Pioneering Caring.com's conversion optimization platform

Assessing assets from the subscription model

As mentioned all of Caring.com's previous concent marketing was educational. "Click to find personal care for your loved ones," "5 ways you can help a family member with dementia."

While this information is definitely helpful, it doesn't drive clicks or action. You can Google this information for free.

My idea was to create clearer call to actions that would drive clicks and conversion. Instead of click to receive personalized care, the message should have a value proposition. More on that later.

Building a conversion funnel from scratch

Caring.com had never conducted an A/B test for any of their landing pages, emails, or software UI. My role extended far beyond just designing landing pages and emails. I surveyed testing platforms, colaborated with engineering, and built systems to track new leads.

Collaborating with Engineering to create the best A/B testing solution

We conducted tons of resarch on tools available and concluded that there was no way to conduct a clean test with tools such as Optimizely. At that time, Optimizely used javascript to control image and copy tests. There would be a flash that would display tested content over the other. How would we know what the user acted on to convert? Image test A or B?

Our solution was to build an in-house URL splitter where we could route specific percentages of traffic to different urls. We could also control which segment, service, (AL or in-home-care), or demographic goes to which landing page so we could build refined pages for different cohorts.

What deserves to be tested?

I was used to getting a directive from my manager to "come up with a bunch of designs and we'll rifle them through our tool". This time around was very different. Andy and I were very calculated about what we wanted to design and build so we woudln't expend resources on tests that may potentially fail.

We assigned an optimum cost per click based on channel, and what it would cost for them to move down the funnel. This would help us determine what to build for which segments, and at what point in the funnel it would be best to spend design and engineering resources.

Time to test, (finally)!

After several sprints of Product Management and setup I was aching to design something.

The higher-ups were apprehensive to go against the grain, even given the user research. We would test two variants against the control.

Control

• Headline: Get FREE Assisted Living Information
• Image: Lavendar Flowers
• Form fields: Long list

Test A

• Headline: Get FREE Assisted Living Information
• Image: Seniors having fun
• Form fields: Long list

Test B

• Find Assisted Living Costs in {location}
• Image: A map of AL homes near them
• Form fields: 2 step flow. Step 1 would capture zip code, who they were looking for, and email address. If they abandoned the page on step two at least we could send an abandonment email.

The landing page I designed and built (in HAML) was a landslide victory. The previous landing page had a conversion rate of 6.64%. The new landing page, deemed "Costs" Had a 12.6% conversion rate.

Moving into SaaS and Product Management

This was an exciting time for me. After a few years of focusing on conversion I was granted the opportunity to dive into full-stack product design. The CEO and Director of Product and I would have quarterly planning meetins to disucss high-level business objectives. It was up to him and I to define and build viable products to solve these problems.

I love product management. I was involved in everything from identifying the epic, breaking it down into stories, pointing these stories, and writing specs for engineers and even myself, since I was still the primary UX/UI designer. Below is a summary of some of the products I worked on.

The Partner Portal and Move-in Reporting Tool

• View leads complete with customer profiles, contact history, and contact information
• The ability to update their rates which would update live on our providers page
• The ability to report move-ins

An entire call center! (We called it Contact)

Our current software, 5-9, was buggy and had frequent outages. It also did not link directly to our in-house CRM, talk. We wanted to add a Salesforce-like layer to expedite the Agent workflow. This product was 6 months in making. It had the agent ui where each sales individual could track their calls, set dispositions, and status, and more impressively, the manager ui, where managers could track things like:

• Agent state - online, offline, on break
• State duration - how long has the agent been on break or on a call
• Queues - what channels agents were currently assigned to
• Tactics - assign a method to the agent based on seniority, cold calling, follow up, etc
• Skills - the ability to assign agents to specific traffic channels