Choozle - Digital Marketing & Advertising Platform

Project Brief

As a part of our larger goal to create a multi-platform ad-tech operating system, we needed to provide our users the ability to upload library assets, specifically geo lists, so they can be used in any buy-side platform. Currently, when a geo list is uploaded into the Choozle platform, the list is emailed to our support team who then uploads that list into TTD for TTD to process. After processing without errors, the list then becomes available in that account's Choozle Geo List Library and can be used in an ad group. Instead we want to store the geo list asset in Choozle and when selected for use within an ad group and send the data to the corresponding buy-side platform. 

My Role

User Research

Data analysis

Information architecture

Ideation

Prototyping

Usability Testing

Final Designs and iterations

 

                                          Project Lifecycle

chlife.png

                                                   

                                                    Research

 

Online Research

I started the research from Pendo - a service that allows to track users behavior online and provides contextual help. I created a segment of Power Users and First Timers, based on number of clicks on the Geo Location List Page. 

I also was able to place a link to a survey Via Pendo. Only the users from created by me segments could see that link. 

 

Heuristic Evaluation

The usability evaluation of the platform revealed 

  • Delays in notifications or no notifications at all

  • Users can't edit, delete uploads manually

  • No error messages when a format hasn't been approved, no confirmation option

  • Instructions for use of the system should be more visible/easily retrievable whenever appropriate.

  • Error messages are all different colors and fonts. When uploading a wrong format, the error message has no explanation of the reason.

  • Document templates - are not very intuitive to use, helpful info is hidden

  • The turnaround time for Geo lists is too long

 

Comparative Analysis

I did comparative analysis of three other digital marketing platforms involved into our platform's work. Was revealed that other platforms:

  • Provide access to searching zip codes by location (county, city, country)

  • Allow users to manually input zip codes

  • Provide error messages instantly upon validation of a zip code

  • Don't require for a template file to be downloaded

  • Allow users to create their lists online

 

Surveys:

I used Survey Monkey for the surveys. I gathered some quick feedback, that I could use for the following work. 

15 people took the survey. Five of them were first time users of the Geo Location List feature. 

 
 

Main Key Takeaways:

 

Overall, how satisfied or dissatisfied are you with the Geo Location List Feature?

 
 

What would improve your Geo Location List creation experience?

 
 

Which of the following words would you use to describe the Geo Location List feature?

 

 

Interviews

I conducted interviews with 8 people. Half of them were internal users (Choozle's employees, who are heavily involved into using the platform) and another half - clients.

 

the Main Key Takeaways

  1. Users use outside map tools for generating zip codes by radius or location

  2. Users want to be able to share zip codes across advertisers, saving it locally

  3. One of the most common feedback - users want to re-use the same list

  4. Users want to be able to manually input Geos

  5. They also want to be notified about errors in their lists

  6. To be able to see the status of the list ("pending approval", "approved", "disapproved due to..." with the error explanation)

  7. Participants requested for access to City/County option in Geo Lists

  8. It is important for clients to input zip codes directly into a field during a campaign set up.

  9. The ability to store and share previously uploaded lists would be very helpful for the clients, this way they will be able to have an upload available for every client/campaign

  10. To have zip codes searchable in Geo Groups (populate for selection).

  11. The ability to edit/delete lists (preferably in bulk) within the platform

  12. It is important to date Geo Lists and to get reports at a zip code level for future campaigns.

  13. Geo Fencing - one of the requests

 

Card Sorting

After gathering all the data, I began my favorite method of data analysis - Affinity Mapping or Card Sorting. It helps to group different data and to clearly see common patterns. 

I did two rounds - for the data gathered internally and for the data gathered externally.

Internal data:

The main key takeaways

  1. There are consistent issues with Geo Lists:

  • It is time consuming to have things done manually (when uploaded the wrong format - have to manually change it and send to TTD)

  • Client can't use county instead of zip code

  • The "download the template" button is not intuitive to use

  • There is no notification about errors on a list

  • There is a traceable lack of information and instructions given to users

    2. Very important finding:

  • Users don't tend to read instructional info

    3. Requests for Geo Lists:

  • The ability to share zip codes across advertisers

  • Saving lists locally

  • Re-use of the same list

  • The ability to draw on a map

  • The ability to specify counties

  • Editing of lists

External data:

I collected all the data from surveys conducted (Power users of Geo Location List and First time Users) and interviews. I organized it in different groups.

The main key takeaways

  1. There are certain issues our clients experienced while running campaigns:

  • There are no error notifications

  • The turnaround time is too long

  • Lack of information on the page/instructions on how to use certain features or international zip codes.

     2. Important findings:

  • Most users wish to be helped by text instructions, tool tip, warnings, video/tutorials, or support center.

  • Most of the clients use other tools to generate zip codes and then they manually type them in the template.

    3. Clients requested for Geo Location List feature:

  • The ability to manually input Geos

  • The ability to use a radius of zip codes and possibly draw it on a map within the Geo List feature

  • They also wanted to be notified about errors in their lists

  • The ability to see the status of the list ("pending approval", "approved", "disapproved due to..." with the error explanation)

  • Many survey and interview participants requested for access to City/County option in Geo Lists

  • It is important for clients to input zip codes directly into a field during a campaign set up

  • The ability to store and share previously uploaded lists would be very helpful for the clients

  • Zip codes can be searchable in Geo Groups (populate for selection)

  • The ability to edit/delete lists (preferably in bulk) would add a value for the Geo Location List feature

  • It was important also to date Geo Lists and to get reports at a zip code level for future campaigns

  • Geo Fencing

After getting the client's feedback with lots of requests. I needed to define what could be built and what should be built.

It's very important to see this edge and to implement the right amount of features.

User can't be overwhelmed. User interface must be simple to provide seamless experience.

 

Personas

After affinity mapping I began creating personas. I went with 2 personas that would be a representation of our users user, based off user research and incorporating user goals, needs, and interests.

 

 

Ideation

Whiteboarding/Brainstorming

Initial sketches were done after white-boarding and brainstorming sessions with my team ( Product Manager and UI Engineer).

 

User flow

Wireframing

First wireframes were based on sketches, and I used them for User Testings internally and externally with clients.

 

Iterations

Testing

I conducted 8 user testings. Getting the feedback after every session, I could see common patterns: users were asking similar questions, were pointing at same things. Based on that data I did several rounds of iterations. 

Initial Library View

New Design Development

The final version of design provides the ability to:

  • View dates/times of list uploads and last changes

  • View a number of zip codes on a list

  • View a status of a list

  • Create a list from the library

  • Use the Action button, which gives variety of options for Geo Lists as Edit, Delete, Clone, Archive, Share

  • Filter lists

  • Expand a list for a preview

 

Create a Geo List (wasn't available previously).

The final version of the design provides the ability to:

  • Create a list within the platform

  • Give a list a name

  • Upload an existing file for further modifications

  • Manually input zip codes

  • Use a Map Tool, which allows to generate zip codes by:

    • radius

    • location

    • drawing on the map

    • Remove/Edit zip codes individually

  • Sort zip codes by:

    • Chronological order

    • Numerical order

    • Source

  • Preview a list on a map

  • Leave notes, comments

It was important, after multiple iterations and testings, to understand what should be built, not just what can be built. The users gave lots of feedback and requested for plenty of features, but I wanted to keep only the ones they really need and will use. 

Edit a Geo List (wasn't available previously)

The final version of the design for editing a geo list provides the ability to:

  • Edit individual zip codes

  • Remove individual zip codes

  • Add new zip codes by uploading a file or manual input

  • Edit a lists name

  • Save a list as a new one, keeping the original one

 

Map Tool (instead of using the offside map tool, users will be provided with the ability to use this tool within the platform)

Map Tool allows users to:

  • Generate zip codes by

    • Radius

    • Location (including county and city)

    • Drawing a shape on the map

  • Remove individual zip codes

  • Reset the map

 

I also worked on validating possible errors up front, so users won't have to wait up to 48 hours for a list to be processed. Upon having any errors users will be prompted to fix them, so a list could be used in a campaign set up instantly. 

One of the biggest findings was that users needed a better way of getting instructions and guidance. With my initiative it became a major project for educating our users and creating an Onboarding Process, not only for new coming features, but for the entire platform overall. 

The new Geolocation Library reduced support tickets by 50% and increased the monthly ad group spend by 25%