The Future of TCPA Compliance and 1:1 Consent

Learn More

Succeeding with AI, Your Foundation Must Be Anything but Artificial

The Next Big Thing in Marketing Still Needs Accurate Identity Data

Artificial Intelligence – AI – is suddenly everywhere in the marketing ecosystem. You read about it in the trade press. Learn about it at conferences. Gain insights in blog posts. And it’s already being taught in business schools. IDC estimates spending on cognitive and AI systems will reach $12.5 billion this year, and top $46 billion through 2020.

All for good reason: AI significantly advances the potential of automated marketing. It promises more accurate segmentation and targeting, teasing out non-intuitive insights by parsing mountains of data. It enables instantaneous delivery of more relevant offers to engage customers and encourage repeat visits. It can even anticipate customer inquiries or issues to shape more responsive service.

But for all its dazzling potential, AI operates with the consumer identity data you provide it – and if that data is inaccurate or incomplete, you won’t get the on-target insights you’re banking on.

You’ll just get faster misses.

Two Kinds of Data Critical for AI Success:

1

Consumer Identity Data that nails down who and where your customers and prospects are, and how to reach them

2

Attribute Data that provides the context for machine learning for modeling and manipulation


Consumer Identity Data: The Foundation for Marketing AI

Consumer Identity data ensures that your customers and prospects are who they say they are, and also informs who you think they are. It is the foundation of all inbound and outbound targeted marketing – by definition, directed towards specific people – whether driven by AI or not.

In the last few decades, the volume of identity data and its velocity of change has multiplied many times – and so has the potential for omissions, duplications and outright errors. Once there was only name, mailing address, and telephone number (landline, of course) – all of which were static unless and until a consumer moved.

Now, that same consumer identity data has been joined by multiple email addresses, social media IDs, mobile and VoIP phones, device IDs and IP addresses. These new identifiers are much less static, subject to changes that are not tied to a physical change of address, making it more challenging than ever to keep them all accurate – and accurately linked.

It’s no wonder that 30% of CRM data is either outdated, incomplete or inaccurate at any given time.
(Source: Infutor Identity Graph 2017)

If your identity data is not constantly updated, standardized, appended and cleaned so that all identifiers are accurate, up-to-date, and properly linked, your marketing AI will be built on a faulty foundation – and your results will suffer.

VELOCITY OF CHANGE AMONG U.S. CONSUMERS

11.5%

Move every year
(U.S. Census)

30%

Change email addresses annually
(Accenture)

17%

Create a new email address every 6 months
(Accenture)

5.59M

By 2019, 2.9 million email users will have 5.59 million email addresses
(Radicati Group)


Attribute Data: The Foundational Context for Marketing AI

If identity data establishes who your customers and prospects are, attribute data completes an individual portrait of each of them. It fills out the details of their likes and dislikes, their attainments, their choices. It’s no exaggeration to say that attribute data enables the targeting in targeted marketing.

If identity data has expanded in the last decade, attribute data has absolutely exploded – in quantity and frequency of change – to enable increasingly accurate and specific targeted marketing across the customer lifecycle.

Key words: frequency of change.

Like identity data, attribute data is dynamic – because consumers live dynamic lives. They get married. They have children. Their interests change. And the marketplace that caters to their needs – the marketplace where you sell to them – is being disrupted and reshaped at an unprecedented rate.

If you want to retain your customers, predict their future actions, project their lifetime value, and acquire more that share their characteristics, your AI needs information that reflects who they are now, and how they’ll change throughout their life’s journeys. And all of it must be correctly linked across all instances of that consumer to the single consumer identity that pinpoints who they are.

If your data is not properly interlinked – or you’re relying on buying patterns and decisions from two years ago, or even 6 months ago – your marketing will be off target.

VELOCITY OF CHANGE AMONG U.S. CONSUMERS

4M

Marry every year
(CDC)

1.6M

People get divorced
(CDC)

4M

Children are born every year
(Accenture)


6 Ways AI Will Power Your Marketing

1. Look-Alike Modeling

Segmenting customer data to uncover identifiable commonalities in demographics, psychographics, ownership, behavior, and other qualities is the basis for targeted marketing. It’s already a growing field for machine learning, as AI-driven tools can process much more data much more rapidly.

AI for Look-Alike Modeling:
When you can identify common characteristics that your best customers share, you can use them to identify your best prospects.

AI allows marketers to parse customer data more finely to gain more accuracy and detail in their models, with the ultimate goal of creating “segment of one” modeling for truly personalized marketing.

The more accurate and personalized the modelling, the more successful the acquisition efforts it fuels.

Data requirements:
Accurate, interlinked, up-to-date customer data is essential; inaccuracies here will send all your targeting models off-target.

More granular data analysis requires fuller and more complete attribute data to deliver accurate, nuanced modeling results – bringing you closer to true one-to-one marketing.

2. Lead Verification and Scoring

AI for Lead Verification and Scoring:
With accurately linked identity data, you can verify and identify each inbound lead regardless of the channel they use to reach you, while also ensuring you have the current and correct contact data to reach them.

Segmentation models can then be used to score each inquiry and orchestrate the proper response across available marketing channels.

With AI tracking and learning from the responses and actions of your prospects in real time, your marketing will become more and more effective, and adapt automatically to changes in the marketplace.

Data requirements:
On-demand access to the most comprehensive identity data available is essential for real-time lead verification.

Accurate scoring similarly requires real-time access to complete and accurate attribute data, interlinked with identity data. And of course, the segmentation modeling underlying your scoring must be based on accurate and complete customer data.

3. Outbound Targeting

Once prospects and leads are identified by segmentation model, you need to serve them content that will interest, engage and ultimately convert them from prospects to customers. By automating many of the manual steps in this process, AI and machine learning enable your marketing to expand in scale while gaining accuracy in targeting – with results improving over time as the AI learns from your prospects’ actions.

AI for Outbound Targeting and Segmentation:
With AI driving your outbound marketing, you can reach more and more finely targeted segments at greater scale, bringing you closer to genuinely personalized, one-to-one marketing at the same time you are reaching more prospects.

AI can also manage versioning in both copy and imagery to match recipient needs and interests, simplifying the administrative work of managing campaigns at scale.

Data requirements:
Real-time access to comprehensive, accurate, interlinked identity data is essential to drive acquisition campaigns at scale, and to suppress duplicate or undesirable consumers.

Up-to-date attribute data is necessary for personalized targeting and messaging that consistently hits the mark and drives stronger response rates.

4. Inbound Engagement

AI for Inbound Consumer Engagement:
The same kind of customization that has long characterized outbound marketing can now be adapted to serve personalized content in real time when prospects click into your website, based on the source of the click.

Offers and imagery can be tailored to prospects’ interests and motivations – increasing stickiness and likelihood to convert.

Machine learning in this application drives continuous improvement in matching content to consumer segments.

Data requirements:
Accurate, complete interlinked identity data in depth is necessary to correctly recognize and identify incoming prospects, who can use multiple channels you reach your business in a single day. It’s also the only way you can ensure you’ll have accurate addressability for follow-up.

Comprehensive and timely attribute data – covering as many dimensions as possible – is equally essential to effective engagement.

5. Recommendations

When you reach out to your newly-converted customers with relevant recommendations that address their evolving needs and interests – and provide informed, responsive service when they contact you – you strengthen their connection to your business and maximize their lifetime value. AI can continually improve personalized recommendations for each customer based on their actions, while also applying learning from customers with similar characteristics.

AI for Personalized Recommendations:
The use of AI to predict what customers are most likely to buy or repurchase next is familiar to almost everyone, even if they don’t realize it, thanks to efforts like Amazon’s recommendation emails and website personalization.

By building a one-to-one relationship, AI-driven recommendations increase engagement and lifetime value – but only if your recommendations are relevant and welcomed.

Data requirements:
The value of accurate, interlinked identity data is obvious; you have to be able to recognize each customer regardless of the channel they use, and have correct data on the preferred channels to reach them.

For the most accurate and effective results, recommendation engines also need the most complete and timely attribute data you can provide.

6. Service and Retention

AI for Customer Service and Retention
AI is already being used to predict why a customer may be getting in touch with your company and what channel they’re likely to use, enabling a more streamlined and personalized service experience.

It can also drive real-time recommendations for upsell and cross-sell offers during each contact – and identify potential at-risk customers for retention offers.

Data requirements:
Complete identity data across multiple channels – phones, email addresses, IP and device IDs – is essential to recognize and reach your customers.

And attribute data must be complete and up to date to drive accurate predictions and relevant recommendations.

It’s All About the Data

AI is here to stay. It delivers tangible marketing benefits across the customer lifecycle – from prospect identification through targeted acquisition, engagement and retention. And the number of providers is exploding, applying its benefits into new marketing niches and making them available to virtually any marketer.

As both supervised and unsupervised AI-driven marketing solutions enter the mainstream, they will change your marketing in both predictable and unpredictable ways, with implications we are only now beginning to understand.

Two things are certain:

  1. AI will make your marketing more personalized, more responsive, and more effective.
  2. The quality and accuracy of identity and attribute data to “ground” AI learning has never been more important. Writing for MarTech Today about the effects of “genius as a service” offerings, Barry Levine notes:
    “Data has certainly been a competitive advantage for some time… But when insights, understanding or highly intelligent agents are available on demand, the quality and amount of data becomes the key differentiator.”

Before you venture deeper into AI, revisit your consumer identity and attribute data and data resources – and make sure both are as complete, accurate, and properly interlinked as they can possibly be.