Even a small website like this uses cookies to give you the best possible browsing experience. By continuing, you agree to the Cookie Policy.

Trop50 Resolution Rescue

Tweets to #ResolutionRescue are dynamically generated into personalised motivational videos

Project details


Raise awareness of the new Trop50 orange juice during winter when orange juice sales are at their lowest


Generate a personalised video encouraging people to stick to their New Year's Resolutions


A custom workflow with moderation, video generation/upload and reply tweeting steps


  • #ResolutionRescue


Pepsico Tropicana


Agency Republic
Senior Creative Technologist


  • United Kingdom


  1. Morning commuters
  2. Twitter users


Launch: 01 January 2014
Retired: 31 January 2014

Key technologies

  • PHP
  • CodeIgniter
  • Adobe After Effects
  • AWS Simple Queue Service


Once Christmas is behind us for another year, it can feel like we're swamped with ads encouraging us to eat this/drink to maintain a healthy lifestyle and rectify any naughty behaviour. So instead of telling you what to do, we decided to help people stick to their own personal New Year's resolutions: whether they were to eat healthier, exercise more or make it through the month without drinking.

To drive traffic to the campaign, sponsored tweets and an integrated campaign (published in London's Metro newspaper) directed people to tweet their New Year's resolution with the hashtag #ResolutionRescue.

User Journey

The user's interaction was simple: tweet us if you're struggling with your New Year's resolution and we will help boost your willpower with a personalised video.


Each tweet went through the following steps to generate a response:

  • Users tweet with the hashtag #ResolutionRescue, which is picked up via the Twitter Streaming API.
  • The tweet enters a moderation queue where it is filtered for profanity or use of trademarks. If approved, it is also classified into one of twelve categories (such as cake, chocolate, etc.).
  • A video is dynamically generated including the users' original tweet, their profile picture, a motivational message and a short animation encouraging the user to stick to their New Year's Resolution.
  • The video is uploaded to YouTube and an automated tweet is sent to the user with the video embedded into the tweet.


Each video animations was customised to contain the original tweet text, the user's profile picture and the user's twitter username - making it personal to them.

The end-to-end process from tweet to response was between 3-5 minutes. Generation times varied depending on the moderation queue and the template response video (as some videos took longer to render). A cluster of 15 servers hosted by Amazon Web Services picked up a job from the SQS queue and generated the personalised videos.

We pre-recorded a variety of animations to cover all possible New Year's resolutions (and some generic ones just in case).

  1. Cake

  2. Chocolate

  3. Couch

  4. Dancer

  5. Drinking

  6. Going out

  7. Magazine

  8. Mountain

  9. Muffin

  10. Trainers

  11. Winner

  12. Yoga


These sample tweets show the end-to-end process of the advertisement, the user's tweet and the personalised video response.


  • Bashkim Isai
    Senior Creative Technologist
  • Al Campbell
    Executive Creative Director
  • James Hodson
  • Jason Keet
    Art Director
  • Tom Cole
    Head of Technology
  • Ian Culshaw
  • Iryna Svitlychenko
    Quality Assurance
  • Victoria Copeman
    Quality Assurance
  • Alistair Millen
  • Stephen Armstrong
    Business Director
  • Richard Templeman
    Head of Project Management
  • Michael Robinson
    Head of Design