Defining the Project
I worked closely with research and product management to define the Project Plan.
Scenario 1: Users already running standard FB ads
Retargeting ads are a logical next step for advertisers who start with standard Facebook ads. They help users reach a new audience (web visitors, including anonymous browsers).
Scenario 2: Users tight on time
With retargeting ads, users save time. They don’t need to configure an ad campaign every time they want to advertise. Users can set up a campaign once, and it will run as long as they put money towards it.
Scenario 3: Users looking for results
Retargeting ads help our users make more money. According to customer interviews, retargeting ads convert well for our users, better than other types of ads.
Goals & Anti-Goals
Help extend and enhance our customer’s existing advertising and marketing efforts.
Help our users convert at a higher level
Allow users to create ads that target audience that has visited their site previously
Incorporate retargeting in a way that’s user-friendly and understandable, particularly for our less savvy users
Allow users to create complex retargeting scenarios (v1 will be streamlined to users who have visited their site previously)
Create an experience that is disruptive and lacking explanation for those less savvy
Break current design conventions/best practices
Primary research (digesting past research)
Comparative analysis (Facebook, AdRoll)
Users adopt and utilize new feature (specific KPIs TBD)
In May, the Ads team performed stakeholder interviews and concept testing on a unified ad editor experience. After concept testing, one participant in particular seemed to best embody the Mailchimp ads customer. The researcher and I used this customer to formulate a fresh persona for ads. The following quotes best describe our user, Amanda:
After a long day at teaching 7th and 8th graders, I have 20 minutes between the time I put my daughter to bed and I pass out from exhaustion.
The less I have to think about it the better. I don't have time to overthink my strategy.
To put in plainly, our ads users aren't particularly savvy when it comes to marketing and strategy, but they feel like they should be creating ads. They are tight on time and fear they may not know enough about marketing, or that the ads process is too complicated. This user mindset frames the decisions made for Facebook Retargeting.
The original research for the Retargeting Ads MVP was conducted in December 2016 and updated in November 2017, but the feature was never quite built. Using past and ongoing research, I will illustrate the problem space.
User Feedback on Ad Retargeting
Retargeting ads help extend and enhance our customer’s existing advertising and marketing efforts.
I knew that people were coming and looking at our site and weren’t purchasing, and I wanted to find a way to convert those drive-by visitors into customers.
Retargeting ads convert well for our customers.
If I were to look at our average for the year we’re probably in the 6-7X return on ad spend.
Designing static retargeting ads in AdRoll is difficult for our users.
So my ads on AdRoll are pretty stale, because, for me, it's a pain to do that, because for whatever reason whenever I do all of the modifications in Photoshop, and then I upload it, for whatever reason, there is always something wrong with the file size or whatever. So it's a very inconvenient system.
Retargeting ads work best when the user has a large enough audience visiting their site.
...now that we have about 100-150K paid users every month, we have enough people coming in that I just do retargeting with AdRoll
Our customers gauge the success of retargeting ads by the amount of money they make.
Success on advertising campaigns for people who we’ve already known and already acquired is purely ROI based. And that’s the one metric and I look at...
The more channels a customer uses to drive conversions or click-throughs, the more complicated attribution for sales becomes.
Obviously click through is worth a lot more because it means it was a direct sale from the ad... but it lead to 20 or 30 view through conversions, which means they saw the ad and then went on to purchase. Based on that data, it led to a $1500 return on $200 spend. And I know that isn’t all attributed to that ad, but it’s significant enough that it’s worth it. And it’s also lifted stuff other than sales, like email sign ups and growth on Facebook, which has helped us draw more organic sales over time from Facebook because of the AdRoll effect.
The Goal: By introducing Facebook retargeting, we want to further extend Mailchimp's multi-channel capabilities, while supporting our users' efforts to convert more website viewers into paying customers.
With these insights in mind, and after examining experiences like Facebook and AdRoll, I presented three distinct directions to my cross- functional Ads team, as well as the design team. For the MVP, we decided to focus solely on website visits and static ads for retargeting. This decision was made based on prior technical research, to eliminate complexity and deliver an experience with the least friction to our users.
Option 1: Front Door Callout and Audience Checklist Edit
Retargeting, in essence, is a matter of our users specifying an audience that has already engaged with their content, to which ads will then be displayed. For now, retargeting will be simplified and limited to people who have already visited a site, after which they’ll begin to receive ads.
As such, I proposed that a new ad type was not necessary to create in the "front door" (the modal in which users select their campaign type). Instead, I propose calling out the new retargeting functionality in the front door and providing retargeting as an audience option in the checklist.
Rationale: Channel driven, semi-savvy user
This approach is informed by Facebook's ad editor experience and implies the following user mindset: 'I am more concerned with selecting my channel, then I'll specify the details as I go. I understand that retargeting is just a matter of audience and that the people who visit my site will be delivered the ads accordingly.' I think with this approach, we may be assuming our users are more savvy than they are, which doesn't necessarily align with our persona.
Option 2: New Ad Type
In this option, Facebook retargeting will be presented as a new Ad type at the front door. Facebook retargeting will closely follow our Google remarketing experience, as a user must begin with a connected site, and the audience remains fixed as those visiting the selected site.
Rationale: Intent driven, not-as-savvy user
This approach was recommended by our researchers in prior findings and caters to our non-expert marketers. It implies the following user mindset: 'I am more concerned about intent and knowing up front that retargeting is an ad type helps me understand that I can reach a whole new set of users... how exciting!'
Option 3: New Front Door– FB/IG Segmentation
In this option, FB/IG ads is segmented into non-retargeting and retargeting options, which follows an existing pattern in Mailchimp's email campaign front door. This approach highlights the new ads functionality, while maintaining the context of channel type.
Rationale: Channel and Intent driven, not-as-savvy user
This approach caters to our not-so-savvy marketers who may understand the idea of channels (Facebook vs. Google) , but also would benefit from seeing at a higher level that they can try a new tactic (retargeting) to reach a new audience. It implies the following user mindset: 'I understand that my ads can appear in certain channels, but it would be nice to begin the process with deciding on my "why." In other words, discovering a new tactic to reach new customers, for a channel I'm likely already familiar with is kind of exciting!'
The favorite, particularly among the design team, was option 3. Here are some of the most useful feedback points I received:
...can we do more to explain why users should create an ad in MC vs on the platform itself (e.g., on Facebook) or in a 3rd party tool? (maybe in front door?)
In comparison to other tools, Mailchimp's ads offering could be viewed as limited/not robust. It is important to explain to our users the benefit of using our ads platform, as it integrates with a larger marketing ecosystem. Also, simplicity benefits many of our less savvy marketers.
Which options align with user's mental model of channel options?
Retargeting doesn’t feel like an audience option. More like an intent.
User research suggests that our ads customer often recognizes that retargeting is an effective marketing tactic, but their ultimate goal is to "convert those drive-by visitors into customers" (intent based mindset). It seems that surfacing retargeting higher in the experience is beneficial given this mindset, instead of relying on our users knowing the minutia of retargeting being an audience type at its core.
Which option is the most scalable? Why?
With this MVP, we've limited retargeting to website visitors and static ads. As we expect the features to expand over time, building retargeting within the regular Facebook/IG ads experience could become quite complicated and make less sense (eliminating Option 1). As ads as a whole continues to evolve, highlighting Facebook retargeting as its own ads type is also concerning, as more ad types may emerge, cluttering the tabs and complicating the hierarchy (Option 2 may not scale well). This left me to recommend Option 3 as the final direction.
Working with Content
Once settling on Option 3, I worked with our technical writer to iron out small kinks in the experience. We planned to utilize existing messaging/flows within Google remarketing, but cater them to the Facebook experience, which was a bit of a challenge. To account for the discrepancies, we opted to make all dialogs related to edge cases generic, so they could be applied to any ads scenario to come and not just Google or Facebook specifically.
We also had to ask small questions like "Do we include Instagram when we describe Facebook retargeting, as technically Instagram is included in the functionality?" Ultimately, we opted to formally refer to the feature as "Facebook retargeting," but reference Instagram as well in any explanatory text.
The front door segment components in Option 3 are part of the current Mailchimp experience (see our automated Email campaign type), but our developer discovered that they were "hacked in" and not available in all views. Our multi-channel team would need to take on this work, which is apparently quite an effort on their side. Due to the technical limitations, we are first building Option 2, then later optimizing the front door for Option 3, when we have the bandwith.
To see the final direction more closely, including updated copy, please click here. Thank you!
Building Front End
Plan to optimize front door for Option 3
Explore educating users more on "why ads with Mailchimp?"
This will likely require speaking to more users/additional research