For a long time working with custom audiences meant that you would need to go through an approved advertising partner.
While this has its uses it was often prohibitively expensive, leaving many small and mid-size businesses unable to use this kind of targeting.
Recently Twitter has changed this, allowing self-serve advertisers to create and upload custom audiences.
I thought it would be useful to show you exactly how to do this.
Before we begin, I should say that it’s worth thinking about the kind of sentiment this type of campaign may generate.
Is there a good reason why these users are following your competitors instead of you? Are they going to be annoyed by your sponsored tweets? Have a clear goal before you get started and make sure you are providing value rather than disrupting people’s experience.
Right, enough moral outrage, let’s get on with it shall we?
Step one: Prepare your audience
@Econsultancy has a fairly large audience, as do many of our target sectors, so here I’m going to save myself some leg-work and use Moz’s FollowerWonk tool to extract some user lists quickly, but if you are dealing with smaller audiences (Say, up to around 5,000 users) then there’s no reason you can’t do this step manually with a bit of patience/an intern who doesn’t mind pressing CTRL+C for an hour.
Thanks to Twitter’s Lookalike audience targeting, you can begin by using your own followers and Twitter will attempt to find users with similar interests to promote your ads to. However, if you’re feeling a bit sneaky, you can also use your competitors followers.
For this example, I’m going to extract a list of people following my personal account.
Log into Followerwonk and select ‘Analyse Followers’.
Type in the account name you want to examine and choose ‘analyse their followers’.
Off to the races…
It’s worth noting that Followerwonk will balk at very large audiences, but it will give you a selection of up to 100,000 users, which should be plenty to get you started. If you have a larger dataset, then you should really be talking to an ad partner.
FW will present you with a short report showing things like location, most active times, social authority and more for my followers (Hello guys!).
I’ve redacted their screen names etc to protect the innocent, but you’ll get a list like this:
Next, dump it to a CSV.
If this is a competitor, it’s likely that a few of them will also follow you. Here you can see that some of my followers also follow @Econsultancy.
You’ll need to get rid of these to avoid wasting your ad budget, so do a quick sort and delete:
Twitter allows you to upload a variety of data sources to custom audiences, including Twitter IDs and emails, but for simplicity’s sake, here we’ll just use their @names.
Delete all of the other information except ‘screen name’ from your CSV and save the file.
Step two: Upload your audience
Next, open up Twitter, and open your advertising account. If you don’t have one, I’ve previously written a step-by-step guide that you can use to open one.
Choose ‘Tools’ from the drop-down menu and select ‘Audience Manager’ from the drop-down menu. Then hit the ‘Create New List Audience’ button.
Give your new audience an obvious name, and add a date in case you need to update the list later.
Choose the type of data you are using, and upload.
Depending on the size of your list, it will usually take a couple of hours for Twitter to process your audience information. It will fire you an email once your audience is uploaded and ready to use.
Step three: Create a campaign
Once your audience is ready you’re all set. Fire up Twitter and head back to the ads section.
Create a new campaign – here I’m using ‘promoted tweets’ – and again, name it clearly:
Now choose ‘Tailored audiences’ and upload your list.
You can refine this further by location, gender, age etc:
Finally, choose which tweets you want to promote (You can also choose ‘compose tweets’ if you’d like it to only be seen by targeted followers, rather than your entire Twitter audience, useful if you are running a get followers campaign, and for A/B testing content in advance), and set your budget.
You’ll notice on the left, Twitter displays a small graph showing how much reach you can expect for your budget.
Here the default maximum payment per engagement is $2.00, but you should test this extensively.
I’ve found that I can usually get this down to around the 50-75 cent mark with smaller audiences and niche content, but this will obviously differ based on your aims (or to put it another way, if you are a broadcaster, you’ll need some ice for your company credit card at some point).
And away you go. Stats and spend will be displayed in the analytics section. I recommend you update your campaign every day with new content for maximum engagement.