As the capabilities exist for marketers to target audiences one-to-one, there are certain roadblocks that limit the ability to grow these strategies worldwide. Writing exclusively for ExchangeWire, Keith Petri, CSO, Screen6, explains the importance of a probabilistic data approach in this new world of GDPR and future regulations, in order to achieve highly accurate results, while still addressing consumer privacy concerns.
There are dual forces at work in the world of digital marketing. One is the growing capability to make one-to-one marketing a reality. Audiences have become more specific and ads can be tailored to an individual person on a specific device. On the other hand, the world of marketing has become global, sometimes borderless, and brands need to expand to new regions to find new consumers. Unfortunately, current market limitations of many probabilistic providers, as well as deterministic datasets, mean that marketers cannot expand their one-to-one marketing strategies abroad.
For brands and modern marketing professionals, accurately understanding their customers, their behaviour, and their needs across regions is paramount. As the amount of time spent on mobile continues to increase in both established and emerging markets, and as marketers think about how they are going to allocate budget, not just across platforms but across regions, cross-device identification has been put front and centre.
As Allison Schiff of AdExchanger points out, the blend of both a probabilistic and deterministic process to achieve this “understanding of the consumer” will lead to the best possible outcome when managing and measuring user data. While deterministic associations can be seen as a more accurate but finite data set, probabilistic data can help provide scale. But there are other challenges to consider, including privacy issues, reach versus accuracy, and the thorny little matter of cookieless environments such as connected TVs and the internet of things.
Naturally, marketers are keen on tracking consumer behaviour, collecting high-level information about their devices, and crunching data to deliver the best possible material to reach potential customers while engaged across all screens. However, with privacy becoming an increasing concern, complying with regulations may be difficult for brands with an international footprint.
With the prospect of General Data Protection Regulation (GDPR) looking to hold data collectors accountable, brands will need to adjust their approach of respecting the privacy of end-users and meeting the interests of their businesses. This will include changing the approach from a standard market monitoring process to being positioned as a liable overseer of crucial data. From here, deploying a hybrid of probabilistic and deterministic practices ensures a set of data that will be balanced and qualified, with the additional consideration for local privacy regulations.
The advantage of probabilistic matching lies in an approach that can be built to adhere to local privacy legislation. Prior to the proposed GDPR regulation, international brands adhered to less-rigid guidelines under the 95 Privacy Directive. The mistrust of online profiling by end-users, the falsification of personal information, among other tactics, were implemented to protect individuals’ privacy. As such, brands now need to look into how to gather, normalise, and share customer data in a way that equally respects their marketing interests and the privacy of users at large.
Though low in profile, consumers now have access to opt out of tracking and identifier expiration dates (i.e. device IDs, cookies, reading history, location information, etc.) to safeguard their activity, if they so choose. Compared to deterministic matching that is harder to scale, this data can be used as seed to create better, engaging marketing campaigns.
According to Oracle, the scale of deterministic device-matching solutions will result in limited reach, pointing to a reliance on a complementary probabilistic approach to improve performance in global markets. To understand the connections between particular devices at scale, your only choice is to use a probabilistic device map to fuel an efficient marketing campaign.
Looking ahead, it will take a solution provider that is agile enough to perform around the delicate intricacies of local market legislation limitations as well as industry working groups self-regulations. The industry should consider relying less on deterministic data, which can be expensive and rife with issues such as PII. By moving away from a reliance on master graph providers, the industry will further ensure compliance with GDPR. Considering probabilistic matching, the benefits are clear in terms of identifying, and even re-identifying, which devices are being used without knowing deeper information about the end-user. Generalizing the collected data, and applying the results, will deliver high accuracy along with great scale, while still providing consumers with the relief and privacy they need.