We ran Unquiet Data from July to December, 2014. In this time we drove multiple changes, optimising the Advertising Creative, Site Design and Content Films. This was about constant optimisation, so we sought a series of uplifts from data insights.

The changes that had the biggest impact on creative were these:

• Better Ads:

We used the best-performing creative assets to drive the brief for new assets across paid social & digital display, then we ensured the ads linked to a specific film, so people saw the content that had grabbed them in the first place.

This resulted in 5x Clicks & 2.4x view rate

• Better Site Experience:

We prompted changes on the site to prioritise the best-performing films, improve user journeys and to prompt another film viewing at the end of each film.

This drove a 28% view rate uplift & 11% uplift in multiple views (which result in 5x sales uplift)

• Better Stories:

By proving with data that the length of a video had no impact on its completion rate or the subscription rate, we helped the creative team (client & production) to argue against opinions on optimum time length (commonly 3 minutes, vs Unquiet Films running 4-7 minutes).

This allowed editors to give each film the optimum time to tell each story.

• Better Edits:

At the same time, we identified where video views dropped, which helped editors to introduce what editorial call the '3rd Paragraph'- ensuring that films had a moment early on that nails why it is important/interesting and, therefore, minimize drop-out rates.

The films that took this approach had a 4x improvement in viewing beyond 25% mark.

As a result of all these changes, view rate doubled from early July to late December. With a proven 3.4x uplift on subscriptions.

With our data united, we were able to build models and data queries that looked at the entire journey from first creative impression through to a final subscription. The key ‘journeys’ for the data were:

First, we proved that the creative content created a prospect pool with a higher likelihood to subscribe. Specifically, we identified that the percentage of people who take a membership after visiting The Times store increases by 3.4 times for people who have watched at least one of the Unquiet Film series, versus the overall store average.

Next, we ran an econometric analysis to understand the relative importance of different stages in the sales journey. Although the digital advertising campaign played a vital role in grabbing attention, this proved that the full-length content was the key – viewing the Unquiet Films themselves accounted for 82% of the subscriptions uplift.

However, the journey between exposure and subscription was frequently several days or more - as we’d expect for creative content which provokes thought more than it stimulates direct action. This gap meant that any conversion tag was a poor variable for optimisation – we needed a leading indicator to drive insight and for programmatic media optimisation.

This was solved with a predictive model for sales based on “View Rate” – the proportion of the audience who view a full-length film, which emerged as the leading predictor of subscriptions conversion uplift. With this as our “pivot point”, we had a short-term variable to drive…

We then started a cycle of “Identify”, “Test” & “Change”. In a series of two week “sprints”, we found data insights that suggested potential uplifts, then the team tested making changes to the creative – across digital advertising, the site design and the films themselves.