The initial release of ACF, though effective in driving advertiser performance, was largely engineering-driven, with minimal design support. As a result, the user experience felt disconnected and cumbersome in certain areas. Additionally, while the initial experiment with managed SMBs delivered promising results, when we tested ACF with unmanaged SMBs, we encountered an unexpected drop in ARPA by 15%.

‍

This presented a major challenge as we couldn’t move forward with a full-scale release to unmanaged SMBs without addressing this issue.Despite this setback, we recognized an opportunity to make ACF more efficient. Our team hypothesized that increasing budget defaults could bring the results more in line with managed SMBs, correcting the ARPA drop, and delivering a better overall experience for advertisers by simplifying the ad setup process further.

My goal was to reevaluate the existing experiment and focus on key improvements. We knew that eliminating the ARPA drop was critical, but I also wanted to improve the overall user experience and bring the flow into compliance with our design system. Through collaboration, we identified ways to streamline the interface, reduce the number of screens in the flow even further, and offer ACF as the default preference at various entry-points throughout the product in order to guide advertisers towards the best the tool.

‍

Part of the redesign involved incorporating elements from Pinterest's Quick Promote Tool—a tool focused on promoting a single Pin—into ACF. By unifying the best features of both tools, we could lay the groundwork for deprecating Quick Promote in the future, simplifying the overall product suite.

The redesign of ACF required a holistic approach, beginning with a thorough analysis of user feedback and performance data. Our priority was to reduce cognitive load by removing unnecessary choices like custom bidding or targeting, while offering data-backed recommendations for budget and Pin selection. We further reduced the number of screens in the flow and simplified decision-making for advertisers. We incorporated elements from the Quick Promote Tool to make ACF more versatile and lay the groundwork for its eventual consolidation.

‍

To tackle the ARPA drop observed in unmanaged SMBs, we introduced personalized budget recommendations, hypothesizing that aligning budgets closer to those of the control group would stabilize revenue. We expected this would also reduce the performance gains, since CPCs usually increase as budgets increase. Fortunately, after launching the new experiment we saw that not only did overall ARPA increase but we also observed positive trends in ad performance (CPC and CTR).

‍

The results were encouraging. Upon launching the redesigned ACF tool for unmanaged SMBs, we observed significant improvements in both revenue and performance metrics. The Average Revenue Per Advertiser (ARPA) increased by 22%, while the Average Revenue Per Campaign (ARPC) saw a 10% rise. Additionally, the Cost Per Click (CPC) dropped by 37%, demonstrating the effectiveness of the redesign in driving better outcomes for advertisers.

‍

One month after the tool’s general availability (GA) for managed SMBs, daily revenue through ACF reached an all-time high of $32K, reflecting a 55% increase in average daily revenue. Over the following nine months, this figure grew further, with daily average revenue reaching $65K, cementing ACF as a key driver of Pinterest's success in the SMB advertising space.