Advertising Week in New York sparked a lot of buzz about “people-based marketing.” In fact, it was a central theme at the conference with an entire track of 11 distinct sessions dedicated to the topic. People-based marketing is the concept of leveraging detailed digital data to tailor advertising messaging to individuals across all of their devices in parallel, rather than to individual cookies, devices, or AdIDs. Now that data has become so prevalent, industry thought leaders are envisioning campaigns in which ad creative will be personalized and dynamically optimized with less risk than ever before.

People-based marketing is a compelling proposition. It’s certainly the stuff of inspiring presentations and there’s no question that the capabilities have been a long-time coming. Unfortunately, successful execution could take longer than marketing visionaries predict.

What’s driving the trend of people-based marketing is not a sudden breakthrough in ad creative talent or better digital ad units. It’s the data. Or more precisely, it’s the perception that digital consumer data has finally become ubiquitous and robust enough to allow this vision to come to fruition. But no matter how much data exists or how far the ecosystem has evolved, people-based marketing won’t work if the data is wrong.

Read More: People-Based Marketing: 5 Ways to Win

Disturbingly, today’s targeting data is wrong as often as it is right.

In a world saturated with ads, successful brands must provide consumers with a personalized experience. It doesn’t matter how tear-jerking or empowering the creative might be — if it’s served to the wrong person, there could be negative consequences. For example, off-target messaging can make the brand suddenly appear less relevant, out of touch, or even offensive. Imagine the impact on the Nike brand if (and when) ads featuring Colin Kaepernick were (and are) incorrectly personalized and served to individuals with passionately-held negative perspectives on the Nike spokesman!

Digital ad data isn’t nearly as accurate as most marketers assume. We’ve previously published data from our own Emodo Institute studies and others that demonstrate the severity of the data accuracy problem. All reliable evidence shows that approximately half of all targeting data is inaccurate.

If you were to run a campaign in which only 50 percent of the impressions were viewable, would you double down and sequence your creative so a follow-up message was served, even to the 50 percent of users who never saw the first ad? Not if you’re focused on a positive consumer experience. You’d have the same issue if you personalized creative using a dataset that’s only 50 percent accurate.

Read More: All You Need to Know About People-Based Marketing

A recent survey of over 500 marketers in the US and the UK found that data quality was the #1 obstacle to personalization. For marketers to fully embrace people-based marketing, they need to be confident in the accuracy of their data. Confidence requires proof that the data is accurate, but right now, there seems to be a proof shortage in the world of marketing data.

Clearly, from the enthusiasm at Ad Week, many marketers would jump at the chance to deliver more personalized, emotionally-engaging brand stories. The one-to-one nature of digital channels and the depth of readily available consumer data make it all seem close at hand. But with such a high potential for waste (impressions and spend), brands betting big on the current state of marketing data are likely in for a rude (and costly) awakening. While the data is hailed as the key to making people-based marketing work, it’s quietly failing at its most basic targeting functions.

The most direct line to true individualized marketing puts the onus on marketers to ask some hard questions about the data they’re using to categorize and target consumers.

• From where is the data sourced?
• Has the data been manipulated at all for the purpose of maximizing scale?
• How might accuracy have been compromised in that process?
• How accurate is the data in each segment?
• Can the data provider provide proof of its accuracy?

The conversation about people-based marketing is exciting and it would be difficult to find a marketer who would disagree. Big ideas like dynamic creative optimization platforms (DCO), better storytelling, emotional engagement, and message personalization are almost always at the core of those discussions. However, if the data is the crucial ingredient that makes it all work, its accuracy can no longer be taken for granted. Once data accuracy is part of the conversation, people-based marketing will finally be as achievable as it is exciting.

Read More: The People-Based Data Lake – A More Nimble Marketing Data Solution