Starbucks cup with multiple name misspellings
PHOTO: sikeri

Last year, Forrester found that 54 percent of marketers are using artificial intelligence (AI) to personalize messaging across channels to drive growth. AI promises to deliver “hyper-relevant” experiences for consumers, making marketing more meaningful and compelling, increasing value and loyalty.

The irony is that most marketers are so focused on the promise of AI that they are forgetting about the most important element of the process: data.

Technology: More Machine Than Man

Only 17 percent of marketers in the Forrester research identified a “well-curated” collection of data to train their AI system as important. Yet most experts agree that data is the key to delivering relevant outputs instead of garbage.

With a term like “artificial intelligence,” it can be tempting to assume that the shiny new tool you just picked up will deliver human-like intelligence at scale at the press of a button. But before you rely too heavily on the algorithms in your martech, you have to look at what you’re feeding it. Like a new baby, if you don’t teach her to talk, she’s not going to learn to speak to other people on her own. The data you choose to put into any technology to increase relevance is the most important part of the process because it's what teaches your technology how to interact with others.

If you’re giving your algorithms data about your consumer, ask a few questions to determine how complete a picture you’re creating. Your data might be outdated and sparse. Maybe you identified the wrong person on a shared computer. Your view of your customers’ behavior beyond their interaction with you is probably limited, even if you are the dominant brand in your category. Only by addressing each of these data issues will you increase the likelihood of offering hyper-relevance.

Related Article: What Data Will You Feed Your Artificial Intelligence?

No Experience Is Better Than a Bad Experience?

Imagine your data is stale, perhaps even a few days old. This immediately creates a “relevance gap.” Some brands — for example, a luxury hotel — might decide it doesn’t matter because most people plan and book well in advance. But for today’s customer, hyper-relevance means always being up-to-date. Within a few days, some percentage of your customers have cancelled their reservations, changed their dates, or upgraded to a nicer room, rendering typical post-booking offers inaccurate. It’s important to go through each scenario to determine if your data (and your assumptions for what your AI can do) will generate a sufficient revenue from relevant offers to offset the cost of annoying consumers with irrelevant content..

Marketers should lean heavily on the teams focused on customer experience such as CRM, loyalty and digital or ecommerce experts to run through scenarios and make smart decisions about how far they can take their data, and where they will need an upgrade to fill in the gaps. Chances are, you've likely received a flood of advertising for an air-miles credit card — which is the exact one you already have. It’s possible that given the available data, no personalization is better than incorrect or outdated personalization.

Related Article: Data Ingestion Best Practices

Start With the Best Option

When you approach your “hyper-relevance” marketing projects with an eye towards your data, not only will you spot the problems, you’ll also spot the opportunities to reach the next level in hyper-relevance. Accenture recently found that 73 percent of CEOs recognize the need to deliver more personalized offerings to consumers, so the competition to deliver the most convenient, accurate and valuable interactions is already stiff.

Consumers’ experiences with one type of service raises their expectations across the board. Here are a few examples of the types of personalization that your consumers are being exposed to — and that your brand, as a result, is competing with:

Sephora collaborated with Pantone to create a program called ColorIQ, which helps personalize makeup recommendations for customers. With a quick in-store scan, a person receives a custom four-digit code that they can use to find complimentary makeup online, on the mobile app or at any Sephora store.

Best Western recently upgraded its mobile marketing with hyper-relevant data. Not only does it pay attention to where consumers are booking hotels, it watches where the consumer is when they open an email or open an app. This allows the company to personalize additional content and offers in real time.

Both of these cases show brands using good data to drive hyper-relevance. By thinking first about what consumers might need, and next what data you have (or have to acquire) to deliver on that need, your quest for hyper-relevant marketing is much more likely to end in success.