Swipe Right to Approve Ads

Creating Boost’s new iOS and watchOS apps, from sketch to development

Branching out into iOS apps was a new frontier for the advertising startup.

Summary

Boost’s platform gives online advertisers the ability to crowdsource and A/B test ad copy and designs for Google AdWords, Bing Ads, and the Facebook Ads Platform.

Boost X and Boost Y are a pair of apps built for iOS and watchOS, over the course of two development sprints (or one month). The apps were built to simplify the most troublesome parts of the web app’s experience.

The problem

What we hoped users would see (left) vs. what they often saw (right): too many things to review.
  • advertisers wanted to see ongoing A/B tests, which only started after new ads were approved
  • advertisers often took days or weeks to approve the new ads submitted by the Boost network of writers
  • without new ads being approved, existing ads couldn’t be A/B tested
  • without A/B tests starting often, advertisers complained that they didn’t see enough value in the platform, and would stop using it

Hopes and goals

Build an iOS app to relieve the bottleneck, drive higher usage back to the web application, and simultaneously engage users on the go.

Toni and Judith on the left, in a meeting with the other engineers

Role

I worked on the app with Toni and Judith, two Boost engineers, during the company’s Innovation Time (similar to 20% Projects). All of us learned Swift/iOS development explicitly for this project.

I designed the iOS interface and prototyped it directly in Xcode and Framer, and we used paper prototypes for the watchOS app. Toni and I were both responsible for the front-end, while Judith took on the back-end.

Challenges designing and developing for iOS

  • developing an app when none of us knew iOS development
  • investigating, prototyping, and shipping after four days (Innovation Time took place on Friday afternoons, specifically)
  • branching out onto a whole new platform

Brainstorming and iterating

Though Toni initially tapped me to design the app’s UI, I started learning Swift so I could mockup ideas and animations in Xcode with him. This turned out to be a huge advantage, and eliminated lengthy handoffs between design and engineering.

Learning Swift was pivotal in collaborating effectively on this project, and let me translate my prototypes to code.

Initial ideas

  • rebuilding the web interface to work natively on iOS, to give advertisers a secondary option
  • gamifying the process of approving ads, to encourage advertisers to approve ads more often
  • sending push notifications when writers had created new ads to approve, to decrease the cognitive load of users checking the app daily
  • making use of the Apple Watch’s actionable notifications, so advertisers could approve ads straight from a notification

Toni proposed a Tinder-like design for the app, allowing users to swipe through cards of ads. Running with the idea, we experimented with how much of Boost’s experience we could translate to mobile effectively.

After prototyping our idea in Framer (left), we were confident that simpler was better. The web app (right) had a lot of functionality, but we chose to limit the scope of our MVP to just approving and rejecting ads. This was key to building quickly, and having a focused product to test with advertisers.

In the web app, users could approve ads, reject ads, edit ads before publishing, or send a request back to the writer for ads to be revised. We opted to stick with just approving and rejecting ads for our MVP.

My mockups for the app’s design — a modern take on Tinder, for advertisers.

Introducing Apple Watch

We loved how effortless it was to run through a large queue of ads on a touchscreen. Curious about porting the app to work on watchOS, we found it difficult to touch and swipe on the watch’s small screen.

Experimenting with paper prototypes (as we didn’t have an Apple Watch at the time!), I came up with the idea of using flick gestures, instead of swiping the screen directly. Users could flick their wrist towards them to approve an ad, or away to reject it.

Designing for the Apple Watch was tricky, but ultimately a fun exercise in interaction design.

After figuring out how to disable the Watch’s auto-lock on raise feature, we found that it was the most enjoyable way to view content and interact with it.

Success

After presenting the app internally, we began collecting feedback in preparation for a full release. Advertisers were impressed with how easy it was to review ads, and account executives at Boost were proud to promote a newer, faster experience for their clients to use.

Putting the app in front of users let validated our simple approach to Boost’s complex platform. Users wanted more options to do specific tasks, like approving ads, more quickly.

Though the app’s release was put on hold, we found it immensely rewarding to research and build a brand new product in such a short amount of time. We’ve since built off of our ideas, and are beta testing a chatbot built into Facebook Messenger, that we’re calling Boo.

Boo is a simple bot that lets you chat directly with the Boost Platform and its APIs.

A big thanks to Judith and Toni for being so patient with me.