Simon Poile, GM of AWS Digital User Engagement, presented at Digital Summit Seattle February 26, 2019. Over the last 10 years, the only thing that hasn’t changed in customer engagement is the value of a trusted relationship. At Amazon, we believe strongly in customer-driven innovation and are constantly striving to provide the best experience for our customers. In our experience, customers are always going to adopt new devices and channels, want personalized outreach, and demand timely and relevant communication on matters they care about the most. Marketers concerned about engaging their own customers must challenge themselves in the same way, evolving and innovating while never losing focus on the most important thing, your relationship with your customer.
Session attendee learned how to leverage innovative technology to:
–Learn more about their customers through a single view derived from disparate data sources
—Create highly personalized engagement experiences
—Better understand when and on which channel to engage their users
It’s not just about adding in their first name to an email, or create an ad that shows their name on a t-shirt.
84%
of customers say being treated like a person, not a number, is very important to winning their business.
Fewer than half (49%) of marketing leaders believe they provide an experience completely aligned with customer expectations.
The definition of a “good” experience has changed. Customers demand not only relevant offers, but to feel truly known and understood as individuals. In a culture of immediacy, they also expect engagement at their exact moment of need. What’s more, this level of engagement is viewed as standard across the entire customer journey, prompting marketers to think well beyond their traditional domain.
—Salesforce.
Not only is Image A a personalization fail, it’s not true personalization anymore. No one gets an email addressed to them and thinks “oh wow, this is just for me!” True personalization in emails is displaying timely content that is directly relevant to the users’ behaviors or preferences.
From “we are having a sale on all home goods” to “we are giving you 10% off this item that you’ve viewed 20 times in the past week”
BUT – don’t be creepy. Image 2 is creepy. You don’t want to take personalization to the level of coming across as knowing too much about the customer for no good reason. Data is a privilege that we are fortunate to have in abundance, but once a customer perceives you as knowing too much about them without actually driving value, you’re breaking down the trust barrier and they might opt for another service or brand instead.
First, let’s define where that line is. Intelligent virtual assistant (IVA) provider Interactions teamed up with The Harris Poll to conduct a survey of 2,000 consumers in the U.S. aimed at determining where the “creepy” line is, and when AI crosses it.
The ensuing report (free with registration) showed that consumers say they are put off when an AI system knows information they didn’t provide directly, or that involves other people in their social networks. The survey found that about half of those surveyed think it’s creepy when:
AI knows other household members’ past interactions with a company (52 percent).
It uses social media data to make suggestions (50 percent).
It knows past purchase history from a different company (42 percent).
“
It’s not just about adding in their first name to an email, or create an ad that shows their name on a t-shirt.
The foundation of incredible digital marketing is data.
The effectiveness of your analytics practice is measured by the value you deliver back to the business.
The foundation of incredible digital marketing is data.
The effectiveness of your analytics practice is measured by the value you deliver back to the business.
DEMOGRAPHICS –> Target based on
Source your data
Load your data
3rd party personalization engine
stream all of your data to a third party recommendation engine that you can then call on when you're looking to fill in content blocks.
extensions -
Coinbase - looking at event streams and doing attribution to the user, not the campaign
better insights into the events.
–netflix - likelihood of finishing a season = if you've watched episode 3.
Understand your customer:Gather data from multiple sources (beyond just marketing) to gain a complete picture of your customer (in a non-creepy way. Use it to find macro trends about how certain groups of customers are engaging with you or your product.
Gather data about what is important to your customers, not just your business or marketing efforts.
Figure out what your customers like in an engaging, non-obtrusive way so that you can tailor their experience accordingly and give them what they need not what they say they want
—don’t be afraid to ask what they want.
This is really important. Customers know what they like and they don’t like, and they can tell us what they think they want, but it is really up to us to read between the lines (and in the data) to figure out what it is that they truly need, and then deliver on that.
Data can help you see what your customers are doing, where they are struggling, and what they are lacking. You as a marketer can use that data to create campaigns or engagement experiences that delight your customers in part because it is what they didn’t even know that they *need* - not want.
This leads to things like recommended items on their shopping home page, or proactive text alerts about their baggage allowances or that their flight is delayed or their gate has changed, or
First-Party Data=
Customer Support Data
Sales Data
Website Tracking
Mobile Application Tracking
Engagement Campaign Data
Social Sentiment Analysis
Device Data
Customer Surveys
Predictive Data
= running ML models on the data that you have to create predictions about what your customers might do
Buying additional insights on your own first-party audience means you learn more about your customers for a more complete view of who you are trying to reach.
What if we had the ability to learn the purchasing behavior of our customers? What could we do if we knew with relatively high probability what their next purchase will be? There are many things we could take action on if we possessed this predictive capability.
Sentiment Analysis can be a great indicator of future behavior
Happy Sentiments —> Turn them into advocates
Sad Sentiments —> Reach out and dive deeper to understand and fix their problems
Let’s look at a retail example. As consumers, we have some intuition of purchasing habits. We might tend to re-purchase types of products we have good experiences with, or conversely, we might drift toward alternatives as a result of unsatisfactory experiences. If you buy a book that is part of a trilogy, there is a higher likelihood that you will buy the next book in the series. If you buy a smart phone, there’s a high probability that you might buy accessories in the near future.
Inferred Channel Preferences –> Predicts what channel your customers are more likely to engage on depending on factors like time of day, type of promotion/content, etc.
CHURN PREDICTION;
-recency of use
-junk/spam email flagging
Sky Alert example
Every day items like fridges, lightbulbs, and printers are now valuable sources of user data that you can use to create really relevant, personalized, and timely marketing messages and to better understand macro trends about your customers
This device data helps you tailor your engagement efforts based on your customer’s interactions with the devices they’ve bought from you.
DEMOGRAPHICS –> Target based on
Source your data
Load your data
3rd party personalization engine
stream all of your data to a third party recommendation engine that you can then call on when you're looking to fill in content blocks.
extensions -
Coinbase - looking at event streams and doing attribution to the user, not the campaign
better insights into the events.
–netflix - likelihood of finishing a season = if you've watched episode 3.
Understand your customer:Gather data from multiple sources (beyond just marketing) to gain a complete picture of your customer (in a non-creepy way. Use it to find macro trends about how certain groups of customers are engaging with you or your product.
Gather data about what is important to your customers, not just your business or marketing efforts.
Figure out what your customers like in an engaging, non-obtrusive way so that you can tailor their experience accordingly and give them what they need not what they say they want
—don’t be afraid to ask what they want.
This is really important. Customers know what they like and they don’t like, and they can tell us what they think they want, but it is really up to us to read between the lines (and in the data) to figure out what it is that they truly need, and then deliver on that.
Data can help you see what your customers are doing, where they are struggling, and what they are lacking. You as a marketer can use that data to create campaigns or engagement experiences that delight your customers in part because it is what they didn’t even know that they *need* - not want.
This leads to things like recommended items on their shopping home page, or proactive text alerts about their baggage allowances or that their flight is delayed or their gate has changed, or
First-Party Data=
Customer Support Data
Sales Data
Website Tracking
Mobile Application Tracking
Engagement Campaign Data
Social Sentiment Analysis
Device Data
Customer Surveys
Predictive Data
= running ML models on the data that you have to create predictions about what your customers might do
Buying additional insights on your own first-party audience means you learn more about your customers for a more complete view of who you are trying to reach.
NOTE - SMS and a time icon for this slide.
Create your top-level customer segments based on your understanding about who wants or needs to know about your message
From there, segment further based on channels that customers in each segment are most likely to engage on.
Finally, segment further to ensure you’ll reach your segments at the right time (time zone, historic engagement times)
PERSONALIZE – For Hannah, list out the 3 albums she’s most likely interested to. Or, for these 3 albums, list out the segments most likely to engage with them.
Create segments based on when users have engaged in the past and target important or timely messages to them at that time
Don’t just segment based on basic demographic data (although that’s a good place to start). Go further with advanced, dynamic segments based on real-time data, deep knowledge of historic actions, and user preferences.
Create opportunities for your customers to organically tell you what they like and don’t like, and then listen to them.
Cater experiences to the audience based on their behaviors and preferences vs. just age, sex, geography.
RIGHT CHANNEL
RIGHT TIME
(event-based triggers, location)
Why reach out? - Understand why you’d reach out to your customer.
Churn prediction
Purchase Affinity
Segmentation: multi-dimensional segmentation at scale.
React in the moment (event-based triggers, location)
Use your data to understand what channels your customers use, what social platforms they care about, and where they engage with certain types of messages
Based on that analysis, decide if there are channels that you haven’t adopted that might make sense to utilize to better engage with your customer.
Analyze your user behavior to gain a deep understanding of what types of events are important to your customer, and then use those events to deliver impactful, timely messages and experiences.
Don’t be afraid to adopt new channels that are important to your customers
Use historic data coupled with predictive analytics to schedule campaigns when users are most likely to engage
Instantly trigger messages based on real-time actions that matter to your customer.
It’s not just about adding in their first name to an email, or create an ad that shows their name on a t-shirt.
IoT personalizing a customer’s experience based on interactions with a device/platforms they have, not just the channel. TIMELY – CROSS DEVICE — PERSONALIZED Roomba = bridge device between the user’s device and the user itself. What if there was an automated campaign – iot system updates messaging system to say the filter is 90% used, and then send them a 10% off coupon, etc.
CHECK ABOUT THE AMAZON BRANDS
And here are a few customers that already trust Amazon to help them to engage with their brands.
Disney Streaming Services are using Amazon Pinpoint to engage tens on millions of users and alter them when a game is about to start or something notable happens in a match.
Netflix uses the same platform to send all their email messages to their customers.
McDonald’s runs their entire user facing digital properties on AWS.
Coinbase
These companies use a combination of our Digital User Engagement services, including Pinpoint, Global SMS, Mobile Push and Simple Email Service to deliver billions of messages and similarly gaining billions of valuable analytical insights into their customers every single day. And in case you were wondering, yes, we also support Amazon.com’s own various businesses to reach their customers too.