Omniscope
A comprehensive forecasting tool that allows ad campaign strategists to analyze and explore impressions, reach, and other performance KPIs in an interactive way.
Unless otherwise stated, all work presented on this page is my own.
My Role
Lead Designer
Timeline
Jul 2018 – May 2019
Design Team Size
2 Designers
1 Product Manager
1 Researcher
Problem Statement
In July 2018, one of the product managers came to me with a theoretical opportunity space:
Our product has access to so much data, but we're not surfacing any of it to our campaign managers in our platform.
Newer campaign managers don't always understand how to use this data to optimize campaigns.
How might we build a tool that not only displays this data in our platform, but presents it in such a way as to answer our campaign managers' key questions?
The key questions we wanted our tool to be able to answer
The Timeline
Jul 2018
At this point, the feature was unofficial and unresourced. It was nothing more than napkin sketches and a conversation between myself and a PM. We bounced ideas back and forth, and once we realized we had an idea that could really become something, I began working on low-fidelity mocks in earnest. At this point, we called the project Codename: Spyglass.
Sep 2018
After getting some guerilla feedback from in-house campaign managers, we polished the low-fidelity mocks into mid-fidelity mocks and put together a product pitch. I did all of the design work. The product manager assembled the pitch deck.
Nov 2018
Spyglass was officially resourced for design! Another designer, my peer, was assigned to the team. The two of us collaborated on interaction flows focusing on input and data exploration.
Jan 2019
The MVP Spyglass was released for internal alpha testing. We collected initial feedback as internal account managers began using it for their real work. Our initial feedback was that the tool remained too overwhelming, so we created an onboarding workflow to help users learn the tool.
Mar 2019
Beta testing. Spyglass was now available to all internal users and a select few external clients. The design was finalized at this point.
May 2019
The feature was released to all users under the name Omniscope.
Final Designs
A Note on Styling
Over the years, the tool received many feature additions across multiple platform rebrands.
The mocks above represent the Omniscope tool following a visual refresh some time after 2023.
This feature was developed during a period of intense brand exploration for the product, and so the mocks I've collected may reflect different design systems.
Core Flow
Step 1 - Input Targeting
An campaign manager opens Omniscope and pulls open the targeting filter panel from the left rail. In this panel are a list of targeting options one can set for their campaigns.
An account manager can enter their intended targeting criteria to forecast the impressions, reach, and other performance KPIs they can expect to hit with our platform.
Step 1A - Import Targeting
If the campaign manager has already set up a line item in an existing campaign, they can instead import the targeting settings from that line instead of entering everything from scratch.
Step 2 - View Insights
The report populates with forecasting data for the campaign manager's inputted targeting criteria. At this stage, the campaign manager can click around the charts to view the data in different formats, or broken down in different ways.
Step 3 - Adjust Output
If the campaign manager wants to tweak their forecasts, they can edit the targeting criteria on the fly with a quick-edit panel instead of opening up the entire targeting panel once more.
Step 4 - Compile Insights
Generally, what a campaign manager would do at this stage is compile this forecast into a report to send to their clients. They can export the full page, or simply copy paste values from tables.
Feature Highlights
In addition to the core flow, here are some additional key features. Some of these screenshots are from concepts or pre-rebrand mocks, and thus may not reflect the above styling.
Onboarding Tutorial Flow
Landing Page
When users first arrive on Omniscope (still named Spyglass in the screen below), they are greeted with a page that walks them through the most common workflows in the tool. This onboarding flow is also accessible via the "Launch Wizard" button near the top of the tool.
Interactive Step-by-step
The onboarding modal walks the user through the flow step by step, using the same components from the full flow so that the user can familiarize themself with the workflow in a stripped-down setting.
Comparison
Accessing Comparison
At the top of the list of filters, there are tabs for Baseline and Comparison. Users can input a different set of targeting criteria in each tab and the charts will update to show Baseline in blue and Comparison in teal.
Charts and Tables
When a Comparison set is active, the charts will update to show both results side by side so that a user can easily compare. The KPIs and tables will also update to show the data in stacked rows.
Customization
Users can also edit the names of these tabs. For example, they could name one "Ages 25-34" and the other "Ages 35+" to better keep track of which set is which. The legends in the chart will update as well to match the user's inputted names.
Additional Features
(that I unfortunately don't have screens for)
Drill Down
Users can interact with graphs to see the data broken out in different ways. For example, clicking on the bar chart for Top Domains will show the results broken down by subdomain.
One can also break down the results by different data types, such as inventory split out by deal or by channel.
Export
Users can export tables as .csv or .xls files, can download charts as images, or can export the whole page as a comprehensive pdf report.
Create Campaign from Omniscope
Under the "Actions" button at the top, there is the option to create a campaign or line item directly from Omniscope. The option takes the user to the campaign or line creation workflow with all of the targeting values prefilled with the targeting settings they were just exploring in Omniscope.
Process
Early Ideation
As mentioned, this whole project started out as a series of napkin sketches that I unfortunately don't have access to anymore. The earliest screens I have are of the low-fidelity conceptual wireframes those sketches turned into.
In these wireframes one can already see a hint of the core layout with the filters and charts. Additionally, the ability to compare two sets of targeting and the ability to interact with the charts and tables featured early on in the ideation process.
The Pitch
The following mocks were used for the pitch deck. Also included is the intro slide for our product pitch, which does a good job of summarizing what our goals were with this feature.
Compared to the wireframes, the screens received a bit of visual polish in accordance with our platform's design styleguide at the time, but the core concept never changed.
Funded!
The pitch presentation went well. At this point, Spyglass was officially resourced as a feature.
In addition to myself and the product manager, we added one other designer and a user researcher to the team and were given one quarter to finalize the design for development.
Research
Methodology
With only one quarter to finalize the designs, the team used Rapid Iterative Testing and Evaluation (RITE) to iterate on mocks.
Instead of creating one set of mocks and having all participants give feedback on the same set, we would edit and improve mocks during usability feedback sessions, and the next participant would give feedback on the output of the previous session.
All in all, we created 8 prototypes over 18 usability tests (we had each participant come back twice), all over the course of one quarter.
The feature was built and launched in alpha testing in Q1 2019 where we continued to iterate as our alpha testers submitted feedback.
146
TOTAL PARTICIPANTS
10
Survey Participants
9
Usability Testers
127
Alpha Testers
Examples of Prototypes
Because the prototypes were iterated upon on the fly, I don't have examples of a full prototype to show, but here is a collection of variations.
To keep feedback focused on the parts of the workflow we were specifically asking about, we left all unrelated UI elements in low-fidelity.
TechPulse Paper
Later in the year, our researcher and data scientist wrote a paper on our process and our findings and submitted it to TechPulse.
Impact
User Satisfaction
2.1x
the number of users preferred Omniscope to the previously-offered alternative forecasting tool
70%
feature adoption in the first month of release
“This is wonderful! This looks like a much better way to pull reports.”
“I can already tell that it's going to be so much easier to get CPM and reach estimates compared to how we did it previously.”
“The layout is great. It's a lot to look at at first, but once you use it once it all makes sense.”
Performance
0.8s
average time to fetch data
1000
queries per minute
2 trillion
bids processed daily
> 1 million
data signals powering forecasting models
Daily active users
The first ramp-up is for internal testing. After May 2019 is the official feature launch.
What's Next?
Omniscope targeting improvements mock created by a fellow designer and is not my work
Omniscope continues to be updated over time. In addition to more features and new ways to interact with the data, we also steadily improved the usability of existing interaction flows over time.
Head over to my Streamlined Workflows case study to see an example of how we improved the Omniscope targeting workflow.
Enter password from resume to view project
© 2025 Eugenia Lee