The recent announcement about changing the emphasis of their ad targeting solutions to conversational keywords reveals that Twitter thinks it can unlock purchase intent from the masses of conversational data that flows through the platform: 500m tweets a day.

This technology seems to offer the ideal scenario that we could all go to Twitter, tweet what was on our mind, and the service would provide relevant and timely offers for us.

But there are still quite a few question marks.

Will keyword targeting be enough to uncover intent?

While it will certainly deliver more value to advertisers than targeting hashtags, true intention marketing requires far more than just keyword targeting, it requires a complex analysis of all the ways in which people express themselves in social media.

This includes looking at their profile information, looking at language they have used in the past. In order to really extract purchase intent from the data there is a huge piece of work around Natural Language Processing (NLP) that needs to be undertaken.

Context is king. Using the example of a consumer tweeting about a “latte” and receiving a promoted tweet about a friendly neighbourhood coffee shop – what about all the possibilities where the word “latte” might be used in a negative or complaining context rather than by a consumer desperate for their first caffeine hit of the day?  

Will an ad then be appropriate or welcome?

Is the promoted tweet the right vehicle when intent is established? As any good lead generation practitioner knows only too well, the level of intent can decay exponentially over time and the ability to deliver the message to the consumer at the right time is often more critical than the intent itself.

Promoted tweets often get lost in a user’s timeline or are only seen when the moment of intent has passed. In these circumstances direct contact from a vendor can be highly effective.

Finally: will the volumes allow Twitter to compete with Google?  

While Twitter has a huge number of active users (over 288m at last count), this represents only 16% of Twitter’s users. Most of those are passive and don’t tweet much. This means that at present there is no easy way to target the vast majority of users by intent.

There are two solutions here: 

  1. Working on search so that the service makes it easier to find conversational content around frequent problems (moving house, getting married, buying a car or a house).  

    That way natural language searches “how do I take my cable TV package with me?” become more likely to find relevant answers, including commercial ones, which would support Twitter’s ecosystem.

  2. Incentivising more people to become active users. The way to do this might in fact be through advertising; not just improved targeting but by creating an ecosystem where consumers can go to Twitter to tweet their needs with the expectation of getting a timely response with a great relevant offer from an advertiser.

    Of course it requires a tech solution to help power this ecosystem. Still, as Twitter starts to become known as a place to go for commercial needs (like Google is now) there will be a huge uplift in consumer use of social media to find products and services simply by asking questions or expressing desires.  

    Then the intention economy could really take off.