According to a study by A/B testing and personalization platforms Kameleoon and Forrester Research, medical brands that implement A/B testing and personalization are 5 times more likely to report faster growth than brands that do not implement A/B testing and personalization . What hinders the development of many brands is the “three Vs”.
- Volume: They feel overwhelmed by the amount of data
- Velocity: Data is coming at them too quickly
- Variety: There are too many types of data
The result? Despite 8 out of 10 healthcare brands saying that they will improve the use of data and analytics to create better patient experiences, only 2 out of 10 do so.
Table of Contents
- analytics and optimization tools
- android keyword installs
- buying android app installs
- google play store ranking
Healthcare brands also struggle to understand the relationship between data type, regulatory risk, and optimization.
The study above also shows that nearly 50% of healthcare brands believe using anonymous data for UX optimization is risky or extremely risky.
All of this sounds very familiar to me. As guardians of some of our most personal data, healthcare brands are right to make data privacy and security a priority. But I’ve learned through my work at Winona that HIPAA compliance does not need to come at the expense of creating a better patient experience.
Winona is a telehealth company that provides anti-aging and wellness treatments to women based on doctor-prescribed medical advice. Like many brands, we set out to deliver exceptional experiences to our customers and be data-driven from day one. However, like most companies, especially those in regulated spaces, we struggled to know where to start.
Management was fearful of regulation. Everyone was busy. When we talked about optimization, our conversations jumped to the most sophisticated ideas, which quickly went nowhere.
Fast forward to today. Through trial and error, we’ve learned that by being customer-centric we can easily measure and optimize patient experience. Our investments in analytics and optimization solutions are being put to good use.
Here are 6 actions we’ve learned can help healthcare companies get the most out of their martech stack.
1. Use HIPAA-compliant martech solutions and services
In addition to HIPAA compliance, work with solutions that can sign business associate agreements (BAA). Rather than switch to a HIPAA compliant solution later, as soon as you are serious about improving conversion rates and engagement, start with one that can sign BAA immediately. You don’t want to start with one that can’t process protected health information (PHI) or personally identifiable information (PII).
There are HIPAA-compliant agencies with strong data security practices to enlist, too. Any agency that will be touching your data also needs to follow HIPAA-compliant practices and should also sign a BAA.
2. Take advantage of available integrations so your entire martech stack works for you
You have the technology at your disposal. Take advantage of its full capabilities by leveraging available integrations. At Winona, we create “cohorts” of customers in Mixpanel, our product analytics platform of choice. Using the Mixpanel-Kameleoon integration, we target cohorts with A/B tests and, where relevant, personalized experiences. Kameleoon sends the test performance information back to our dashboards in Mixpanel in real time.
In the screenshot below, the users who saw variant #2 in the Onboarding A/B Test saw a significant lift in both monthly active users and revenue generated from that group compared to the control group. No one had to pull a report, work late, and make a presentation deck explaining the business impact of our optimization.
3. Start with a simple opportunity affecting your target customer
To get the most out of your analytics and optimization solution, focus on a challenge/opportunity that your primary target customer is having. Start at the end by working out what the ideal solution would be for that customer.
At the beginning of our optimization journey, we were excited to do omni-channel personalization, where every user would receive a personalized experience, but that’s very complicated and caused us to stall on execution.
Instead, we decided to start with a much simpler opportunity to implement and provide immediate value: a classic funnel optimization.
Through our analytics, we knew that we were experiencing a significant drop off (60%) when patients had to upload an image of their ID (what doctors need to see in order to speak with our customers), so we decided to A/B test new ways to optimize this step in our funnel. In one of our variations, we added a “Skip For Later” button. Within 3 days, we found out this variation created a 9% lift in our conversion rate.
This is the moment the team realized we didn’t need to leap to advanced personalization to increase conversion. Instead, we just needed to continually be making small optimizations that helped out our customers.
4. Use simple and consistent data naming conventions so that everyone is talking about and looking at the same thing
If a company creates an event called “booked appointment” in their analytics tool, but refers to it as “booking” in their optimization tool, you’ll set yourself up for confusion, especially as your company grows and changes.
It may sound minor, but to help you get the most out of your martech solutions, naming conventions must be consistent. If not, the team could feel like they are tracking twice as much data than they actually are. You’ll constantly face questions about data accuracy, wondering which platform is the “source of truth.” At Winona, we wanted to prevent this issue by ensuring everyone was aligned on what our data meant, how it was collected, and—most importantly—how it was named. To do this, we created an internal data blueprint document called a tracking plan.
Our document listed out all the events, their triggers, and the metadata associated with each event. With any martech tool used, we ensured that the data integrated into the tool strictly followed the naming convention as listed out in the document.
The effect was tremendous. It allowed members within our team to have more informed conversations about data, it removed confusion, and it allowed our team members to move faster when switching from one tool to the next.
5. Anonymize as much data as possible
The best way to secure a patient’s data is by anonymizing it. While initially all our A/B tests relied on anonymous first-party data, we built a practice of anonymizing patient data even if we didn’t immediately need to in order to solve a UX challenge affecting our conversion rates. This quickly paid off as we later began to target cohorts with A/B tests that relied on their geolocation—i.e. information that could personally identify them.
We anonymized using Unique User Identifiers (UUID). Unique User Identifiers or UUIDs are a system-generated set of letters and numbers that are produced and attributed to a user when a user creates an account. If a user ever switched devices, we would set up the implementation for our data and optimization tools to recognize and identify the user’s UUID when they logged in.
6. Use analytics to find out who is NOT having the intended experience
At Winona, we make it a priority to watch out for users who might not be having the intended experience.
In our case, an unintended customer experience comes in many forms. The most common cause is when a patient does not read an important message from their doctor. Often, this prevents the patient from receiving treatment—i.e. our product.
Using Mixpanel’s cohorting feature, we can track cases like this by easily identifying users who took certain actions, as well as users who did not take certain actions.
For example, we pulled together users who received an important message but failed to ever open the message.
This dynamic list of users is then shared with our customer care team to follow up with the customers to get them to check their outstanding messages and offer a better patient experience.
We also A/B tested new ways to get the number of missed messages to decline.
The performance of the tests fed directly into our analytics tool, so we knew if we had solved the problem.
No one says they don’t want to build a data-driven, customer-centric culture
What often gets in the way of being data-driven and customer-centric is the misconception that all data is equally valuable or risky. It isn’t. Customer data has value when you’re able to apply it to solve a customer problem. By starting with a clear opportunity to improve a customer experience, healthcare brands can tune out the cacophony of data signals and focus only on the data they need to improve one specific customer experience. By practicing customer-centricity, and using HIPAA-compliant tools that integrate in impactful ways, healthcare brands can steadily optimize any digital experience and rely on science to guide their decision-making.
What healthcare brand wouldn’t want this?