Feb 11, 20196 min read

Know your options: Identity resolution beyond the walled gardens

The first rule of business is “know your customer.” But, these days, it’s also one of the most difficult problems for brands.

The big walled gardens — Facebook, Google and Amazon — have a kind of identity resolution nirvana. They can approach each user as a single identity, since those users are usually logged on.

But that has made many brands and data providers outside the gardens envious, leading to a variety of approaches to come up with unified identities for customers and prospects.

Studies bear out that customers desire consistent identities with brands. A 2018 Personalization Pulse Check from Accenture Interactive, for instance, found that 91 percent of consumers “are more likely to shop with brands who recognize, remember and provide relevant offers and recommendations.”

Persistent ID mapping

The problem, according to Richard Foster, UK managing director at identity resolution provider LiveRamp, has been that “marketers must either target audiences in the same walled gardens where their preferences are shared, or use cookie data [for web interaction] without much context beyond their customers’ browsing history.”

Option one limits the available channels, he has written. “And option two relies on impersistent data.”

To counter the dilemma of impersistent data, a brand typically uploads its customer list to LiveRamp, which matches it with other data sets in its repository of offline and online info. Usually, LiveRamp connects the various datasets through a “persistent identifier” — a piece of reliable information that is present in the different sets, such as an email address.

If the approach to identity resolution is a continuum, one end — a full multichannel set of data about an individual, matched with a high level of confidence — would be represented by the solution LiveRamp offers to brand advertisers.

Account-based identities

But many marketers are turning to a variety of other solutions, many without persistent identifiers to match up datasets, because not every brand needs that level of identity resolution.

Calabrio, for example, is a brand that provides contact center solutions to businesses. Its key identity need is to identify key people at targeted businesses so it can get new contact center business.

For existing accounts, Calabrio CMO Rebecca Martin said, her company tracks users at existing accounts through Terminus, an account-based marketing solution for targeting key accounts. Terminus lets Calabrio know, for instance, what kinds of web content users from Company A are visiting, as determined by IP addresses, cookies and other signals, whether or not they go to Calabrio’s website.

For new accounts, Calabrio employs Demandbase, which similarly uses corporate IP addresses to flag which new accounts are interested enough to visit Calabrio’s web site. Zoom’s reverse IP lookup then helps Calabrio determine the specific people to be approached at the new businesses, based on their corporate IP addresses.

That, essentially, is the extent of Calabrio’s need for identity resolution: a web-based understanding of what existing accounts are interested in, and what companies are checking out Calabrio’s product line.

Genesys is a brand that sells call center technology and is in a similar business to Calabrio, and thus has a similar need for what might be described as a relatively thin identity resolution.

Like Calabrio, Genesys also works with Terminus to identify what specific corporate accounts are visiting its web site, but it uses data provider Clearbit to look up visitors’s IP address, match it to the company at that IP, and then determine individuals to approach.

As with Calabrio, that’s the extent of Genesys’ identity resolution needs. CMO Merijn te Booij said that, unless the user provides a name, his company does not try to match site-visiting identities with other identities on the web or in other channels, like apps, physical store purchases, calls to contact centers and so on.

Resources on identity and paid media tools

Connecting through street address

For a furniture retailer like Serena & Lily, however, there’s a need to understand its customers and would-be customers in the real world — where furniture lives — as well as in digital realms.

VP of Direct Marketing and eCommece Carla Rummo said her company tracks identities through customer data platform AgilOne, where the golden master of customer data resides. A street address acts as the “persistent connector” that is often used to tie together identities from the different realms, although sometimes an email address is employed. Since one of the company’s primary marketing methods is a snail-mail catalog, the company has street addresses for its customers and many would-be customers.

In the case of a brand like coupon and marketing services provider Valassis, it’s important to connect a variety of diverse data sets, particularly so it can engage in remarketing that follows up on site visits with non-web actions, like direct mail or email. SVP, Product and Marketing Michelle Engle said this includes connecting such diverse data sets as “offline coupon redemptions, addresses, daily intent signals, daily location signals and households across devices.”

To create a single identity, Valassis matches a user’s IP address, cookie ID or mobile device ID to more than 150 million US street addresses, although many of the profiles also have email addresses. The match is possible because Valassis, or one or its data providers, has seen that IP address, cookie ID or mobile device ID when a user logged into a site or was otherwise identified.

Identity resolution through self-identification

A company like Claritas creates audience segments for brands, but, SVP of Digital Solutions Barney Marvin said, those multi-channel identities are not resolved into one unless a customer has self-identified at some point, such as logging in or filling out a form.

For the campaigns of client brands that are focused on a channel like web or in-app ads, Claritas doesn’t need more extensive identity unification.

For SAP Customer Data Cloud, which also conducts targeting for brands, individuals themselves similarly drive identity resolution.

Head of Product Marketing Sonny Dasgupta said that registered customers can link their other selves — such as their social media accounts — to a registered account, specifically so they can better communicate with favorite brands. But it’s left up to the individuals, instead of conducting the match through IP addresses, street addresses or email addresses.

“By providing transparency and control to your customers, over the data that you have on your customers,” he said, “you have a foundation for trusted personalized relationships.”

Probabilistic matching

A persistent identifier like a street address or email address offers a hard match to connect separate datasets about the same individual, thus providing a highly accurate definitive match known as deterministic.

But many brands — and their identity solution providers — do not have that persistent identifier across the data sets. Instead, like customer data platform RedPoint Global, probabilistic techniques are used to build likely identities from behavioral patterns for brands.

The same IP address at night from different devices, for example, can mean those devices probably live in the same place, and are probably owned by one or more individuals in the same home. Probable identity down to a household makes assumptions about individuals in that household.

Another identity solution provider, customer experience management platform Pimcore, similarly creates probable visitor profiles based on behavior, particularly the behavior of web site visitors.

For its brand clients and for itself, Marketing and Partner Manager Christian Kemptner said, Pimcore employs analytics platform Matomo to create unified visitor profiles based on site behavior, such as the same cookied browser coming back to the site and visiting the same product pages on multiple occasions. It then uses a de-duplication service to automatically remove duplicates and merge visitor, customer and third-party data.

Without a persistent connector like a street address or email address, the Pimcore profiles are based on likely matches, which meets the needs of its client brands.

The ID initiatives

While all these brands — and their solution providers — have different needs, approaches and levels of accuracy, several initiatives may soon supplant the need for these kinds of identity resolution by brands or their solution providers.

There are three initiatives underway to populate the web and other channels with a single ID. The Interactive Advertising Tech Lab’s DigiTrust and The Trade Desk’s efforts are currently focused on the web, while the Advertising ID Consortium — closely connected to LiveRamp’s multi-channel IdentityLink — seeks to create an ID across other platforms as well as the web.

Although those efforts are driven by the advertising industry’s desire to avoid the inaccurate and costly effort of massive cookie-matching that currently takes place, the eventual result might also be to provide one or a few ways for brands to track individuals via an ID.

If so, identity resolution would get a whole lot easier.

This story first appeared on MarTech Today. For more on marketing technology, click here.