We were pleased to have early access (and a free voucher!) to the small(er) business offering of Twitter Ads. Previously, you had to spend at least $5,000 per month to run any ads / promotions on Twitter, so it was beyond the reach of most people’s budgets.
A number of people have asked whether we could write up a quick review of the experience with some screenshots of what the platform looks like.
We’ve experimented with it with mixed success. Here I’m using the example of a recent webinar we ran, where it was quite successful – other promotions we’ve run have had far less impact/conversion/ROI.
Here’s what you’re presented with after you log in and set up an account (NB click images for larger version):
So the biggest challenge is that unlike Facebook, you can’t geotarget users beyond ‘Australia’. So if you’re a local / state based brand, Twitter Ads still has a long way to go before it’s useful. The only choice you have (for any targeting) is location. In the US you can target to states and ‘metro areas’ so I’m sure that it will (eventually) flow onto us.
After this, you can choose to either promote your Twitter Account or promote Tweets. This is pretty similar to Facebook’s ability to promote your Page vs Sponsored Stories. The tool currently doesn’t allow you to promote Trends.
Here are the 3 types of promotions and what they look like to users using Twitter Web:
Promoted Trend (you can’t do these using the Twitter Ads tool currently)
Promoted Tweet (this example was in a search result for #summerofcricket above)
As you can see, they’re labelled as promoted to users.
Promoted accounts look like this when you set them up:
As you can see, you set a per day budget, and a per-follower budget. Assumedly this operates on a bid-per-follower basis, where the highest ‘bidders’ will be shown to more users and more often (and pay more for the priviledge). Twitter recommends a maximum bid of $2.50, but we were still getting new ‘follows’ at $1.30. The ‘you can expect’ line I’ve found to be wildly useless and not a good indication at all.
When you have run a campaign, you can see your Twitter Follower growth over time, and it splits the graph into those sourced in a paid context and an unpaid context (NB: we weren’t running a Promoted Account campaign at the time so this is just our organic growth). You also see the number of impressions (people who see the promoted account) along with the follow rate, number of followers and the amount you’ve spent.
Certainly running Promoted Accounts increased our twitter followers, but given that it was vastly untargeted, we haven’t been using them as we have quite a specific target audience. From the looks of who followed us though, there does seem to be some trends toward advertising to people who follow people who follow you – we got a lot of PR/marketing agency types following us.
We ran a small (~$100) campaign to promote our webinar that we recently ran on social media strategy. So the results below are from the 3-4 days that the promoted tweets were running.
As you can see, we opted to manually select our tweets rather than have Twitter do it automatically, as we wanted to specifically promote the webinar. Additionally, the tweets we promoted were both a ‘naked’ one along with several from popular Australian hashtags including #socadl, #npau, #gov2au and #hcsmanz.
Similar to Promoted Accounts, you set a spend rate and cost per click. Recommended maximum bid is at least $2.60 above but we got fine enough results for $1.50 (imho). Again, the expected clicks were wildly inaccurate, with the campaign running for 3-4 days (not spending the full daily budget) and costing about $1.14 a click, with about 25 per day. Had we used hashtags with more people following them, or timed them to occur around the tweetchats that occur – I’m sure we could have increased our clicks. But our click through rate was pretty high – 1.24% isn’t too bad at all and is better than a campaign we ran on Google AdWords for the same thing (we experimented with $70 and got a CTR of about 0.6% – but this wasn’t exactly optimised).
You can also see the detailed stats for the tweets:
Beside anything else, it was certainly interesting to see which hastagged tweets got the most impressions: #hcsmanz and #gov2au had far more impressions than #socadl and #npau – this could become a good way to measure the number of people viewing and/or visiting hashtags. Also the platfom allows you to experiment with a few different ad variations and revise them as you go, which is always useful.
The end effect
We got a good conversion from the above to webinar registrations (so it met our outcomes) and also increased our number of followers quite a bit (it won’t tell you how many of those followers were via a Promoted Tweet, so it could’ve been other sources) throughout the few days, so it certainly worked well.
We’ll keep you updated about the latest, and provide any updates if we experiment with other things in the future!