How do you determine who was involved in a conversation?
Conversation targeting is powered by a human-tuned, rule-based keyword algorithm. These rules are tuned for precision and scale by carefully examining keywords for phrase or word matching, denylist matching, boolean expressions, author authentication (i.e. Nick Foles plays football, but not all his Tweets are included in the football conversation), time of Tweet validation, language matching, and more.
We consider a user to be engaged in a conversation if they have Tweeted or engaged with a Tweet that mentions a conversation topic, or dwelled on a Tweet about that conversation topic. There is a lookback window of 28 days.
Learn about our Twitter Ads policies and how to manage your privacy settings.
How does conversation targeting layer onto other targeting types?
Conversation targeting is an example of additive targeting, also known as "OR" targeting.
If you select 5 conversation topics and 35 follower @handles in your campaign, your campaign is eligible to serve to anyone who fits any part of that 40-piece targeting. For example, someone participating in Conversation 1 OR Conversation 2 OR Conversation 3 OR @handle 2 could be eligible to be served your ad. We call this additive targeting because the more pieces of targeting you add to your campaign, the larger the potential audience.
Conversation targeting will also take into account any subtractive, or "AND" targeting, you've chosen in your campaign. Gender, geo location, age, and language targeting are all "AND" targeting types, meaning that they will ALL be respected. For example, let's say you set the following targeting parameters in your campaign: Females in the United Kingdom who are 18+, with 3 conversation topics and 35 follower @handles. Only 18+ females in the UK who are participating in Conversation 1 OR Conversation 2 OR Conversation 3 OR @handle 20, etc, will be eligible to be served your ad.