Visualizing Intent
Header Image
Case Study By
2 months
Project Title
Visualizing Intent

Design Team

  • Partner and chief scientist: Mark Schindler
  • Design lead: Andrew Liebchen
  • Managing partner: Angela Shen-Hsieh

Client Team

  • Co-Founder and CEO: Jonathan Mendez
  • Co-Founder and CTO: Richard Shea
  • Chief data scientist: Soren Macbeth
  • Senior front-end developer: Eric Solen
  • Software engineer: David White

Ed. note: This case study is a selection from the 2012 “Justified” competition, in which an esteemed jury identified submissions that demonstrate the value of design in a clear, compelling and accessible way. It serves as an example of how to explain design thinking to clients, students, peers and the public in general, based on specific metrics.


Yieldbot is a startup company tackling a big data problem. Their web-based software product analyzes website traffic data to help online publishers optimize their ad revenue. was engaged to design the user interface for their initial software product offering. Yieldbot’s technology pulls in their clients’ clickstream data and applies their own proprietary analytics to interpret visitors’ “intention” based upon keywords used and other data. Having insight into visitors’ intentions enables publishers to better match advertising and tailor content to those visitors. “We tell publishers why visitors want to view every page of their website and make those findings actionable in realtime. To do this we collect massive amounts of data and solve complex problems with it,” says Jonathan Mendez, Yieldbot’s founder and CEO.

The application would bundle keywords and phrases into intent “segments” and serve these to marketers to analyze and choose which ones to implement. Yieldbot’s technology could not only track the historic performance of keyword phrases (page views, landings, bounce rates, etc.), but would also predict future performance. Users would need to see the data around these keywords, be able to analyze and manipulate the keywords and segments and implement their choices.

Slideshow Image

Yieldbot’s technology analyzes keywords as a proxy for intent. These views illustrate the dynamics around a group of keywords used to access one publisher's website. The volume of page views is represented by the bar heights and page depth (how many pages into a website a user goes) is represented by the length of the gray bar in the center column. Keywords related to the primary keyword are show in the far right column. The rollover state highlights a single intent segment. (Design Team)

Slideshow Image

This sequence describes the process of building a marketing segment around a single keyword. First, the publisher sees an overall cluster map of keyword-based intent generated with data from his or her site. The user can then zoom into a heavily trafficked keyword. The publisher can then “grow” this keyword into a segment of related keywords (i.e., the surround nodes) and see important summary statistics dynamically calculate (on the right) to test different combinations of keywords. (Design Team)

Slideshow Image

Once the marketing segments are created, their performance is tracked against actual site analytics. Segments are compared by their page view performance. More detailed analytical information for the segment can be quickly accessed by expanding the segment row. (Design Team)


Total design and development budget unknown


Research included detailed study of the client’s technology and data.


The challenge of this project was to make information and data that advertisers had never seen before—a wholly new concept around online advertising analytics—engaging, understandable and actionable. This meant making the vast amounts of session clickstream data and the algorithms which defined intent segments understandable to a user in such a way that they would have confidence in adjusting their site content and strategy. If the black box could not be demystified—and in a way that was simple and accurate, yet still conveyed its inherent complexities—there would be big challenges in selling the product. As Jonathan put it, “None of the hard stuff matters unless we can present this data to publishers in a simple way that they will understand and trust.”

Strategy’s design strategy was what we call an “apps approach.” Typically, outputs from a business analytics system would be tabular reports—lots and lots of grids of data and a series of menus which would allow the user to query the database. These are usually consolidated under a “Reports” tab or button, where users go to print out these lists of data. Users often transfer this data into Excel to do analysis, and then refer to printouts of their Excel models when working in, or enabling actions, in the software. Our approach was to enable, through visualization, appropriate views of the data in line with their analysis process—in the same way an iPhone application like Yelp knows to show you all the nearby options when you search for “Italian restaurants,” and then, once you’ve found the one you want, provides you with directions from your current location. The user’s experience is then much more fluid, as they are guided through workflows supported by the data.

This required the design of a visual summary of the underlying analytics in the technology and rich, interactive visualizations that would expose the key context, dimensions and dynamics of the keywords to enable easy interpretation and analysis.


These visualizations and the design of the user interface gave an intuitive and accessible form to a powerful and complex technology. “The design process made that happen.” The design work enabled the company to bring the technology to market and land key customers. “We work with some of the largest media companies in the world and see data on over 500 million page views a month through our platform,” Jonathan reports. This face to the technology has also helped fundraising and the company reports that “we have raised $5.5 million to date.” The implementation of this product can ultimately improve the experience of web users everywhere by connecting them with more targeted content through less effort (multiple keyword searches). It enables online advertisers to spend less and reach the right audience.

Integrating visual analysis also alleviates the need to print reports.

Tags Case study information design Justified