Paid social automation across accounts and channels

More automation use, fewer support queries

Approximately 20% more feature usage, 60% fewer automation-related support queries and 6% more customer spend managed through the feature.¹ Spotify used the redesigned automation for more than a year and increased the budget it managed through it.

I redesigned a seven-year-old automation system for performance marketing teams: reusable Trigger Sets across accounts and channels, trigger creation inside reporting, and an activity history explaining what ran and why.

An automation feature people were struggling to use

This was the first and longest-running project I owned at Smartly.io. Experienced performance marketers used campaign triggers to improve return on ad spend, but new-customer adoption was declining and support requests were growing.

The system was seven years old. Smartly was expanding across Meta, Snapchat, Pinterest and TikTok, while triggers remained tied to individual accounts. Users had to clone logic to reuse it, and the backend limited what we could change in the experience.

I was the sole design owner for the redesign, working with product management and engineering on the workflows, interaction model and product direction.

Two groups needed different things

Advanced users wanted to reuse automation across accounts and channels. They could handle detailed controls, but repetitive setup and rigid logic made their work harder.

Newer users struggled with setup and navigation. They also had trouble seeing what an automation had done and why. Some returned to native ad tools when Smartly took too much effort to learn.

Our redesign needed to make the first setup easier while keeping the controls experienced marketers relied on.

My role

I led problem framing, workshops with Product and Engineering, scope decisions and prototype validation. I worked with leadership, Sales and Customer Success to connect customer needs with the engineering work required to support them.

Understand the constraints before changing the interface

We analysed support tickets, reviewed the bug backlog and compared usage across customer cohorts. I used those findings to examine which problems came from the interface and which required changes to the automation engine.

Over several weeks, I worked with product management and engineering to agree on the first phase and what we would defer. A few interface fixes would have left account and channel restrictions in place, so we chose a larger redesign despite the added engineering effort and longer timeline.

We tested prototypes, ran alpha releases with selected customers, including Zalando and Spotify, and reviewed feedback with Sales and Customer Success.

Three decisions shaped the redesign

Group reusable logic into Trigger Sets

Trigger Sets gave users a way to group automation logic and reuse it across accounts and channels. This addressed the repeated cloning that frustrated experienced marketers.

We also agreed to introduce a new trigger engine. It supported cross-account and cross-channel automation and advanced scheduling, giving us room to change the experience beyond the old backend’s limits.

Let users create triggers from reporting

Users already checked campaign performance in reporting. I brought trigger creation into those views, so they could act on what they were seeing without navigating to a separate automation area.

Proposed product workflow

Track the usability work we deferred

OKRs and capacity meant we could not address every refinement in the first release. We recorded those decisions and tracked the remaining usability work so later teams could understand what still needed attention.

The redesigned workflow

This diagram compares the proposed navigation with the previous version.

Users could reuse automation across accounts and channels, start with templates and inspect an activity history showing what ran and why. I simplified the hierarchy and layouts, and we adopted several of the component patterns into the wider design system.

Close

1 / 3

Results after launch

The redesign took 18 months, through changing OKRs and several leadership transitions. Our decision records helped new product managers and engineers pick up the work when the initiative resumed.

In the first year after general availability, we recorded:

• Approximately 20% more feature usage.
• Approximately 60% fewer automation-related support queries.
• Approximately 6% more customer spend managed through the feature, with several hundred thousand dollars in incremental revenue.

Spotify, one of our largest enterprise customers, used the redesigned automation for more than a year and increased the budget it managed through the feature.

Product leadership recognised the work, and I received a promotion and a change in responsibilities following the redesign.

What I carried into later work

The project changed how I document decisions. Across an 18-month effort, the reasons for a choice needed to survive changes in the team. Keeping the deferred usability work alongside those decisions made it easier to resume without repeating earlier discussions.

Bringing engineers into customer conversations helped us make choices together. Working with the team on analytics also meant we could measure usage, support demand and spend after launch.

Close