| Led our platform’s implementation of personalization

A little
more personal

The win:

Built a hyper-personalization feature that let users launch personalized A/B tests in 7 minutes, down from 10–14 days.

My Role:

Director of Product Design:
Player / Coach

Stakeholders:

VP of Product
EVP of Engineering

Timeline:

2.5 months

Team:

Myself
Product Designer, Platform
Senior Product Manager, Platform

00.

Prologue

For the best context of this platform case study’s, please visit the prequel.

Dugout is Centerfield’s proprietary software that manages our end-to-end customer journey to acquire new customers on behalf of major brands.

Within Dugout, we built the capability to personalize these website experiences in real-time with backend customization.

01.

Testing typically took two weeks from design to dev.

We wanted to now make personalization simple, at scale, for our internal users by
building the functionality into
the platform:

02.

First was the fight for the ideal solution vs. a band-aid that’d deepen tech debt.

I advocated for Domains as the intuitive entry point — even though Clusters were easier to surface from the legacy structure. The legacy path posed real abandonment risk.

03.

With the origination sorted, we got to work on our first draft of building personalization within an experiment:

Step 01.

Step 02.

Step 03.

Step 04.

04.

The trigger feature lived within our rebuilt experiment set up and needed a target audience defined for personalization, so we did usability with media buyers:

I led the prioritization of findings across six themes for the total experiment set up with the PM and platform designer, where we focused on our most critical vulnerability that that video shows: rules targeting.

4/5

subjects truggled to understand the rule set up

The problem, visualized:

3.4

minutes is the average time it took for users to just establish rules

05.

Facing implementation constraints and reduced timelines, we simplified the UX/UI and rethought the personalization flow architecture.

Step 01.

Step 02.

Step 03.

Step 03B.

Step 03C.

06.

The results:

2:50

we reduced the time spent building rules by ~1 minute

4/5

completed an experiment setup without moderation

-30%

we reduced the time spent building an experiment from 10 min to 7 min

Visit the prequel

Or, learn about how I