KaTE McManuS
Users enter JourneyFinder here. On the landing page, users begin their search for customer journeys using the search bar.
Users enter JourneyFinder here. On the landing page, users begin their search for customer journeys using the search bar.

JourneyFinder

  1. Google
  2. UX Design
  3. 2021

Overview

JourneyFinder is used by 3,000+ Google employees to navigate over 320k customer journeys, consisting of 800M touch-points, across 30 verticals. Beyond empowering Sales and Marketing teams with examples of real journeys, the internal tool enabled core insights in Google's narrative for their $100B Ad Marketing business. These insights were presented in executive keynotes, featured in Think With Google articles (collectively receiving 256k UPVs), and referenced at CES and Shoptalk. I led the tool's UX Design. The timeline was very short—I had a 2-week design sprint to hand-off final designs. (Note: Journey data wasn't internal Google data, but purchased from a third-party where participants opted in to use tracking software in a paid study.)

User goals & tasks

Leading up to the sprint, I joined user interviews led by our UX researcher with internal Sales and Marketing folks. We observed the participants interact with an early iteration of JourneyFinder and learned about the role of customer journeys in their work. This groundwork informed our efforts to define their critical goals and the set of tasks to achieve those goals. Establishing user goals and tasks kept our team aligned and accountable for UX quality as we strove to create a useful and usable product.

This page displays customer journeys for users to browse. Using filters and sorts, users can refine the search results to find customer journeys that are relevant to them.
This page displays customer journeys for users to browse. Using filters and sorts, users can refine the search results to find customer journeys that are relevant to them.
Users are taken here when they select a customer journey from the search results. This page summarizes the selected customer journey's key metrics and lists all of its touchpoints. Using filters and sorts, users can find the touchpoints that are relevant to them. They can also select specific touchpoints to export a customer journey visualization to a Google Slide.
Users are taken here when they select a customer journey from the search results. This page summarizes the selected customer journey's key metrics and lists all of its touchpoints. Using filters and sorts, users can find the touchpoints that are relevant to them. They can also select specific touchpoints to export a customer journey visualization to a Google Slide.
We defined our users, their critical goals, and the set of tasks to achieve those goals.

Process

To start, I organized the data attributes in a shared spreadsheet, categorizing and adding notes. The user research sessions had illuminated which attributes deserved priority, paired nicely, worked well for filtering vs. sorting, etc. Our team worked in the spreadsheet together. Engineers clarified capabilities, research fielded user questions, etc. A source of truth, it cemented our shared understanding and was foundational in later decision-making. For example, mutually exclusive attributes translated to radio buttons, whereas attributes supporting multiple choices called for checkboxes. Our attribute descriptions supplied copy for subheads and tooltips.

When the sprint began, I took advantage of established Google resources like their Material Library. Adhering to best practices and utilizing building blocks, I could rapidly shape a sophisticated tool. Without time for usability testing the new tool, it was helpful to tap into Google Flights, an enormously successful and tested Google tool with much functionality overlap. We referenced the live site frequently when finalizing interaction decisions.

Prioritization

Our rapid timeline meant we tabled some features. Our choices were dictated by the value each feature would provide to users and technical effort required. For example, we didn't include filtering results by “Purchase Sites.” User interviews revealed that this would be valuable because clients are often interested in journeys where customers purchase items from their site or their competitors' sites. Unfortunately, this required significant data cleaning, rendering it out of scope.

Attributes supporting multiple choices called for checkboxes.
Mutually exclusive attributes translated to radio buttons.
Informative subhead copy was supplied by the attribute descriptions we wrote in a shared spreadsheet.
Tooltip copy was supplied by the attribute descriptions we wrote in a shared spreadsheet
Each filter chip and sort drop-down required custom design decisions. These decisions were straightforward once we defined and categorized the various data attributes. Module interactions were informed by Google's Material Library and live tool, Google Flights.
The original design for the customer journey search results page. The “Purchase Sites” filter was removed from the design for the initial release due to substantial technical overhead and time constraints.
The original design for the customer journey search results page. The “Purchase Sites” filter was removed from the design for the initial release due to substantial technical overhead and time constraints.

Search

When designing the search bar, we decided to incorporate additional functionality beyond standard global search. The customer journey data had a 3-tier hierarchy. Industry verticals (e.g. Auto, Consumer Electronics, Health, etc.) at the top broke down into vertical subcategories, and again into further subcategories. During research sessions, many users sought journeys pertaining to a specific industry vertical because they largely worked with clients in a particular industry. Therefore, we believed it'd be advantageous for users to be able to navigate the data tiers. We also believed traversing the hierarchy would help users form a mental model of the data.

For a next step, I would have prioritized user-testing on the search bar design as it deviates from convention. In lieu of this, I leaned heavily on best practices outlined by Nielsen Norman Group (NN/G). We applied their advice to always allow users to submit a standard search query. If users search using an industry vertical or subcategory (scoped search), NN/G recommends to explicitly state where the user is in the hierarchy at the top of the results and to enable 1-click access to move between tiers (enlarge the scope). I used breadcrumbs to address this advice.

Team

Product & Engineering: Nathan Beddome; Project & UX Research: Katie Herskovitz; Engineering: Victor Zeng

The customer journey data was organized by industry vertical. These industry verticals broke down into vertical subcategories, and again into further subcategories.
The customer journey data was organized by industry vertical. These industry verticals broke down into vertical subcategories, and again into further subcategories.
When designing the search bar, we incorporated additional functionality beyond standard global search. Users can navigate the customer journey data through browsing industry verticals, industry vertical subcategories, or further subcategories. Users can also submit a standard search query (non-scoped), as seen here with the term “travel.”
Users can click the breadcrumbs resting above the search bar to navigate across the different tiers of the data hierarchy. These breadcrumbs address 2 recommendations from Nielsen Norman Group: explicitly state where the user is in the hierarchy at the top of the results and enable 1-click access to move between tiers.
An email from a happy user on a Google Analytical Team