Lighthouse

Perfecting tool for data collection and validation.

team

1 Product managers, 4 front-end and back-end engineers, 2 data strategists, and me

tools

Sketch, Principle, InVision Studio

my role

UX designer and prototyper

time

Mar. 2018 - Aug. 2018

what is Lighthouse?

Through a couple of rounds of iterations. I redesigned a tool, so called Lighthouse, to help our clients submit better quality of data and have a better understanding of the data collection and validation process in the data on-boarding.

why we need lighthouse?

At Neustar, to build the most accurate and effective attribution models, it’s important to capture as many events as possible that our customer has with their brand. So it is very important to pull in data from many different sources to paint a highly accurate picture of our users’ customers’ journeys.

So, our data team needs a tool to help bridge the gap between what our clients can provide and what we expect to receive. This also contributes to our end goal of automating the entire process and greatly improve our efficiency in collecting data and shorten the on-boarding time for our clients.

As the UX designer for this project, I am responsible to redesign the whole experience of data collection and data validation.

Who am I designed for?

  • Our data strategist who expect to receive high quality of data coming from our clients and validate the data so that they can pass the data to modelers.
  • Our clients who expect to have a portal to send all their data easily and fast, so that they can have a short period of time of deployment and start using our marketing analysis tools as soon as possible.

What problems are they facing?

I interviewed 5 data strategists in our data team, each have worked with multiple clients and frequently communicated with them. I wanted to understand why our current app and process impede our clients from submitting good quality of data we could use and how the daily communication goes between them and the clients. I consolidated all the findings into a few pain points below:

Pain point 1:

Our current tool is just a glorified FTP, clients simply asked for white glove service after they submitted the files. Clients didn't participate in the data collection process. Our data strategists need to proactively request data, solve data issues and track the data collection progress with clients using emails, which are very time consuming.

Pain point 2:

Our clients don’t know what data we expect to receive.

Pain point 3:

Our clients don’t know what is happening with their data when it is in data collection.

Pain point 4:

Normally, we have a tight schedule, the process is very time sensitive.

Main focuses

Before jumping to design, I decided to focus on three main points. This will ensure the design scope to achieve a viable and consistent design.

My goal of the design is to provide our clients an interactive way of uploading and validating data.

This new flow empowers them easily track their uploading progress, submit better data and understand what is happening with their data when it is in the system.

  • Be Transparent, it is very important to let our users know where they are in the data collection process, how good is the data, if there are issues, how can they solve them.
  • Be Effective, because the data collection process is time sensitive, this tool should allow users to submit all of their data in an effective way.
  • Be Delightful, the data collection process could be cumbersome, if we invite our clients to participate in this process, it is important to make the whole experience fun and easy to follow.

In order to improve transparency to the user, I should let clients know which set of data they can upload or download. Also clients should know whether the data collecting in progress? For which period? By when they are expected to submit all the data? When they are uploading, it should be easy for them to understand how much work is still left and which files need to resubmit. And finally when they are done with uploading, I should clearly show they have completed which portion of the data collection.

In order to make the tool more effective, I should allow users to submit multiple files at the same time. Give them a very clear structure to manage their files. Also the entire flow should be very straight forward, easy to follow and quick.

In order to make the tool more delightful, the copy should be easy to understand, general visual style could be playful.

Data Collection

After equipped with user research, I am confident to start the design process. There are two phases in the uploading process, data collection and data validation.

First, I started designing user flows for data collection phase. I organized the data files into a three-layer structure including dashboard, folders, and files/sheets.

The user flow below illustrates how users upload files across these three layers, and how it helps users to upload and manage their data files effectively.

After having a holistic picture in mind, let’s take a deep look into key design changes for data collection phase.

Dashboard

old designnew design
old designnew design

In the current design, it is very hard to keep track of the progress of the data collection. Users don’t know which files they didn't upload in which folders and there is only one progress bar at the top of the page that is not informative to the users.

Key Changes:

  • Added a card on top of the page reporting the current data collection progress, including number of days till the end of data collection, folders that are in progress collecting data and folders that have completed data collection.
  • Remove requested and received date columns, instead, added a new column showing the collecting status of the folders.
  • Not only show how many files within the folders, but also show files that have issues.
  • Added searching function, so it is easy to find folders when there are too many of them.
  • Added a notification panel to notify users about all the recent changes, issues that need to solve and data collection progress.

Upload Flow

old designnew design
old designnew design

Current uploading process is like ftp, users can only use the tool to download previously uploaded files and upload new files. The new design is a completely revamped flow involve assigning meta-data.

Key Changes:

  • The new flow is a guided flow that welcome users to upload files and assign meta-data.
  • Users can upload multiple files, if the spreadsheet contains multiple sheets, users can choose which ones they want to upload.

Folder View

old designnew design
old designnew design

Currently, there is no folder views, there is only one modal per folder that allow you to upload and download files. The new design includes a completely new page for folders, where users can manage their files and track their uploading status of files within that folder.

Key Changes:

  • A new page for managing files within folders, users can click the file name to have a closer look at the files, reviewing the results or fixing data issues.
  • Add a column to show the status of uploading files, for example, data accepted, irregular values etc.
  • Users can close the folder if they think the collection process for this folder is completed. This will be reflected in the dashboard.

Data validation

The second phase for uploading process is data validation.

After users successfully uploading all their data, they want to

  • validate their data to see how good their data is
  • explore data in any aggregation level to help discover data issues.
  • quickly fix common data issues by reuploading the file
  • know the progress of solving data issues.
  • reach out to our data specialists if the issues cannot be solved on their end.
  • have a good communications with data specialist throughout the process.

The user flow below illustrates the general interactions between our data specialist and clients, and how the new flow can help streamline the communications and solve data issues faster.

Again, let’s take a deep look into key design changes for data validation phase.

Data Validation flow

In the new design, I introduced a new data validation flow after files have been uploaded. Users can quickly solve common data issues.

Users can check if the headers for the data spreadsheet is as expected, are there any missing headers or headers that have been assigned incorrectly. Users can also check if the data types are wrong, what are the expected data format and what are the issues about the current data.

Then users can fix the data file by reuploading that file again or send the file to our data specialist for help. Below is the user flow for reuploading.

validation report dashboard

old designnew design
old designnew design

Validation reports are generated after data collection to help users explore data from any perspective so as to find discrepancy and outliers in data. Users can start conversation with our data specialist in all reports.

Currently, the reports dashboard does not help users track the status of solving issues in validation reports. Also users need to click into each report to view the up to date conversions. So users have no context of the data issues and cannot easily access to them.

Key Changes:

  • Added scorecard on the right side, including pending conversation, reports that need review and reports that have signed off, in consistent with the data collection dashboard, these metrics help users on track of their data validation progress.
  • Added a new column showing the status of the reports. So it is very easy to tell what reports still need review.
  • Show all conversations when users click on one report, resolved conversations are hidden by default so users only need to work on pending conversations. They can access to the conversation with one click.

conversations

old designnew design
old designnew design

Currently, the comments dropdown overlap the report, so users cannot review the report and write down the comment at the same time, also the metrics panel on the left side takes up too much horizontal space, users cannot immediately get on the report when they come to this page.

Key Changes:

  • Combine metrics panel with comments into one, since users will only either tweak metrics or write comments.
  • Move report to the left and put panels on the right side, so users can focus on report when they come to this page.
  • Make it easy to add a snapshot of the current report and share it with others, this can help users provide a context for the data issues in the conversation and get answer faster.

Email templates

Finally, In addition to all the design changes, I design a set of email templates to help notify users about

  • the data collection that is in progress.
  • pending conversations.
  • newly published reports that need review.
  • reports that get signed off and conversations that get resolved.

These emails serve as the entry points for the users to keep track of their data collection and validation progress. It helps increase user engagement with the app and facilitate the communications between users and data specialists. Examples are shown below.

Outcome

Unfortunately, with some key members leaving the team, and shifting priority for the year of 2019, this project was put on hold, but the good news is this project has been rebooted by the end of 2019, and I have been working with all the talented people on the team to move this project forward.

What I've learned

  • It is important to set up design goal before dive into the design and prototype phases, this can help the design on track and help me make decisions that contribute to the goals during iterations.
  • Before starting the redesign process, it is important to ask why we want to conduct the redesign, also a redesign can run the gamut from a visual refresh to a complete user experience overhaul, if it requires a complete overhaul, I will be responsible for understanding every single piece of information that must go into the redesign and how that information fits together.
  • Moreover, for a redesign, it is important to improve upon the prior design while avoiding radical changes that may confuse the existing users.
Up Next
Neustar Academy
Open navigation