Leave a comment
Get the GH Bookmarklet

Ask GH

I recently got a question on Twitter asking why I was using Twitter Ads when Outbrain offers better engagement for a cheaper price. I think the short answer is that a channel is only as good as the person working it. And I wasn’t able to crack Outbrain to get it working for me. On the other hand, Twitter Ads are working really well for me.

Twitter offers the benefit of potentially having a Tweet go viral. It’s particularly good for GrowthHackers.com since our audience overwhelmingly shares on Twitter rather than Facebook (see these survey results: http://kiss.ly/1gyhXgt ). The first time I used Twitter for ads, my ad did go a bit viral. Unfortunately I had a high daily budget so paid for every click.

Lesson 1: So my first lesson was to have a low daily budget. $25 seems to be a good budget because it gives the sponsored Tweet enough opportunity to be seeded for the day but the spend is relatively low.

My second lesson came because I unselfishly shared this little hack on GrowthHackers in this discussion: (http://growthhackers.com/twitter-vs-facebook-how-top-social-networks-rank-on-ad-performance-wordstream/#comments ). In the spirit of reciprocal giving, someone shared a little hack back with me.

Lesson 2: They suggested that I include an image with my sponsored Tweet. I did this and saw the engagement rate go up quite a bit on my sponsored Tweets. So lesson #2 is to include an image with your sponsored tweets.

Lesson 3: My final lesson was to include the link in the middle of the tweet rather than at the end. I learned it from this post on the Hubspot blog (see point 21): http://blog.hubspot.com/marketing/create-perfect-social-media-posts-slideshare .

Here’s my sponsored Tweet that generated 1049 clicks today for $25: http://pic.twitter.com/BBNfo8Eskv . While today was particularly good at 2 cents per click, the overall campaign has still averaged less than 4 cents per click.

  • SE

    Sean Ellis

    about 3 years ago #

    UPDATE: I moved this "comment" up to the question details so it wouldn't get lost in the comments.

  • MW

    Matt Wheeler

    about 3 years ago #

    Great results! I like the desktop trick.

    Were the 'clicks' the ones in the Twitter ad interface or actual clicks on the URL? Clicks on the picture also get counted towards what Twitter reports. We found adding pictures decreases actual URL CPC in some cases (blog posts) but increases it in others. (direct response landing pages). I'm sure the picture and tweet content heavily influence the effectiveness - https://www.driftrock.com/blog/should-advertisers-use-pictures-twitter-campaigns

    • SE

      Sean Ellis

      about 3 years ago #

      I've had problems in the past with tracking less clicks than Twitter is reporting, so thanks for heads up on this potential cause. Yesterday I tracked well in excess of what they were reporting though.

  • SE

    Sean Ellis

    about 3 years ago #

    The longer I run this, the cheaper it's getting. Today I got 2500 clicks for the $25.

    • GD

      Guillaume Dumortier

      about 3 years ago #

      If you keep on tweaking, updating and storytelling this ad, you could run it for a year.
      I think the 'storytelling' factor of having this tweet out everyday make it more appealing in the end and you're almost 'forced' to click at some point to discover the intriguing content. It would be interesting to know how many times they saw a "$25 Twitter ad with 1000+ clicks" promoted tweet before clicking it.

      At the end of the day, it's brilliant to have an evolving promoted tweet with an evolving story as it goes.

  • FM

    Francois Mathieu

    about 3 years ago #

    This channel is relatively new and cheap, I wonder how long it's gonna take until there is much more demand and the prices go up. One thing for sure, there are many tools that are freely available like the Twitter cards that should be taken advantage of. Why asking people to go get lost on your website when you can get (free) email subscriptions right in the Twitter feed. If you don't have your Twitter Ads accound, I highly suggest you go get it now: https://ads.twitter.com.

  • SM

    Stuart McKeown

    about 3 years ago #

    I think this example also goes to show the power of including good images in Twitter posts now that they show them in the feed. I've seen some really creative uses of these lately & it is skyrocketing engagement.

  • CC

    Chris Conrey

    about 3 years ago #

    So let me ask the question that bounces around my head whenever someone posts a "I got X clicks/likes/views/whatever" is that the extent of the conversion you are aiming for? Or was there a goal of more users at GH? Signups? Or purely eyeballs?

    • SE

      Sean Ellis

      about 3 years ago #

      Hey Chris, My goal is to drive visits that act like organic traffic when they hit the site. As a network effect business, I'm still working to prime the pump on the community. While I don't want to waste the money, my total daily spend is only $25. I'll continue to be "full court press" on priming the community pump until organic growth reaches a level where it is no longer necessary.

  • RF

    Ryan Fujiu

    about 3 years ago #

    Very impressive. What was your max bid? Just wondering if it was a bunch of clicks at a very low bid, or a few key retweets which got the ball rolling.

    • SE

      Sean Ellis

      about 3 years ago #

      Bid was $1.02. Actual cost per click has been tracking about 2 cents though.

  • A.

    Also .

    about 3 years ago #

    Sean, thanks for sharing .. hehe :-)

    What was your bid value? I've just made a quick attempt to set up a campaign for our product and the suggested bid was $1.50. That seems rather high in the context of your post.

    • SE

      Sean Ellis

      about 3 years ago #

      My bid was $1.02. But because my budget was low, they can't stop the traffic from the retweets. My engagement rate today is over 10%. So despite by $1.02 bid, the clicks are coming in at less than 2 cents. My strategy is to "seed" the tweet every day with a $25 budget and then hope that it gets some good retweets. So far it's working.

      But even if you don't get a ton of retweets, a high click rate can still push it down your cost per click to twenty something cents. Hope this helps.

      • A.

        Also .

        about 3 years ago #

        Makes sense. Thanks. I will let the campaign run for a couple of days and perhaps post the results.

        • SE

          Sean Ellis

          about 3 years ago #

          It seems like it is really dependent on being able to get some RTs. Interestingly my current campaign had 0 RTs first day, 16 RTs 2nd day, 7 RTs 3rd day, 13 RTs 4th day, and 15 RTs so far today. So don't give up if it doesn't work first day. Obviously followers on the RTs makes a big difference too. But that's the reason why I'm doing it was a daily budget of $25. You still need to get lucking with the right RTs. Obviously quality of the Tweet and content makes a big difference. Good luck!

      • MW

        Matt Wheeler

        about 3 years ago #

        Not sure if I'm reading this right but you don't need to limit cost from RTs with budget. Twitter only charges you for the first engagement. For example, if someone RTs a promoted tweet then you don't pay for any interactions of that tweet from the followers of that person. Effectively a RT makes it a standard tweet after that. This is probably why you are seeing more clicks than twitter is reporting because it's not charging you for subsequent viral interactions. Check out the timeline activity report. You should see all interactions, not just the paid ones so you can compare the 2.

        The seeding tactic is still great though for viral content. We have seeded tweets for $100 that then become viral as the tweet gets more RTs than we pay for.

        Il write up a detailed post on how the Twitter auction works and post it to GH if you like?

        • SE

          Sean Ellis

          about 3 years ago #

          You're probably right that this is how it technically works. But the reporting calculates all the clicks. Here's how the tool tip states it: "The total includes billed engagements as well as additional engagements you aren't billed for across Retweets, replies, follows, card engagements and other clicks." I guess the issue is that it's better to seed with $25/day than $500 one time, since $25/day gives you a daily opportunity to rekindle virality. One of the reasons that I think results keep improving is that people see 190 retweets and 140 favorites. This social proof makes them feel that it's OK to retweet it. First day I received no retweets, but today alone I received 27 retweets for the same $25 spent.

          • MW

            Matt Wheeler

            about 3 years ago #

            That's some awesome stats. Great point about the extra RTs from social proof. Most tweets die out, but I guess if the volume is there and the engagement is high then it keeps going! I'm going to give it a try too with a blog post.

  • HA

    Hana Abaza

    about 3 years ago #

    Thanks for this Sean, just started playing around with promoted tweets and this is super helpful. Have you narrowed down the targeting at all, or keeping it pretty broad?

  • EP

    Eric Paine

    about 3 years ago #

    Great read. I'd love to know how many RTs came from current followers vs. non-followers. I'm not sure someone with only ~400 followers (me) could achieve the same results, but I'm going to try nonetheless.

    • SE

      Sean Ellis

      about 3 years ago #

      Good point Eric. I do include my own followers in the targeting for this very reason. My followers are a lot more likely to retweet than someone who doesn't know me. In fact, that's partly how I stumbled on this way of using Twitter ads. This friend https://twitter.com/TammyCamp tweeted one of my early sponsored tweets and shot my costs up. I thought it was ridiculous that I had to pay for clicks on her retweet to 150,000+ people. Then I thought that I'll just lower my budget to $20/day and they can't charge me when this happens. I recently moved it to $25 to give it a little more seeding.

  • KD

    Kristin Dziadul

    about 3 years ago #

    Hi Sean,

    I saw you already noticed my post, but I really felt inspired by your suggestions here that I had to try this test myself! Definitely saw a huge spike in performance after less than a day.

    I would love to hear your thoughts, though, on how you think the presence of an image and the increase in clicks is actually beneficial to the bottom line goals of a company running a promoted tweet. As I mentioned in my post (http://kdmedianow.com/spent-100-twitter/), Twitter includes the clicks to expand a tweet to view an image in the total Clicks, which can be misleading to some. For those aiming for conversions, do you think this strategy would be misleading to them? How would you recommend they approach that?

    • SE

      Sean Ellis

      about 3 years ago #

      Hi Kristin,

      Thanks for sharing your experience! I agree that image clicks don't provide any real value. Ultimately my goal is to get the GrowthHackers.com community to be self propelled. Like any "network effect" business, this requires a full court press effort until the pump is primed enough for the community to take over. Since launching on September 30th, I have worked every day to make this happen and rarely go a 6 hour period without checking in on the community.

      I'm anxious to "tip" the community ASAP so I can take a day off. Considering my time has a value (several people have paid me $900/hour for consulting on Clarity.fm), it makes sense to consider spending money to accelerate toward this tipping point.

      We did some research and it turns out the majority of growth hackers share and find articles on Twitter. From day one, I've used my Twitter following to start seeding the community. In October I began experimenting with Twitter ads to help prime the pump on growth. I spent $2058 from Oct 17 - Oct 23 and tracked 3724 engagements in Twitter's reporting for a eCPE of $.54. Since I've run the Upworthy campaign for 11 days, I'll include a full 11 day period for tracking the total visits that we received from Twitter. From Oct 17 - Oct 27 I tracked a total 4,236 visits from Twitter using Google Analytics for an average cost per visit of $.485. Average length of visit from Twitter was 2:24. I felt like this was a waste of money, so I paused the campaign on Oct 23rd to rethink it.

      Then I did a little experimenting and studying about Twitter ads (as outlined above).

      On December 5th, I started the Upworthy campaign applying all the learning. Since that time I have spent $367 on Twitter ads ($300 directly on the campaign and the rest on little experiments). During that time, my twitter visits have been 12,298 with average length of visit 2:40. So the average cost per visit is $.029.

      At any given point when I check realtime Google Analytics for the site, we generally have between 10 and 50 people engaging with the site. We still have a lot of optimization that we need to do to maximize engagement from this traffic as described here: http://conversionxl.com/optimization-experts-share-their-favorite-google-analytics-reports/

      Hopefully this clarifies why I consider this a good use of money for me.

      I'll post this on your article as well. Thanks!

      • KD

        Kristin Dziadul

        about 3 years ago #

        Sean,

        That definitely clarifies -- and wow! Definitely seems worth the spend to optimize the way you did since you realized such an improvement in results.

        You've definitely given me hope in Twitter ads, as one of the first campaigns I ran didn't have great results so I was discouraged about the platform as a whole (because Google Adwords had great results in the same time period).

        Very different channels, though, and I'm looking forward to Twitter further differentiating itself so it can fully be its own unique PPC channel.

        • SE

          Sean Ellis

          about 3 years ago #

          It definitely needs experimentation. Good news is that you can experiment for very little money. I've written a couple ads there duds since, but it didn't cost me much to figure it out. One shortcut, if I don't get retweets from my own free account, it probably won't work as a paid campaign.

          • KD

            Kristin Dziadul

            about 3 years ago #

            Yeah that is the benefit of a lot of PPC channels (Adwords included). You can test for messaging, early success, etc. with just a few hundred dollars.

            Great point about if it doesn't obtain engagement organically, it may not when promoted. I'll keep that in mind!

  • MW

    Mitch Wainer

    about 3 years ago #

    Sean - Thanks so much for sharing. This is hugely helpful. We will be implementing lesson #2 and #3 immediately.

  • DA

    David Adeyalo

    about 3 years ago #

    I love how the community here is willing to share their "secrets" in order to help one another. This is so "rare" in the startup community, and I'd love to see more of this.

Join over 70,000 growth pros from companies like Uber, Pinterest & Twitter

Get Weekly Top Posts
High five! You’re in.
SHARE
33
33