People-Based Marketing Gets Complicated Quickly In The B2B World

"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Brian Jones, global head of performance advertising and biddable media at DWA Media.

B2B advertisers always want to know where they should invest their marketing dollars. For a long time, the answer has been, “Depends on where your audience is.” If they could figure out where their target audiences spend time, they could focus their marketing efforts there.

But this recommendation does not apply in our fragmented media world, nor is it the most advanced way for B2B companies to build marketing strategies. That’s why I generally recommend that B2B marketers ask a different question: How do we build the right audience?

Channel The Channels

Most people have social media accounts, watch TV, browse the web and use Google to search daily. Is one channel more important to marketers than another? Not really.

B2B marketers must target a very narrow and finite audience. It’s hard to use standard advertising targeting methods, such as those that are interest- or context-based, to serve ads only to a select group of individuals.

Unfortunately, many B2B marketers still use these tactics and hope for the best when they should be taking better advantage of their first-party data, which sets the right foundation for building the right audience. Almost all channels have some type of first-party data integration, including Google, Facebook and demand-side platforms (DSPs).

Digitize To Capitalize

Many B2B marketers still don’t use data onboarding to transfer offline data to an online environment. Companies like LiveRamp and Oracle Data Cloud can connect offline customer records to online users by matching identifying information.

Once a brand has identified its offline customers online, it can perform audience segmentation to create cross-selling and upselling opportunities. Quantcast and others are developing products that identify high-risk customers and take pre-emptive action to prevent them from disengaging with a brand.

Until recently, B2B marketers who did data onboarding had low match rates, which limited scalability. This is quickly changing with products such as Acxiom’s AbiliTec, which matches a consumer’s email address to his or her business email address, significantly increasing overall match rates for B2B marketers.

Prospect The Best

What should new customers look like? A lot like existing customers.

Data onboarders have integrated with all major advertising platforms, including DSPs, social networks and Google. Once an audience is transported to these marketing platforms, they can build lookalike audiences and market to them.

In many cases, B2B marketers’ first-party data has limited attributes, which then hinders their ability to build highly qualified lookalikes. In these cases, B2B marketers can enhance their first-party data sets to help build new models by layering on third-party consumer data and offline data.

Learn About The Journey

Once the correct audience is identified, B2B marketers must understand the buyers’ journeys. Brands must be present in their audiences’ conversations throughout that journey to raise awareness, generate interest, create desire and encourage action.

Although most seasoned marketers start by mapping the process with a bit of intuition and some historical data, it is important to constantly learn and pivot based on data and analytics. Marketers should start by using their analytics attribution.

Google Analytics, for example, offers access to a wealth of attribution data, such as detailed information about the paths that lead people to conversions. It’s possible to see how media types, such as display, social and search, interact and work together. There is also data on how many touch points or interactions with an individual occurred before a conversion.

These insights can help B2B marketers understand how media types interact and influence in order to build a media portfolio mix. They will also offer clues to whether or not B2B marketers should develop and deploy a comprehensive attribution model in-house.

Match The Graphs

Ad tech companies are building identity graphs using deterministic and probabilistic data to match one individual to all of his or her devices.

DataXu’s OneView product, for example, bridges the identity gap across devices and builds a single, comprehensive view of a consumer’s path to purchase. Drawbridge and Oracle Data Cloud are also beefing up their identify graphs to create better user experiences and increased ROI through the enablement of more intelligent advertising.

When choosing the right vendor, there are many questions to ask. The more informed B2B marketers are, the better decisions they can make. Do they work with multiple DSPs or within one marketing stack? If they are consolidated into one marketing stack, more than likely they can use that particular vendor’s tool set if available.

If they are decentralized and use multiple DSPs or online platforms, they might need to use an independent third party, such as the Oracle ID Graph, to connect the dots between disparate systems and channels.

Whether you are a B2B marketer or B2C marketer, the best advice I can offer is to truly understand the data sources, how old the data is and whether it is deterministic or probabilistic. Marketers must start asking questions to each vendor and hold them accountable. They are very good at using lingo that leaves marketers feeling as if the vendors hold the keys to success. Test, test, test.

Ask For More

Post-campaign, people ask how they know they were successfully targeting the right group. Most marketers go straight to their ad platform and web analytics data. Until recently, these choices provided traditional metrics, such as impressions served, click-through rate, conversion data, page views or bounce rate, but that information won’t suffice for B2B marketers.

But B2B marketers need other metrics, such as type of company, industry and size of company. Bombora, Dun & Bradstreet and other companies continue to make advancements daily and help B2B marketers verify that they’re targeting the correct audience. With this information, they can validate and justify continued ad investment.

Follow DWA Media (@dwaTechMedia) and AdExchanger (@adexchanger) on Twitter.

1 Comment

  1. When choosing a x-device, identity-graph vendor there are several other key considerations:
    1) What is the vendor's source of ground truth? What's the data source, for example logins to a set of websites? How large is the truth set? How fresh is it? How representative is it? Does it scale across geographies? A matching algorithm is only as strong as the data on which it is trained and tested. Garbage in / garbage out.
    2) When a vendor quotes an accuracy rate -- "we make matches with X% accuracy" -- what does it actually mean? How do they evaluate Type 1 vs. Type 2 error, i.e. when they miss making a connection that should exist or they make a connection between two devices that are actually unrelated?
    3) How does accuracy change with scale? I can make a handful of x-device connections with 100% confidence, but how does that scale to 10million people, 100 million people or a billion people? Quoting a static "accuracy rate" is often misleading.
    4) What about consumer notification? As an internet user, do these techniques feel "shady"? Is there consumer notification that their activities are being used to stitch together all of the devices in their lives? If there were legislative action or a consumer backlash, would you feel comfortable that your brand is associated with these techniques?

    With the proliferation of browsers, devices, and apps, there has been a massive fragmentation of a Person into multiple digital identities. But putting Humpty Dumpty back together again is rarely as easy as the sales pitch would indicate. Caveat emptor.

    Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>