Twitter Brings Keyword Targeting to Social


Twitter has announced a significant enhancement to its Promoted Tweets. “Keyword targeting in timeline” allows brands to show ads to users who have tweeted or interacted with a tweet that contains specific words and phrases. This new functionality gives advertisers a paid search style model within Twitter similar to Google’s.


The process works similar to paid search. Advertisers create a keyword list, set a bid, and when a user tweets or interacts with a tweet that uses words that match something in the keyword list, they are served an ad. Unlike search, the user is not served the ad immediately; instead, it will show up in the user’s timeline within the next several minutes. For example, Bob sends a tweet about looking for a new computer. If Dell has the words “new computer” in its keyword list, the next time Bob refreshes his timeline he gets served a Promoted Tweet from Dell. Further, if Bob does not send a tweet, but instead interacts (replies, favorites or retweets) with someone else’s tweet that references a “new computer,” he could also be served a Dell Promoted Tweet.

Twitter will launch keyword targeting with two types of keyword matching:

  • Phrase Targeting – A series of words in the exact order, with no words in between.
  • Unordered Match – A series of words that appear in any order, with words in between.

These targeting options are not as robust as Google or Bing, but they are a good first step. Twitter will offer some keyword tools at launch to help marketers with word tenses and plurals. Also, bulk keyword management is supported immediately. Users of Buddy Media and SHIFT will see the changes incorporated into both technologies within the week. The biggest issue with match types is the lack of negative keywords. In SEM, negative keywords keep brands from wasting spend on irrelevant queries or being associated with unfavorable keywords. In social, there is an additional use: sentiment. Twitter will use an algorithm to help identify negative sentiment within a tweet, but language interpretation can be extremely hard for an algorithm, particularly when hashtags are involved. It is easy to see how an algorithm could miss the nuance of a tweet such as: “Can’t wait to eat at McDonald’s again! #notreally #neveragain.” The addition of negative keywords should help to ease these concerns, but there may never be a 100% accurate sentiment solution.


Keyword targeting in timeline is a smart move by Twitter that will further enhance the appeal of Promoted Tweets. Performance remains to be seen, but, in theory, keyword targeting will be a large boost to the efficiency of Promoted Tweets. Twitter is now a “pull marketing” platform where the user first expresses intent. Maybe most exciting, marketers can still make use of three targeting options – location, device and gender – in conjunction with keyword targeting, this gives a great level of relevance. The enhancement also allows for scale and speed of what was once a very manual process – responding directly to Tweets. It will allow marketers to have larger keyword lists and to serve ads to users within minutes of their tweet.


Even with concerns about match types taken into account, keyword targeting in timeline is a win for marketers. Keyword targeting in timeline will allow advertisers to quickly respond to users whilst the interest is still there, and will prove especially valuable in when used in conjunction with Twitter’s other targeting tools.