WE’VE HELPED THE CHARITY TURBOCHARGE DONATIONS AND STEP UP THE FIGHT AGAINST CANCER - OUR ANALYTICS LED TO ELEVEN MAJOR CHANGES TO THE WAY WE, THE CREATIVE AGENCY AND THE CLIENT PLAN FOR RACE FOR LIFE.
What’s working hardest for us?
Cancer Research UK has been holding Race for Life events since 1994. Every event is promoted by a hugely diverse mixture of media – from national TV campaigns to localised ad-hoc promotions. So a key challenge for the marketing team is to understand the impact of each activity on race entries. This is made more difficult by
interdependencies between different types of marketing. For example, there is a relationship between online search and traditional media such as TV, radio and press.
HOW WE RESPONDED
We modelled the true contribution of marketing channels.
We created a series of nested, segmented entry models, splitting entries per race down to a more granular level. And we set ourselves up so we could measure media interaction effects using tracked data for each of the media. We split the tracked data for each of the media. We split the media into those playing an ‘assist’ role and those with an ‘attraction’ role. In total, we created a series of 23 models, using a pooled modelling approach. Race entries were classified according to the type of race location (city, suburban, rural) and the level of local support received (high, medium and low, since no actual monetary value could be attached to such activity). These were modelled against the various marketing factors, together with the calculated underlying natural build in race entries (calculated by looking at the pattern of ‘base’ race entries in terms of their distance from the actual race date).
Within these models, the tracked response data from certain media became a model in itself. For example, paid search drove entries, but paid search responses were themselves driven by other media such as TV, press and radio. Even with so much marketing noise around the events we could directly pinpoint the impact of
over 90% of the marketing activity taking place. We successfully measured the remaining 10% of activity (i.e. the unobservable local support activity) by comparing the base level of entries generated from the different segmented race models. For example, we could compare the value of the base, per race, in a ‘city’ race that
receives high levels of support versus a ‘city’ race that receives low levels of support.
Eleven major changes to the way we, the client and the creative agency plan the activity.
The analysis prompted 11 major changes in how Cancer Research plans for the race, including:
- how the client budgets - we found we could effectively upweight the budget by 20% to maximise turn out and donation
- distribution of budget – we’ve moved support from lower performing races to stronger performing ones
- seasonality of spend - we now start activity much later in the year
- creative messaging – we’ve changed the outdoor creative to make the medium more effective
- media choices - we substitute less efficient media channels for more efficient ones (eg upweighting regional radio and TV)
Analytics now sits at the very core of Race For Life planning, with the forecasts generated from the models used to scenario plan against targets, as well as being used to monitor performance on a weekly basis.