Promo Automation

Automating Enterprise Tools

 

Challenge
Current process of moving a promotion from a marketing request to production involves many redundant steps that allow for errors and potential customer-facing issues

Hypothesis
Finding small and actionable opportunities to streamline the promotion-building process will have a big impact without affecting testing or functionality within the active system

Metric Goal
50% reduction in promotion build time

 
    • Site Ops Team

    • Cut down on repeated manual data entry to focus on larger initiatives

    • Marketing, Creative, Technology, and Compliance teams

    • Reduce the build time of a promotion

    • Gather information and create documentation

    • Discover opportunities and scope

    • Create wireframes & prototypes

    • Demo and iterate from feedback

    • Process Audit

    • User Interviews

    • Confluence - Documentation/Research

    • Jira - Dev Collaboration

    • Overflow.io - User Flows

    • Figma - Wireframes & Protoype

 
“Building a promotion involves a lot manual entry and bouncing between tools. It’s easy to make a mistake”

-Internal User
 

Promotion Request / Building Process

Marketing
- requests a promotion by filling out 1 of 19 different forms (depending on the request)

Site Ops
- sees that request, selects the template (if available)
- builds the promotion by re-entering all the data (~30-70 attributes) in the Promotion Service

This process on average takes 3-4 hours and has the potential for costly human error that can be customer-facing.

List of different marketing request forms and their functions

When all elements are in place, this process will take internal users 25 mins to complete. View Flow - >


User Set Up Issues

  • Builds are being set up incorrectly

    • Forgetting to create Events

    • Forgetting to include Tied Constraint
      (likely selected incorrect Template)

  • Accidentally getting tied by external processes (CDE, Campaign Ops)

  • Tedious build and testing process

Flow detailing the considerations from the ‘tied by’ attribute. View Flow - >


Costly Pain Point

While in discovery, I learned that users had been making a costly error by selecting incorrect templates.

Looking closely at the template selection process, I began to see opportunities to streamline this process and prevent this error from occurring.

This error was found during a regular touch base with users where I ask them about their process and discuss potential enhancements.

The template name is cut off and is not clear to the user


New Perspective

While sharing the template selection issue with developers, I floated the idea of using the unique ID of a promotion to atuo-select the template.

I created user flows that show how our system could pull data from the request service.

 

Early flow where user enters unique ID into a promo build

 

Covering All Scenarios

I iterated after getting feedback that there would be some scenarios that would not be covered by this experience.

Accounting for all known scenarios, I decided to have users enter the unique on the main promotion tab at the top of the page.

Once we had the experience in a staging environment, I had users go through the process and provide feedback.

After this round of iteration, I felt ready to deploy the new experience.

Placing the unique ID field on the promotion covered all scenarios

When entered, a modal would pop where users would visually confirm the template

This new process cut down on the chance of selecting the wrong template and creating customer-facing issues.


Automate Template Data

After seeing the benefit of users auto-selecting templates, I wondered what other data we could pull in.

Pairing with developers, I discovered we could automate the first half of a promotion build. This would cut down on timely copy/pasting and reduce errors.

I interviewed users to understand how and where this data was entered and if there were cases where it wouldn’t be available.


Prioritized List

After discovering that we could use a promotion’s unique ID to auto-select the template and pull in the static data, I wondered how I could streamline this process even further.

I preposed the idea of having a ready-to-be-made list of offers, where users can start a build and have the static data auto-populated.

Users and developers saw promise with this approach and I created a prototype to get everyone’s feedback.


Test the MVP

One of my primary questions was what elements were needed for the offers list. Knowing there were still a handful of technical unknowns and blockers, I had to focus on what would get us up and running.

I needed to confirm that the prototype was intuitive and streamlined the users’ process.

One major concern was the data flow and ownership. The design needed to communicate the status of each promo and the people involved since it could fall out of compliance at any moment.

 

MVP prototype for the Offer List

 

“This new tool saves us hours of daily work!
I can’t imagine going back to the old process.”

-Site Ops Team

Alerting Users

Once we began to pull in data from the requesting service, I knew that some promos may have some values change after being built.

I designed and implemented some indicators to inform users that a promo has fallen out of compliance. This is crucial to help prevent customer-facing issues.

I shared the idea of creating a notification system that would alert users via email, but the team deemed that out of scope.

Alerting users at every step to keep them informed of a promotion’s status


Impact & Next Steps

Auto-selecting promo templates and bringing in static data decreased the building process from 25 mins to 5 mins while also preventing costly human error.

Next, I’m iterating on flows to automate offer data that will exponentially speed up content build time. This is a primary goal for the business and has many potential uses that will make the systems more integrated and provide more personalized experiences for customers.

Future iteration with styling updates

Early Design of a modal that shows users associated offer data