Smart Marketer -
Let AI Manage Ad Buying
A product designed to make marketers’ work easier by using AI to generate marketing campaigns, create digital ads (search, display, and social), and manage ad spend.
By analyzing historical marketing data, the platform would create accurate forecasts for future campaigns, freeing marketers to focus on higher-level tasks and strategies instead of getting bogged down in the complex details of placing and managing ad spend.
My Role
Platform: Web App, B2B, Marketing Automation
I was the lead designer on this project: I helped create the user requirements, wireframed, prototyped, and tested designs with users, and I lead a visual designer in the creation of the design language for this project.
Project Goals
We knew that our marketing team, which was comprised of brilliant people, was stretched super thin. Our intention with this project was to take some of their duties off their plate in an intelligent way so that they could spend their time solving more complex problems.
We narrowed down our goals to:
Use historical data to generate personalized marketing plans
Automate the ad creation and buying process using AI
Research & Discovery
When we started this project, I interviewed the marketing team leads to understand pain points for the team and then took those findings to the head of Product to develop the business requirements for the app.
After setting out initial requirements, I met with the marketers to get more insights into how they created marketing campaigns, what their spreadsheets looked like, and I got to hear their struggles. These conversations turned into the user requirements.
Learning: This could be easier
We were at a loss for how to generate a year-long marketing plan for brands with individual niches. For example, brands that sell men’s shoes have a very different life cycle than brands selling school supplies. Where was the commonality?
Through our interviews, we learned that marketers were using last year’s campaigns and sales data to create upcoming campaigns and forecasts.
We figured that if we could connect to their historical data, we could use AI to create future campaigns, including info like start-end dates, expected spend, channels, and expected revenue. Problem solved!
Before getting too far deep into this project, we looked into different automated marketing companies to see if we could find anyone doing anything similar, and we couldn’t. This was both exciting because we could be the first, but also kinda signaled a red flag. Why hasn’t anyone done this yet?
We decided to move forward with the project, and even though we wanted be able to auto-publish ads to sites like Facebook, Instagram, YouTube, and TikTok, we decided to focus on Google Adwords first.
Design, Testing, & Validation
I created rapid wireframes and prototypes for proof of concept and tested them with potential customers to see if their needs and expectations were met. After receiving feedback, I iterated and brought the updated prototypes back to the testers, and they were all very excited about the changes I made. The feedback I received ranged from “I’d like to control what part of the marketing funnel is focused on; I have a lot of inventory I need to sell, and don’t need to find new customers” to “I want to see my projected spend and returns down to the day so I can compare with the same time last year.”
At the same time as working through the prototypes, I led a designer in creating the new brand look and design library for Smart Marketer. I incorporated the new designs into the prototypes when testing with users to gauge their responses to the aesthetic.
At the project kickoff, I led brainstorming sessions where we sketched the types of content and functionality throughout the app
By end of day, I translated the sketches to wireframes with notes/concerns that the group had. These initial wireframes were presented to potential users for feedback.
The first round of prototypes was very bare. I abandoned the idea of a mini calendar (unhelpful) while adding the funnel functionality.
After learning we could use historical data to generate campaigns, I brought back the marketing calendar, but this time populated with our campaign recommendations.
After more feedback, I added the ability to control how spend was allocated to different channels, as well as example ads that the system would generate for upcoming campaigns.
Encountering Skepticism
During the testing phase, we encountered skepticism among potential customers regarding the platform’s ability to effectively manage their marketing strategy. Why should they hand over credit card info without trusting the product?
To address this, we introduced the “Marketing Audit” feature, which demonstrated the tangible benefits of our solution by providing historical data analysis and identifying areas for improvement. We focused on things like ”This is the distribution of spend across all your channels, but only 3 channels resulted in significant returns. If we focus on those 3 channels, we can have higher returns.”
The response to this feature was highly positive. Testers were enthusiastic about the insights and recommendations. By showing an analysis of their previous campaigns, and pointing out how to improve, they were more comfortable using the platform.
Learning: Marketers want more control
We discovered that marketers prioritized different parts of the sales funnel. Some wanted to be more aggressive in reaching new prospects, while others preferred to focus on proven customers. Beyond approving a budget for ad spend, marketers wanted more influence on the focus of the spend (both in the funnel and channels).
In response, we included a feature that allows marketers to adjust the spend strategies of their forecast, and they could instantly see the impact on their projected results.
Implementation & Development
We worked with an external team to build the product, and we successfully got to an Alpha version of the product: it could generate Google Adwords and control the spend, and we were in the middle of building out the AI that generated future campaigns before leadership decided to shelf the project till a future time.
All in all, it was a lot of fun to design, and I learned so much about the marketing process.