/ Customers

3 Personalization Use Cases for eCommerce

As you look at well-orchestrated content, influencer, and subscription marketing campaigns from top e-Commerce retailers, you may wonder - how do you work your existing website to get more revenue from it? How could we turn the intent signals into behavioral campaigns that generate revenue?

Hint: it’s not distribution, it’s data.

Let’s walk through 3 eCommerce examples.

Use Case 1 - The Home Depot 

Content marketing works. A revenue driving content strategy guides your visitors through the sales funnel without making them feel like you’re pitching. Your customer controls what pieces of content they interact with, while you enjoy the climb in conversion rate.

Websites which implement content marketing are experiencing an average conversion rate of 2.9%, compared to a 0.5%conversion rate of websites without a content strategy. That's nearly a 6x increase.

Known for being the go-to place for DIY projects as well as serious construction, The Home Depot's marketing agenda incorporates a clearly defined strategy of content prospecting.

Browsing the "DIY Projects and Ideas," section doesn’t feel like pitching to the visitor. A visitor could convert or leave, all the more grateful to the Home Depot for demystifying a project.

Their website facilitates ordering tools and supplies for delivery, while also providing a virtual library of instructional videos and info for DIY projects. Some examples:

  • View an assortment of different instructional pages for a variety of projects both indoor and outdoor - appliance selection tips, how to build a fence, and even basic tasks like caulking, under the tab "DIY Projects and Ideas,"
  • In the "How to Caulk," page, watch a how-to video to see how caulking should properly be done, and then scroll down to easily view what supplies are needed for the project.

Instructions are clearly outlined with a photo step-by-step. It’s an undaunting way of setting up a visitor for a successful project - without ever having to leave home.

Home Depot DIY Page.pngThe Home Depot's "How to Caulk" Instructional

What If?

The Home Depot could harvest (at scale) the vast amount of intent data that such (engaging) content produces. We’re talking valuable interest data. Based on how visitors engage (and don’t engage), a marketer can anticipate the triggers that will be effective once the visitor is considering a purchase. Here are some ideas on how The Home Depot could potentially leverage the data to turn initial engagement into (later) revenue.

  • Collect and store data. First, let’s store every interaction with a visitor to understand their behavior. Imagine that we’re collecting data from the "DIY Projects and Ideas," page instructing "How to Caulk." Based on a visitor journey and via predictive modeling, we can tell that a visitor is highly engaged with the Bath Projects section of the website in a marketer quantified way - they’ve watched at least 2 videos over the last week covering How to Caulk, and Caulk and Sealant, and are predicted to browse caulking supplies in the future.
  • Create visitor segments. Using a platform like Intempt, where we blend past behavior (fact) and future behavior (predicted), we can specify a target group. For example, we create a segment that contains visitors who have watched Caulking videos at least 2 times (behavioral property Count of) after browsing the "Bath Projects" section (behavioral property Has Done), and are predicted to not add a product to the cart (predictive property Will Not Do Add To Cart).

Home Depot Segment-1-765420-edited.png

Creating a segment inside the Intempt Platform

  • Set up a notification campaign to target currently active visitors from our target segment with notifications at key drop-off points. In this example, a visitor may not get into the segment on their first visit of the “Bathroom Projects” section. They might make it after a few visits. What happens next? If they return to the store area and browse supplies, they meet a trigger event (for a campaign) which gets the notification out. Here’s what it could look like:

Home Depot Store.png

Simulated Scenario - Predictive Notification

 

 

 

Home Depot Store zoomed.png

Zooming In

Why This Notification at This Time?

Because the visitor’s past browsing behavior, purchase behavior, and their current context is indicative of what will happen:

  • They watched 2 caulking videos, so we use the suggestion for caulking supplies in the notification to cater to visitor's interests.
  • They have browsed the Bath Ideas section, indicating there may be an interest in bathroom remodeling supplies, so we suggest to pair the current items.
  • The platform predicted that the visitor is not going to add an item to the cart (Will Not Do Add to Cart predictive property), so the “Free Shipping” offer appears to encourage the visitor to proceed with a purchase.

Keep in mind, this isn’t blanket targeting that clutters the experience. They are targeted with a personalized message based on the project of their expressed interest. Additionally, they are offered a shipping promotion since the visitor is predicted to not add these items to the cart based on their browsing behavior, and potentially (the model knows) their past purchasing behavior.

Use Case 2 - Birchbox


Subscription eCommerce can have its cake and eat it too. Beauty and grooming product companies in particular, benefit from a direct-to-consumer acquisition framework and a recurring revenue stream. Consumers enjoy a personalized experience. Everyone wins. You know this segment has promise since the beauty industry is witness to an eye-popping billion dollar acquisition.

Effective marketing technology is crucial to the success of the subscription e-commerce model. Marketing technology needs to aim at the biggest challenge that these companies face on a recurring basis: acquiring subscribers and keeping them engaged through the subscription cycle so they can lower the rate of “churn,” or membership cancellations.

No wonder AI-based marketing has proven to be a subscription e-commerce friend. If a company relies on marketing that accurately predicts what products a customer will like, and engages with the customer in a real-time manner to cater to that discovery, then the probability of cancellation diminishes considerably.

As a marketer, you can actually understand your subscribers and articulate what they want better than they ever could. How? By going Predictive.

Birchbox is perfect for window shoppers who enjoy endlessly browsing beauty and grooming products, but can't quite make a decision on which product to get. Birchbox makes it easy and affordable to enjoy variety without having to make a decision.  Visitors can either shop products or become a subscriber. Subscribers receive their monthly box of 5 samples every month. The company created a buzz in the beauty industry with their monthly boxes, made popular by well known influencers, such as Emily Schuman of Cupcakes and Cashmere, who has a following of 300k+ on Instagram. The subscription comes in the form of a monthly / yearly subscription.

Birchbox home page.png

Birchbox Homepage

By subscribing, members have access to 5 beauty samples that include makeup, hair and skin care. Subscribers receive points redeemable at the online Birchbox store (available to be shopped by non-subscribers, as well) when they add full size products to their boxes, and after the second month, they can begin to choose the samples they want to receive.  

Birchbox Explained.png

  Birchbox Perks

What If?

Let’s explore an idea of how Birchbox could acquire new subscribers. We know visitors browse around 2 pages per visit, and visit more than once, on average.  We are looking to create a campaign that will leverage the power of product interest by converting window shoppers into subscribers, and prevent Birchbox from losing money on retargeting advertising. Here’s what we do:  

  • Collect data. First, we collect the data of a curious visitor who is browsing in Birchbox, leaves without subscribing, and returns for the typical second visit. Using a platform like Intempt, we specify a tracker for the domain to understand the visitor's behavior from the very first visit, store every interaction, and leverage the data in the future campaign.

Birchbox Tracker.png

Creating the Birchbox tracker

 

  • Identify a drop-off point. We’ve assumed that visitors leave the website after viewing 2 pages. To back our hypothesis with data, we slice two narrowly defined segments out of the funnel and compare how many visitors have fallen into each segment. We create a Segment 1 for visitors who browse 1 page (behavioral property Count Of) and a Segment 2 for visitors who browse 2 pages (behavioral property Count Of). Here’s what it could look like:

comparing_segments.png

Segment 1 audience vs Segment 2 audience (marked in red)

Based on the gap we noticed between the two segments, we know a visitor is likely to leave, so we prevent the drop-off with a targeted notification.

  • Segment our visitors. We specify a target group by blending past behavior (fact) and future behavior (predicted). For this segment, we are focusing on first-time or returning visitors who haven’t subscribed (behavioral property Has Not Done), have browsed 1 or more pages (behavioral property Count Of) and are predicted to drop off without subscribing (predictive property Will Not Do.)
  • Time to campaign! For this campaign, when one or more pages are reviewed, we know the possibility of a drop-off is high based on the average number of pages per visit. It’s a good time to send a notification inviting the visitor to check out the Birchbox subscription. Final look:

 Birchbox Store.png

Simulated Scenario - Predictive Notification

Why This Notification At This Time?

  • The visitor has engaged with at least one page;
  • The visitor is new to the subscription model;
  • The visitor is predicted to drop off without becoming a subscriber.

In this scenario, we’ve leveraged a visitor’s intent of exploring new products and used that intent to convert them into a subscriber.

Use Case 3 - The North Face

3% of people generate 90% of the impact online, such as views, linkbacks, likes, etc.. Using “Influencer Marketing” to leverage the power of those 3% is a valuable strategy for anyone looking to accelerate reach. As consumers are becoming immune to traditional digital advertising, they are demanding a more authentic medium that marketers haven’t turned into a growth machine. Yet.

94% of marketers who use influencer marketing believe it to be an effective strategy. Search volume for "influencer marketing" has steadily climbed in the last few years, according to Google Trends.

Here’s the expected catch - 78% of marketers said that determining the ROI of influencer marketing was a top challenge in 2017. While some of influencer marketing benefits are clear - it drives engagement and brand awareness - we’re still challenged in making campaigns successful in terms of hard metrics - conversions and revenue.

Let’s look at The North Face and their influencer marketing. On the “Athlete Team” section, the visitor can select an athlete’s profile that inspires them to run/climb a little “extra”. The visitor checks out the athlete’s motivation, expeditions, videos and - most importantly - the gear they are using. A visitor will convert naturally due to the ease of product placement and flow to purchase funnel.

NF3.png

The North Face "Explore" Section

What If?

How could The North Face harvest (at scale) the vast amount of intent data that such engaging influencer marketing produces? Based on how visitors engage (and don’t engage), a marketer can anticipate the triggers that will be effective once the visitor is considering a purchase. Here are some ideas on how The North Face could potentially leverage the data to turn initial engagement into (later) revenue.

  • Collect and store data. First, let’s store every interaction by a visitor to understand their behavior. Imagine that we’re collecting data from the “Explore: Athlete Team” page dedicated to Alex Honnold. Based on a visitor journey and via predictive modeling, we can tell that a visitor is highly engaged with the “Climbing”  section of the website in a marketer quantified way - they’ve watched at least 3 videos covering athletes' expeditions over the last week and are predicted to browse climbing products in the future.
  • Create visitor segments. Using a platform like Intempt, where we blend past behavior (fact) and future behavior (predicted), we can specify a target group. For example, we create a segment that contains visitors who have watched Alex’s expedition videos at least 3 times (behavioral property Count of) after browsing the “Climbing” section (behavioral property Has Done), and are predicted to not add a product to the cart (predictive property Will Not Do Add To Cart).

segment_small.png

Creating a segment inside the Intempt Platform

 

  • Set up a notification campaign to target currently active visitors from our target segment with notifications at key drop-off points. In this example, a visitor may not get into the segment on their first visit of the “Explore” section. They might make it after a few visits. What happens after? If they return to the store area and browse gear, they meet a trigger event (for a campaign) which gets the notification out. Here’s what it could look like:

NF2final.png

Simulated Scenario - Predictive Notification

   

nf6.png

Zooming In

 

Why this notification at this time?

Because the visitor’s past browsing behavior, purchase behavior, and their current context is indicative of what will happen:

  • They watched 3 videos of Alex’s expedition, so we use the athlete's name in the notification to cater to the visitor's interests.
  • They have browsed the Climbing section, indicating there may be an interest in climbing gear, so we suggest to pair the current item.
  • The platform predicted that the visitor is not going to add an item to the cart (Will Not Do Add to Cart predictive property), so the “Free Shipping” offer appears to encourage the visitor to proceed with a purchase.

Keep in mind, this isn’t blanket targeting that clutters the experience. The visitor is not exposed to a wide range of climbing pros. They are targeted with a personalized message based on the athlete with whom they have an affinity. Additionally, they are offered a shipping promotion since the visitor is predicted to not add these items to the cart based on their browsing behavior, and potentially (the model knows) their past purchasing behavior.

To Recap


In order to enable marketers to build predictive experiences, the best personalization platforms must be:

  • Data-Ready

AI-driven predictive marketing tools that offer native data-capture to save time and resources by eliminating the need for ETL. This means your site or app data is ready to process.

  • Modeling-Ready

Machine learning needs to be built into the fabric of your platform, rather than being something you have to bolt on.

  • Notification-Ready

Eliminating the need for engineering and operations teams to get involved (to make changes to your site and apps), models process and notifications deliver, automatically.


With AI based predictive marketing tools, e-Commerce companies can drive revenue - and rest assured that intelligent data tracking, modeling and notifications are built into every consumer interaction. Set up a discovery call to see predictive marketing platform in action.