Twitter has recently opened up their self service advertising products to advertisers based in the UK, finally allowing us to delve into the system and start testing every element. One initial problem that we are experiencing is with the targeting capabilities of the platform, which at present is currently lagging far behind that of Facebook.
First, lets discuss the four main options available in terms of targeting on the platform, namely:
- Interests & Followers: Reach people with specific interests or who are similar to followers of specific accounts
- Keywords: Reach people that search, tweet about, or engage with specific keywords
- Television: Reach people who engage with specific television programming
- Tailored Audiences: Reach people using your own data
The interests & followers area is similar to Facebook’s ‘precise interest’ targeting that has proven so powerful when used correctly. However, one key difference is that the targeting works based on similar followers, with Twitter explaining it like this. “Targeting @usernames allows you to reach users with interests similar to followers of any of those accounts”. This brings a slight element of vagueness. On Facebook, for example, I can choose to target all of my competitors fans, whereas on Twitter I can only target a more vague concept of people with similar interests to people who follow my competitors. This will give a bigger reach, the only issue is we aren’t told what Twitter class as people with similar interests and therefore this all feels slightly outside of our control. In an age of exact targeting, advertisers demand to know exactly what is happening with their ads, and this concept may not be warmly received by many.
One area where Twitter has huge potential is in the Keyword targeting. Targeting by conversation content is a concept that no other mainstream platform has yet rolled out, and if used correctly has huge power. There are a lot of good options in Twitter’s platform around this, with great features such as:
- Negative sentiment filtering
- Keyword volume estimates
- Bulk keyword import
- Keyword match types
The only thing missing, in my opinion, from the keyword targeting section is the ability to choose timeframes. I would assume Twitter prioritise the serving of ads to people who have mentioned your targeted keywords most recently, but it’d be great to able to specify this. For example being able to target people who have discussed the phrase “Buy a Macbook” is powerful, but being able to just target those who have discussed it within the last 24 hours is even better.
The Television targeting segment is clever, and plays into Twitter’s strength as the leader in second screen conversation. Targeting conversation around a TV show before, during and after the show is a great option if you have the right kind of brand or appropriate content to be relevant within that conversation.
Tailored audiences hold great potential on Twitter. Targeting using tailored audiences is the equivalent of using custom audiences on Facebook. Whilst Facebook are making custom audiences more and more accessible with launches such as custom audiences from website visitors recently, Twitter’s tailored audiences are only available through third-party providers at present. That being said, if you can get access then including your current customers in your targeting set makes complete sense.
What would be great to see from the next iteration in Twitter’s advertising platform would be the ability to use all of these three options in conjunction with each other to make the targeting more and more advanced. For example, being able to target users who follow a certain celebrity or brand and have discussed a relevant product is better than either part in singularity. A real life example of this could be attempting to target marketing professionals interested in Twitter advertising. Targeting people who follow Social Media Explorer and other relevant industry blogs who have mentioned ‘Twitter Advertising’ is better than just targeting people interested in Social Media Explorer (as some may be un-engaged followers, or just interested in Facebook for instance), and people who talk about ‘Twitter Advertising’ alone may be general users discussing the ads they are seeing, not necessarily in a negative frame. Put the two targeting elements together and you have a much more focused targeting set. I personally use this kind of combined targeting on other platforms and it drives the best success rates and lowest CPA’s so we can’t wait to see this reach the Twitter platform.
In conclusion, the platform is a great first step for Twitter and is far better than Facebook’s initial self-serve platform. However, there is work to be done and areas where improvements could be made for the benefit of all advertisers. We’d love to hear what kind of targeting variants are proving most successful for you in the comments so please share!