About RealSelf

RealSelf is a leading healthcare marketplace for people to get factual, unbiased information on cosmetic procedures. Their mission is to create a world where every investment in modern beauty is worth it. RealSelf aims to be the go-to advisor for medical and aesthetic procedures by making the experience easy, personal, and trusted.

If a user has an aesthetic concern they’re looking to get fixed, they can browse RealSelf’s platform by concern or procedure, and then get connected with a physician to begin a consultation. RealSelf also works closely with the physician, offering a variety of opportunities—like featuring them on their website or showcasing them within their emails. 

Due to an unwieldy and ineffective in-house AI model that led to a decline in engagement and high opt-out rate, RealSelf was in need of a more reliable, automated solution that would leverage AI to deliver their messages at the most optimal frequency.

The Challenge: The Ugly Side of Low Engagement

RealSelf relies on user-generated content (UGC) to drive sessions—which is why customer reviews make up 95% of their emails. Sessions are their biggest revenue driver, making UGC crucial for the bottom line. A session is counted when a user logs into their platform, both on the website and app. The goal is to increase sessions, with the hopes of increasing the number of contacts who book an appointment with a physician—a key conversion for RealSelf. Since June 2021, the company had seen a steep decline in sessions due to low email engagement as a result of a cumbersome and unsuccessful internal AI algorithm that determines email send frequency.

RealSelf’s AI Needed a Facelift

RealSelf had built an internal AI algorithm that determined how frequently content would be sent based on user engagement with their emails and on their website. However, this model was time-consuming to manage and didn’t prove to be successful with improving their user engagement and ultimately driving more sessions.  

One part of the algorithm involved excluding users with Microsoft Outlook—a notoriously difficult email service—who haven’t opened an email in greater than 30 days. Another part was excluding users based on a lack of email engagement—any user who has not opened an email for 180 days or more would not receive these emails. However, with Apple Mail Privacy Protection changes that went into effect in the summer of 2021, they could no longer rely on email opens as a predictive metric for engagement. What’s more, these exclusion lists became complex to independently manage for every campaign.

Solidified Send Frequency with Iterable AI

To resolve the challenges with their AI model, RealSelf turned to Iterable’s Brand Affinity, which uses AI to automatically calculate a loyalty score based on a user’s recent interactions across their marketing channels, and converts these into affinity labels tied to each user profile. For example, users labeled as loyal are those who are highly engaged with RealSelf and can be considered brand advocates.

Using these affinity labels, the RealSelf marketing team gains a unique advantage by intelligently translating user engagement data into deeper, more meaningful connections. This helps them maximize retention with loyal customers, nurture customers through lifecycle stages, and mitigate churn with at-risk detractors.

RealSelf has replaced their in-house AI model and now uses Brand Affinity to determine how many times per week their users will receive their UGC emails in hopes of improving email engagement to drive more sessions and conversions. 

For instance, the RealSelf team segmented their loyal and positive affinity cohorts as well as any new user to receive UGC emails seven times per week, or one email a day. The neutral group—those who are sometimes engaged and who occasionally interact—receive UGC emails four times a week. The negative group—those who are typically disengaged—receive the emails only once per week. RealSelf then allowed Iterable’s Brand Affinity model to auto-adjust the frequency of sends within these groups on a rolling weekly basis based on real-time engagement levels. 

Using Brand Affinity, RealSelf was able to reduce their opt out rate with their neutral group. Iterable AI showed this group was underperforming (compared to other groups) so they reduced the send frequency of this group from 4x a week to 3x. Within one month, the opt outs declined by 33%. Prior to Brand Affinity, RealSelf could not make quick send frequency adjustments to a specific group due to potential data flow issues with their in-house algorithm. 

Within the first month of launching Brand Affinity, RealSelf has seen sessions and contacts increase by 30% and click-through rate increase by 13%.

Iterable Helps RealSelf Put Their Best Face Forward

With the help of Iterable AI, RealSelf spends significantly less time and effort reconciling and managing their user engagement data, and has completely eliminated the need for their exclusion lists, leading to better utilization of resources. To illustrate, RealSelf is saving their data engineers about two hours each week when downloading their user engagement to their data warehouse. Furthermore, RealSelf’s marketing team has been able to completely eliminate back and forth internal messaging and meetings about possible data flow issues and setting user frequency of emails.

What’s Next for RealSelf?

Using Brand Affinity, RealSelf plans to launch a re-engagement journey for users who move from loyal to negative in an attempt to re-engage them and mitigate churn risk. 

Individualization is Key

RealSelf also plans to launch phase two, which will include implementing Iterable Catalog with Brand Affinity to send more relevant review content based on individual interests and behavior.  With Catalog’s easy-to-use Collection Builder, RealSelf will be able to drive more conversions with intelligent content and automate millions of unique messages across channels.

For example, RealSelf collects customers’ birthdays. With Catalog, and this customer data, they can promote popular cosmetic treatments to get in your 30’s, 40’s, 50’s, and more. Additionally, RealSelf can match physicians to users based on cosmetic category of interest and geo-location. A California-based plastic surgeon specializing in rhinoplasty would be matched with a user who is interested in a rhinoplasty procedure who resides in California. The RealSelf team will also individualize content based on where the user is within the journey—from research phase to procedure and beyond.