Empower advertisers to make the right performance decisions
Mid-Flight Recommendations for advertisers for better ad performance
Company: Facebook
Facebook Ads Business Products
overview
I spearheaded User experience on ads business product, from conception to launch, coordinating all UX Design related efforts, leading the project to collaborate with various cross-functional teams, while placing the user at the core of the product design lifecycle.
Success Metrics
We shipped high impact solutions for advertisers accomplishing 1.00% top-line StAV lift (Standardized Ads Value) and 0.87% revenue lift which is approximately 600MM incremental revenue annually.
platform
Desktop App
Project timeframe
3 months(Oct 2020- Dec 2020)
My Role
Senior Product Designer
Team
Content Designer, Product Manager, Data Scientist, 3 Engineers, and Product Designer (me).
Tools Used
Proton (FB design tool) and Figma.
CONTEXT
Mid Flight Recommendations (MFR) alert advertisers to problems that are currently impacting their performance and provide them a clear actionable way to resolve them.
Auction Overlap is one of the MFRs when multiple ad sets in an ad account have the same object and similar audiences, causing them to enter the same auctions for the same people, as a result some of the ad sets are shown less and may get fewer results.
PROBLEM OVERVIEW
The current design is 50% lower than the “control” in adoption rate.
As per user research,
Advertisers often fail to both diagnose and fix problems that impact their performance.
Lack of knowledge of actions they can take to improve their performance.
Don’t know or trust the value of taking actions on recommendations.
HYPOTHESIS
Currently there is a “Control” version and “MFR Framework” version running as a test. MFR framework is a consistent design for Ads Manager.
The Hypothesis is to test another variation of Auction Overlap in MFR Framework, concurrently with the existing design by improving the resolution flow to encourage higher adoption rate.
Project goal
Improve Auction Overlap design so that the new design variant matches the adoption rate of the “Control” version.
aligning user goals with business Goals
Advertiser goals
Improve performance of their campaigns (primary).
Help them set expectations for future performance.
Help them prepare for future disruptions, ex. signal loss.
facebook goals
Make advertisers their trusted partner.
Help advertisers make their business successful.
the design process
Our team followed the Agile Scrum format with daily scrum meetings, and we went through various MVP cycles to come up with the right design solution. Below is my design process that I followed for this project.
Empathize and Discover: I teamed with product manager, data scientist, and content designer to strategize accurate and deep intuitive understanding of the user and translate concepts into features that addressed user’s information needs and behavior.
User Interviews: I attended user interviews (performed by UX Researcher) to gather and understand user requirements and constraints. I collaborated with Product Marketing Manager and Product Design leads from Ads manager team, to gain insight into advertiser behavior and expectations. I consulted the user researcher to understand advertiser pain points.
Hunt for Data Source and Utilize Metrics and collaborate with diverse teams: I collaborated with Data Scientists to get adoption segmentation analysis, and key metrics to track so we could gauge the success of the project. I also collaborated with partner teams to understand advertiser segments.
Strategy, Vision, Ideation and Iteration: I executed design workshop, with different cross-functional team members and partner teams to define strategy and vision.
Testing and Validation : We tested and validated the designs early and often during various stages of the design process.
Who are we solving for?
There were a number of stakeholders involved in this project like Executive Sponsor, Legal, Operations, and Client Management. Stakeholder feedback helped understand the Business limitations and goals that I used in designing the product for this project.
Adoption Segmentation analysis
Our recommendations serve most common verticals on Ads Manager (Retail, Professional Services, Entertainment) as well as Ecommerce.
Factors associated with higher adoption:
Locations: Vietnam and Brazil
Unmanaged
SMB
<1 year tenure
Ads using on-going schedule
Tail/Torso
Findings:
4% of guidance cards are completed using the native workflow in which Auction Overlap adoption rate is 5.1%
Data Metrics
I collaborated with data scientist to understand adoption rate of Auction Overlap in comparison with the control.
Identify opportunities:
Auction Overlap resolution flow had 50% lower adoption rate as compared to the control.
Persona
I created three personas based on advertiser segments. Pros/Advanced Advertisers are primary persona for Advertiser Guidance.
CONTROL VERSION
On right is control version of Auction Overlap which has high adoption rate.
CURRENT MFR VERSION
On right is current MFR version of Auction Overlap which has 50% low adoption rate as compared to the control.
DESIGN EXPLORATION 1
Using In-Table MFR card to refresh and pause ad sets just like the “control”.
DESIGN EXPLORATION 2
Using “success toast” after pausing ad sets and providing second CTA to reallocate budget via success toast which is then triggered via bottom sheet.
Challenges: Currently Ads Manager success toast does not support primary CTA, this would need AIP process and Ads Manager approval which would need more time.
DESIGN EXPLORATION 3
In this design, we allow users to pause ad sets in L2, and then let them opt for “re-allocating the budget” CTA via model. This will also help maintain parity with “Control” performance.
Challenges: Model would need AIP approval and MFR Framework team approval which would need more time.
FINAL MVP DESIGN
Using In-Table MFR card to refresh and pause ad sets just like the “control”.
Team agreed upon this design as a final MVP to test if it is in parity with the control before adding more feature like “re-allocating budget”.
LAUNCH & SUCCESS METRICS
The main success metric we were monitoring and comparing with the control, i.e.
Pause & Publish (final adoption of resolution flow)
After launching we saw Auction Overlap design change performing as expected.
Adoption rate of Pause & Publish improved by 4%.