Creative personalisation in a privacy-first world
In the midst of significant industry change, it's increasingly important to understand how the technologies that underpin creative personalisation will be affected. Data strategies should consider what can be done to move to a model in which ‘creative’ is leveraged as an optimisation lever in and of itself.
The brand should own the DCO integration
Since its inception, dynamic creative optimisation (DCO) has been in flux. As we realised the power of onboarding new and different data sets outside of those available in the ad server and DSP, the integration that allowed us to change our creative ‘on-the-fly’ was born again as vendors offering deeper personalisation in previously inaccessible channels. Moreover, simple interfaces meant DCO was no longer the reserve of specialist teams who understood how to code. As we move towards online ‘data privacy budgets’ and deprecated online IDs, data centralization and democratization will be key to leaning into these changes. Vendors will be restricted, not only in the data they can collect and pass, but also in the amount. Furthermore, personalising within a vendor-owned silo leaves little room for strategic data enrichment across a business. The most effective way to achieve this ownership is to move the DCO integration back into the brand-owned stack via the ad server or DSP.
‘Creative’ as a data point that transcends the gardens
There is a lot of conversation about moving toward contextual indicators within our data-driven creative strategies. These indicators remain important and effective but on their own do not provide the same level of ‘one-to-one’ addressability that we’re used to.
Machine learning offers additional insight through the quantification of creative detail. Using natural language processing we can understand how the syntax and sentiment of our language is perceived. Computer vision offers an opportunity to pick our creative apart by generating creative metadata (colour codes, asset positioning, facial recognition, etc). Coupling the output of these technologies with a brand’s owned & ads data enables its enrichment and unlocks brand-specific data science in the form of creative scoring and propensity modeling. Linking this research back to creative ideation will see brands raise their creative baseline, starting from a place of research as opposed to relying on in-platform optimisation cycles to tell them what works.
We want more fusion between our creative and media teams, not less
Among various uncertainties, it holds true that good creative messaging is one of our most effective tools, and can be leveraged and optimized regardless of the channel or platform we’re speaking to customers in. As data strategies evolve, we should consider how we think about DCO as it pertains to data-led creative and, ultimately, personalisation at scale, in order to circumvent the key pitfall of working on DCO within a creative silo.
Aligning DCO to creative ideation will eventually limit our headroom for innovation by restricting how easily we’re able to work with data partners, and existing brand or partner innovations. As a technical integration that is used to generate data, including it within our overarching data strategies allows us to break down silos, making for more successful marketing in the long run. For example, centralising creative analytics data and overlaying it with brand and partner assets allows us to generate scores for creative permutations. In time, this unlocks the ‘pre-qualification’ of creative based on a brand-specific grading system. This approach creates fusion between agency partners as learnings can be fed back into the ideation process in order to help fuel it.