As the marketing world moves into a cookieless reality, the importance of strong first-party data is being emphasized on a number of fronts. In particular, there’s a tremendous amount of momentum behind leveraging first-party data onboarding as a means of continuing personalization efforts in a cross-channel capacity. But that’s only going to be part of the larger equation.
It’s true that first-party data will be key to success in a post-cookie world. But marketers mustn’t limit their opportunities by thinking about onboarding as simply a path to personalization. First-party data can go much further on behalf of a brand when marketers embrace deterministic and probabilistic onboarding as a part of the same strategy.
A broader view of data onboarding
When most marketers think about data onboarding, they think about matching their offline data to persistent personally identifiable information (PII). And without a doubt, there’s great value in this deterministic pursuit when it comes to improving customer personalization in online marketing channels, refining user experience and bolstering retention efforts. But in reality, there are other kinds of onboarding that marketers need to be prioritizing—not just one-to-one tactics in reaching known customers who have already engaged with one’s brand, but one-to-many.
To optimize results moving into a privacy-first future, marketers are going to need to employ a smart mix of deterministic, probabilistic and contextual targeting. First-party data provides the basis for both the deterministic and probabilistic pieces of this equation, but each use case serves a distinct purpose in the overall marketing picture.
Deterministic data has been put on a pedestal within U.S. marketing circles in recent years. And indeed, if you want to retain, upsell and cross-sell existing customers, one-to-one deterministic data onboarding is a key piece of the puzzle. However, if you’re looking to find new customers and ultimately grow your business, you need to be thinking probabilistically when it comes to data onboarding.
Probabilistic onboarding—a key part of building out audience cohorts for targeting—is an essential component of a strong prospecting and acquisition strategy. Going forward, this one-to-many approach is going to become even more important when it comes to deriving the most possible value from first-party data.
The power of probabilistic onboarding
Outside of the U.S., probabilistic onboarding has become standard practice. This isn’t terribly surprising, given that the rising tide of consumer privacy legislation reached global shores first. Cohorts and probabilistic data have been helping companies operating internationally broaden the applications of their first-party data on a number of levels, all in alignment with privacy restrictions around the obtainment and use of PII. As a result, many leading international companies already think of personalization and prospecting together.
As the U.S. privacy landscape shifts, both in terms of regulations and the policies of titans like Google and Apple, deterministic-centric strategies are going to become far more balanced with probabilistic techniques. And that’s a good thing as it relates to stretching first-party data further.
When companies rely solely on deterministic onboarding, they lose a lot of control regarding what data is onboarded and what isn’t. For example, if a women’s fashion retailer wants to onboard its offline database of 1 million customers, deterministic methods would look to match those profiles to online identifiers, and that match rate might be only 30%. The resulting 300,000 profiles aren’t necessarily representative of the retailer’s customer base.
It’s possible that the retailer’s offline customer base is 80% women, but its matched profiles, when deterministically onboarded, might end up presenting a 60/40 split between women and men—an incomplete and narrow view of the retailer’s true audience. Sadly, this loss of control is entirely out of the hands of the brand or the quality of its first-party data.
With probabilistic onboarding to find net new customers, allof the retailer’s data would be onboarded, whether or not it can be matched to an online key. In this manner, the retailer retains a true view of its actual customer base, which can then be leveraged to build cohorts for scalable lookalike targeting.
Probabilistic onboarding, at its core, helps brands find net new customers. In this sense, it shouldn’t be viewed as something that U.S. marketers are being driven to by the shifting privacy landscape.
Rather, it’s a more-complete and sophisticated approach to first-party data that to date has been lacking in too many brands’ strategic playbooks. In pivoting their thinking for a cookieless future and partnering with companies that have been helping global companies unlock the full value of their first-party data for years, brands have the opportunity to align their data strategies for a future in which personalization and prospecting walk hand-in-hand.
—
Rob Armstrong is a proven product innovator with a 10-year track record of building high-performance organizations and driving the development of next-generation and disruptive marketing and advertising solutions with big data and artificial intelligence.
