have a great site, a nice toolset and make no promises. I am suggesting (spoiler alert) that sponsored tweets rank in profitability somewhere below hiring alchemists to try to turn lead into gold.

Sometimes, I just gotta take one for the team. I’ve been skeptical about sponsored Tweets for a long time. Why on earth would anyone click an ad that says ‘Ad’ in it and zips by in a stream of 140-character messages?

But I wanted to see for myself. So I busted my piggy bank and took a shot. I figured, hey, maybe it’ll work!

Yeah. No.

The test

I chose my SEO training, the Fat Free Guide, as the test subject. And I used ad.ly as my advertising platform. I signed up for an ad.ly account and bought a single ad for $370.

Ad.ly lets you select a particular Twitter celebrity for your ad, then write the copy (as long as it’s not in the first person). Their site’s super easy-to-use, and I had everything set up in a few minutes.

I selected Shoemoney as the publisher. That meant he’d send out my ad to his entire Twitter audience. In internet marketing, he’s got the reputation and the pull, plus he could kick my fanny in an MMA match, I’m sure, so I figured, why not?

I paid. Ad.ly approved my ad and Shoemoney posted it within a few minutes:
tweet from shoemoney

I wrote the ad based on three things:

  • Most people who sign up for the Fat Free Guide search on my name at some point.
  • The $7/month lifetime price has been a great draw.

The results

To break even, I needed about 55 orders.

Right now, the Fat Free Guide site long on brains, short on looks. It has a hideous landing page:
Fat Free Guide landing page

But it does OK. 13% of people who read the ‘Join’ page sign up. Shoemoney has 112,000 followers on Twitter. Even if I my conversion rate plummeted to 3%, I only needed 1800 clicks. That’s a 1.6% clickthru rate.

That’s pretty high, but it assumed a big drop in conversion rate, so I figured my odds were good.

7 hours after Shoemoney sent out his Tweet, guess how many clicks I got? Just guess!

.07%

75 clicks

Wooooow that took me back to the good old days, when I created banners using CorelDraw and Comic Sans.

75 clicks?! Are you fracking kidding me?! I could’ve gotten my kids’ school to click the ad and gotten more than that.

Cough.

I’m OK. Not thrilled with the results, but OK. This was all in the name of science, after all.

Who sponsored tweets might help

There is an audience that could use sponsored tweets: I think political organizations, sports teams and folks in need of a fast media hit could capitalize on ad.ly and similar tools. Even if the tweets have [ad] inserted into them, organizations and politicians, etc. can still get their name out in front of a few more people for pennies per impression.

Sponsored tweets feel a lot like display ads to me: Potentially good for branding, but a conversion nightmare.

Oh, also: ad.ly has solved the Twitter monetization problem, though: Sell Twitter ads!

Yo, ad.ly

If I did anything wrong – if there’s something I should change to get a better result – tell me. I’ll do another test, on my dime.

The one thing you can’t tell me to do is change my landing page. I want to see better clickthru – at least .5% – before I change anything else.

tags : conversation marketingtools

related articles

14 Comments

  1. You know, it raises any interesting question – what’s the real relationship between followers and impressions? Shoemoney has 112,000 followers and, somewhere in our brains, we hear “112,000 impressions”. How many people are really listening, though? I’m not even talking about spammers or people who tried Twitter for a week. How many of his legitimate followers are paying attention to any given Tweet at any given moment? 1%? 0.1%? If only 1,120 people were listening, your CTR was almost 7%.
    I’m not picking on Shoemoney – I mean this in a very general sense. We as a community still have little or no idea how to effectively measure social media.

  2. Ian

    @dr_pete clearly they can’t all be quality followers. But even if only 5000 are good ones, you’d expect to see, I dunno. ONE SALE.
    But yes, you’re right on target – we’ll never be able to measure social media. You can’t measure basic communications, and that’s what this really is.

  3. Daman

    I wouldn’t be surprised if 30,000-35,000 of these are dead accounts.

  4. Juergen

    Nice test, very interesting. Thank you for sharing.
    Let’s say your ad has a CTR of 5%, IMHO that’s a reasonable rate, so 1500 people are actually reading your ad. Now estimated 25% are dead accounts (probably even more), still 84000 followers, if about 2% of them are really reading the tweets that’s your 1500 ad-readers.
    In general I guess that your CTR is higher than the percentage of the readers of most tweeters out there. Not the fault of the tweeter but it’s because of the followers. It’s easy to follow, it’s a bit more difficult to read…

  5. rishil

    You wis to retry the test (free) via my tweets? Would love to try it.

  6. That’s surprising, even for an obvious advertisement. I’d be interested in how the click-throughs might change if Shoemoney tweeted that directly, without the Ad.ly and (Ad) label.
    Jason Cohen mentioned yesterday in his blog the idea of valuable and engaged followers versus pure numbers ( http://blog.asmartbear.com/chris-brogan-blog-strategy.html ). I don’t follow Shoemoney on Twitter, and only occasionally read what he writes, but this makes you wonder just how small a percentage of those 112,000 followers are listening. There’s probably never going to be a perfect way to measure. We could start by monitoring CTRs in a person’s tweets, retweet rates, mentions, but these only paint a small picture as many on Twitter are “listeners” only.
    Either a small number are listening, or a large number are turned off by the Ad aspect. (Only way to find out is get rid of the Ad mentions.)
    Advertisements in the past have found their way onto our TVs and radio, our highways, our clothing, and our Internet browsing, but now they’re invading our (online) conversations.

  7. Sambo

    I would imagine that this may be due to the target audience.
    The guy likely has seo people following and those people may not want another guide to teach them how to suck eggs. (I may be wrong, I havent looked, no time).
    I would say that if you were selling Hair straighteners and could get onto Britney Spears (ridiculous I know) the CTR and ROI would be alot better.
    S

  8. Derek

    Some people already pointed this out, but just look at Shoemoney’s followers: http://twitter.com/shoemoney
    I went through 10+ pages of his followers and I am not sure if even 5% of them are legit accounts.
    On top of that, the ad that was written looks super spammy so I am sure that the handful of real followers Shoemoney has wouldn’t click on it even if they saw it. Also, to be somewhat effective, this thing would needed to be tweeted several times throughout the day. If it wasn’t then it will most likely get lost in all of the other tweets on a users stream.

  9. gareth

    I agree with Sambo, the followers of Shoemoney are probably
    – people who are part of his network
    – collegues
    – rivals
    – some bots
    I’d try with people in another “somewhat related” niche like
    – technology
    – small business
    – management
    I’m thinking about business people who want to do “The SEO” but can’t really understand this kind of activity. Your guide may help them understand much more this field of activity.

  10. Brad

    How many magazine ads, TV ads, PPC campaigns fail? Millions. Is it always the fault of the medium? Sometimes, but not always. Talking about one test without a control or more data reckless. You should try multiple times before drawing a conclusion like this.
    In addition to what’s already been said, my two cents…
    You already mentioned, but the landing page horrid…why would you EVER use that as a test?!?!?!.
    Have you tried that copy and landing page as a PPC ad w/ success? IMHO the copy of the tweet felt spammy, lacked credibility, and gave no promise/benefit to the consumer. I would have never clicked on such and obvious ad with no promise.
    Who is Ian Lurie? What the hell is “The Fat Free Guide to Internet Marketing?” What do I actually get for $7 month? What is this guy REALLY going to do with my CC#?
    Think why Marketing and SEO people use Twitter. To gather and share information and to network.
    Chum the water, throw a baited hook, then reel them in. Tease, offer a taste of what the site is about, THEN sell.
    Sample Copy
    “Practical SEO Tips to Drive Revenue”
    “10 SEO Tips You’ve Never Heard Before”
    “Advanced SEO: you know the basics, so does everyone else. Tactics to set yourself apart from the competition”
    “SEO tips to set you apart from the competition”
    “SEO made easy, tips everyone can use”
    Give them an article that delivers on what the tweet promised. Then offer “more” by joining for a small fee. Maybe even bait more with a free trial.
    Or if you prefer the direct sale method what’s in it for me? Include the value proposition. What do I actually get for $7 a month? ROI, # of articles, types of articles, access to SEO experts, etc.

  11. Wow man that’s rough. I used to get excited when somebody with 3-4 thousand followers would retweet something or share a link to something I had going on. Blah. Few clicks, that’s about it.
    Like you said, “Who sponsored tweets might help” has a lot to do with it too.

  12. Krista

    Hey Ian — first of all, thanks for giving Ad.ly a try, and we’re sorry you were dissatisfied with your test.
    Clearly there were a lot of factors / variables involved, as you note — creative copy & offer terms, landing page, high expectations on CTR, etc. — as discussed in your comments section.
    Here are some reflections and recommendations:
    1.) While we are still learning in real time (one year in), we are finding that B2C products and services that appeal to a broad audience — such as consumer electronics, media & entertainment, automotive, etc. — perform far better than B2B offerings that appeal to a targeted niche audience.
    Yes, there are a lot of SEO pros on Twitter, but how many of them follow Shoemoney; how many of them are going to actually see your offer; how many of them are going to respond well to the copy/call-to-action, etc. The numbers tunnel down quickly.
    2.) There are also a number of endemic-to-Twitter variables that we are learning. For instance, time-of-day can be a major factor, as can burying tweets (for instance, Shoemoney tweeted four times right after your ad, which no doubt hurt your click-through). 
    We work closely with advertisers to put these learnings to work on campaigns everyday, and we are working on ways to programmatically build them into our self-service system as well (including engagement performance guidelines).
    3.) We also have some key learnings on creative:
    — Write a great headline rather than sales copy. The message needs to grab attention, not try and sell straight away.  E.g. “Check out this cool new SEO guide with 5 killer secrets to crush it on Google” vs. “Get a Lifetime Subscription for $7/month.”
     
    — Discounts work best for well-known brands. E.g. Free $50 AdWords Coupon would probably work well, but the Fat Free SEO Guide is less well-known.
    Best,
    -Krista
    VP Marketing
    Ad.ly

  13. I think most of us Tweeps would be shocked to find out how many of our thousands of followers have stopped using twitter. I am feeling, at this moment, that twitter might be more of a fad then anything else. It is, after all, just a giant chat room. If Twitter monetizes the site (which looks like they are preparing for) more and more tweeps will be saying ‘adios, baby’ to twitter . IMHO

  14. Thanks for sharing this experiment, Ian. The comments were spot on. It all has to do with the level of engagement from the ad carrier’s followers. When you see someone with 100k followers, unless they are very high profile and the followers are fans – you can almost always assume that most of it is junk. A nice exception would be someone like @MayhemStudios Calvin Lee who spends 8 hours a day on Twitter. You can tell a lot from the number of RT’s your ad carrying candidate generated.
    @LindaSherman

Comments are closed.