Title	LINKREASER
Product/Service	CURATION TOOL FOR TWITTER USERS
Category	A01. Innovative Technology (incl. standalone solutions not in association with a brand or creative campaign)
Entrant Company	OGILVY PARIS, FRANCE
Advertising Agency	OGILVY PARIS, FRANCE

Name	Company	Position
Stéfan Galissié	Ogilvy Paris	Chief Data Officer
Adrien Leygues	Ogilvy Paris	Head of Technology
Pierre Lebret	Ogilvy Paris	Technical project manager
Fabien Chasseigneaux	Ogilvy Paris	Lead Front End Developer
David Raichman	Ogilvy Paris	Associate Creative Director

The Brief

The project originated based on two principal challenges that we noticed our clients facing, and that we, as consumers, faced: 1.More than half of followers on Twitter are fake or inactive, whereas having an engaged follower base on Twitter is very important for a company and even a consumer. 2.No tool existed that could offer a real curation system of Tweets focused on user’s interests. This gave us the idea to develop a type of direct marketing tool with a sophisticated scoring engine that has an ambition of developing a consumer or brand’s Twitter activity by interconnecting accounts that are similar to one another. The tool’s main objectives therefore were to grow an account’s Twitter audience with a more relevant audience while giving them access to more relevant content. The team consisted of members of our data, technology, and creative teams. It launched in France, but is available for any French or English speaking language. The total budget was around 100K euros, and the product is currently in a Beta public release.

IMPLEMENTATION The data from Twitter allowed us to create the brand and platform Linkreaser, a creative direct marketing interface for Twitter users. Twitter users arrive to Linkreaser.com and connect their Twitter account with the platform. They then enter the top ten keywords that are most interesting and relevant to them. Linkreaser uses these keywords to automatically find other Twitter users who Tweet about the Linkreaser user’s keywords, as well as Tweets that contain these keywords leveraging the sophisticated recommendation engine. Linkreaser then proposes a list of Twitter accounts for the the Linkreaser user to follow, as well as a list of curated Tweets that could be interesting to them. They have the option to let Linkreaser automatically follow these users and favorite these Tweets or to select them manually. Since Twitter’s data is open, we connected directly to the source via their API to collect the data. The first step consisted of creating a test database that allowed us to structure the data from Twitter, and apply text mining and natural language processing techniques in order to set ourselves up for statistical modeling. This test environment lasted about six months which allowed us to optimize our recommendation engine and our predictive algorithm. Once we felt that our recommendation engine was at a sophisticated level, we tested it on specific Twitter accounts during a pre-Beta period to understand how the tool reacted and to be in a constant optimization period. We then moved into a Beta testing period for about one year where we handpicked Beta testers and gave them free access to the tool to test it and provide feedback. After approximately one year of the Beta period, Linkreaser was open to the public officially in March 2015, launching as the first B2C product for an advertising agency in France.

OUTCOME The vision of Linkreaser is to change it from a free tool to a subscription tool where users must pay a monthly or yearly fee to have access to the service. There is also a B2B version that is custom made for the agency’s clients. Linkreaser has an average of a 75% follower increase in 3 months, and has 567 total members to date.