You might have noticed that we talk about reinvention a lot. It's on our website, in our marketing material, in the name of events we've hosted (Growth through Reinvention, in case you missed it). We also discuss it a lot amongst ourselves, because we're making it a central component of how we work with our clients.
Looking back at the growth of our business - and of the businesses that we partner with - we realised that working in a state of continual reinvention was at the heart of success. Sometimes, that reinvention came in the form of an overhauled brand identity, or a dramatic change of direction in marketing. Other times, it was something more focused: a problem solved with the introduction of a new approach to bidding in paid media, or with a change in organic search strategy.
Building a culture of persistent challenge
Reinvention does not need to be radical. It can be a gradual, subtle process. Woven tightly into the fabric of our daily work, it can be a culture of persistent challenge, of constantly rethinking how to do things in order to make them more effective. At its heart, it is a methodology for avoiding the curse that every agency-client team dreads: familiarity. Familiarity breeds comfort and complacency, and of relying on received wisdom for decision-making. It is the death of good work, and we do everything we can to avoid it.
The starting point for reinvention: Croud’s Impact Model
We're building a toolkit for reinvention to ensure that we're continually challenging our clients and proactively developing their marketing. At its heart is 4D, our planning cycle that covers research, strategy, planning and measurement. If there is a starting point in that closed loop it is what we call the Impact Model, a benchmarking methodology that quickly positions brands in their category alongside their competitors, and is a vital part of the discovery work we undertake at the beginning of any planning projects.
How it works
Competitive data is notoriously hard to get hold of, especially for digital channels. Our approach - using relative values for brand search and website traffic - gives us a quick to use, accessible, high frequency and near-global dataset to use to direct decision-making. By showing us a high-level view of the relative strengths of a brand’s purchase consideration (share of search) and acquisition efforts (share of traffic) we can begin to make directional recommendations about the roles for media.
Alongside that, we can suggest to clients the ‘enabling’ projects that should accompany it: data and analytics work, or strategic analysis, like segmentation, brand positioning or measurement projects. Regular tracking of the category shows us (alongside a fully-formed measurement framework) whether we’re making progress, and also allows us to contextualise our media performance in terms of gains or losses the competition might be making.
The model in action
We used this approach with our client Equifax, a major credit agency. Using the Impact Model, we showed them that their share of search had declined over time, and that they were lagging the rest of the market. A drop in relative brand search volume suggested a weakening brand, which we were starting to see reflected in our direct response media performance, as paid search CPCs had increased by 1.2x year-over-year, and we were having to spend progressively more for the same return. Against the backdrop of an overall drop in budget YoY, we made a case for additional spend in brand media and rebalanced their media mix, introducing BVOD as one of the ‘safest’ investments in terms of brand building. Overall, we saw great results:
- We diversified the media mix by 13%, helping to future proof their media strategy
- Brand demand increased by 40% in the regions which had BVoD live
- Site sessions were up 30% year on year in our exposed regions
- We saw +2.5 pt uplift on ad recall for people exposed to the ad, along with a 9.8 pt uplift in ad recall
Of course, the model is not without limitations, and we need to be careful to configure it as accurately as possible in order to get the most accurate results. When using it I’m reminded of George Box’s claim that “all models are wrong, but some are useful.” We like to think that it’s useful, and that it could be for you, too.
