CASE STUDY

Streamlining Workflows for Adobe AdCloud Users

Adobe Advertising Cloud for Search is Adobe’s platform for search marketing management. AdCloud provides marketer's with the tools needed to manage and optimize campaigns that hit performance goals.

As an Adobe employee, AdCloud is one of my most utilized products. When Adobe began prioritizing Generative AI this year, including the launch of Sensei for the Experience Cloud, I was surprised to see AdCloud excluded from the product rollout.

There is an opportunity for Adobe to build out Generative AI capabilities across all solutions, including AdCloud, and brand themselves as consistent leaders in the ever-evolving tech space.

TIMEFRAME

6 weeks

TOOLS

Figma

MY ROLE

UX Designer

UI Designer


THE PROBLEM

  • High volume of companies competing to launch the next big AI product

  • Adobe's AdCloud has not been included in Adobe's generative AI rollout

  • Daily task management in AdCloud is very manual and time consuming

THE SOLUTION

  • Design an AI-backed feature integrated into Adobe’s AdCloud platform

  • Build a new AdCloud feature that improves user workflows in AdCloud

RESEARCH

How do marketers use AdCloud to optimize campaigns?

Understanding AdCloud super users

To kick off the research phase I recruited five Adobe colleagues that are also Marketing Managers. All five participants currently use AdCloud and have tried AI technology before.

Interviewing marketer’s with a mix of seniority and experience, I was keen to understand what aspects of campaign management they owned and how they incorporated AdCloud’s existing feature set.

β€œIf there were any AI tools to help with reporting, i’d jump on it.”

β€” KN, Growth Marketing Manager

β€œI mostly use AdCloud for pulling reports, and then i’ll review the data to map out key insights before creating an optimization plan”

β€” JB, Senior Marketing Manager

β€œMy team spends 6+ hours a week pulling manual reports from AdCloud into excel - looking for insights is very time consuming”

β€” AR, Search Marketing Strategist

Marketing tasks remain very manual

Data analysis proven the most time consuming task

As a former digital marketer, I know how time consuming campaign optimizations can be. I wasn’t surprised to find that most of my colleagues felt the same way. 

The majority of participants struggled the most with analyzing data and synthesizing insights. This task is completed multiple times a week and is very time consuming, leaving less time to innovate.

The majority of participants were also looking for a better way to QA campaign errors and track changes. This is a valuable part of campaign management that often gets overlooked.

Last, all participants were excited about the idea of using AI technology to supplement their work.

View Affinity Map β†’

4 out of 5 participants say reporting & analyzing insights was the most time consuming part of the optimization process

5 out of 5 participants want to review all AI recommendations first before implementing

5 out of 5 participants were excited at the idea of using AI to supplement their work

4 out of 5 participants say reporting & analyzing insights was the most common optimization in AdCloud

Who benefits the most from AI-enhanced AdCloud?

Developing junior and senior-level personas

Based on the research findings, I developed two personas. These personas represent two types of marketers who use AdCloud - a Manager and an Analyst. Managers drive the strategy and collaborate with stakeholders, while Analysts drive the day-to-day campaign performance and optimizations.

Persona 1

The Manager, Strategist

Persona 2

The Analyst, Insight Finder

Persona user journey mapping

Visualizing a day in the life

Mapping out the user journeys for each persona was helpful in identifying their largest pain points. For both personas, easy access to campaign performance and insights were essential to their daily roles. Managers need to report performance back to their stakeholders, while Analysts need to make accurate campaign optimizations using data.

Persona 1

The Manager, Decision Maker

Persona 2

The Analyst, Optimizer

How do competitors use AI technology?

Understanding the competitive landscape

To go a step deeper, I analyzed four competitors that offer AI-based solutions for the most frequently mentioned needs (i.e. data visualization, ad creation, keyword-content mapping, keyword recommendations).

STRATEGY

Dual focus

Designing for the entire team

It became apparent after the research phase that better reporting and optimization tactics were the most central need across all marketer’s. β€˜How might we’ statements were created with both analyst and manager personas in mind.

How might we help analysts build better reporting processes that generate more impactful insights?

How might we improve manager optimization strategies to drive stronger campaign performance?

Key Features

Defining the MVP

After reviewing the research results with the team, we spent some time ideating on features. It was important to enhance the existing feature set by brainstorming on new ideas. After gathering the list, I started prioritizing what features were necessary to launch the product.

Must Haves

The core set of features I would need to design

Data Filters

Report Dashboard

Notifications Page

AI Chat Box

Change Log

Nice to Haves

Additional features that are build in addition to the core feature set, timeline permitting

Edit Report Insights

Share Access Button

Share Report Button

Reporting Templates

Instructional Pop-Ups

Minimize Reports

How will users complete key tasks?

Building three primary task flows

Tasks flows were built out for three core tasks. It was important to understand when and where a user would interact with the primary feature - an AI chatbot. Defining what a GenAI chatbot could and couldn't do within AdCloud was essential to staying within the project scope.

Ultimately, I decided that AdobeChat could be able to create and change both notifications and reporting views. These aspects of AdCloud are very rules-based and would be realistic for GenAI technology to actually implement.

View Task Flows β†’

REVIEWING NOTIFICATIONS

CHECKING IN ON PERFORMANCE

SHARING REPORTING

DESIGN

Ideating on feature layout

Low Fidelity Wireframing

Sketching out the core pages, I tested multiple layouts for the must-have features. Experimentation with filters, tabs, notifications, and icon type ultimately let to selecting the layouts that were easiest for users navigate.

Building the layout

Mid-Fidelity Screens

Three core tasks flows representing the most essential features were mapped it. I incorporated a combination of the low fidelity elements to create the most harmonious layout that best aligned with the style guide.

View Mid-Fidelity Screens→

REIVEWING & EDITING NOTIFICATIONS

REVIEWING & EDITING REPORTS

V1 Designs

High-Fidelity Screens

Testing

In-Person Usability Testing

Validating the design with usability testing

The goal of testing this prototype was to understand how users would utilize a generative AI chat assistant to customize notifications and reports. Five Adobe employees and AdCloud super users were taken through two key tasks;

  1. Check-In on Notifications (Review, Remove, and Add Notifications)

  2. Check-In on Performance (Review, Update, and Share Reports)

Positive Usability Feedback

Analyzing the test results

While one user did not successfully complete all of the tasks without error, overall  product feedback was extremely positive. All of the participants found AdobeChat enjoyable to use and expected the feature to reduce their manual workflow significantly.

4 out of 5 users completed both tasks without any errors

5 out of 5 users found the AdobeChat feature enjoyable to use

5 out of 5 users believe that using AdobeChat would reduce their time spent on manual task work

Optimizing for Visibility

Frequency to Severity Mapping

Using a frequency to severity map I was able to organize participant feedback and prioritize iterations I would make. The more frequent and severe the feedback, the more likely I was to implement the iteration.

View Frequency-Severity Map β†’

Making Iterations

Naming Conventions

The change log title needed to be simplified to help users clearly understand the features purpose

Automated Instruction Pop-Ups

Instructional pop-ups were added to key screens to inform users of the new AI chatbot feature and teach them how to use it

Saved Notifications

The saved notifications were added to the Change History to be able to track previous and existing notifications in one place

Reporting Metric Drop-Down

Create an option to select different reporting metrics to provide easier report customization

Edit Icon

The edit icon was added to the reporting insights to help users quickly edit and customize their reports

Final Prototype

Key Learnings

Users want to be informed at each step of an interaction

  • When a new technology is released within a mature product, like AdCloud, it can often be overshadowed by the existing feature set. Long time users will stick to their routines and are less likely to test out a new feature. This was a huge insight identified during usability testing that led to the addition of pop-up notifications for clearer education and direction.

Future Considerations

Because this project was completed in a 6-week sprint, there are a few things I would do differently with more time and resources

  • Align with Adobe employees working on the GTM strategy for GenAI features to better understand the company priorities and refine the feature roadmap

  • Interview a larger set of digital marketing professionals, outside of Adobe employees

  • Perform another round of usability testing with the post-iteration screens