← Work

Redesigning Instagram ads to protect creator identity

Small businesses and creators weren't afraid of ads. They were afraid of what ads would do to their reputation. I designed a new ad model to address this fear.

Team

1 PD (consulting) • 1 SWE • 1 PM • 1 DS

Impact

Launched new Instagram ad model generating ~$450K daily revenue • Boost completion increased +8%

Overview

"If my content starts feeling like ads, I might lose my audience."- Small business owner, Clothing store

The boost flow had a 91% abandonment rate, users explored the flow, previewed their post as an ad, and left

The correct diagnosis wasn't "the flow is hard to use", it was: boosting a post asked users to alter their identity in public, and that felt like a cost they weren't willing to pay.

I led product strategy and end to end design for a new ad model, ad-only posts, that decoupled promotion from identity and created a new advertising category on Instagram.

My Role

Problem reframe
Drove diagnosis through behavioral analysis and dogfooding, surfaced an identity problem disguised as a conversion problem
Problem direction
Defined the ad-only post concept and core interaction model, shaped how the team framed the strategic opportunity
Alignment
Led the org pivot from boosting posts to ad-only content, a meaningful investment requiring buy-in across product and engineering
System design
Delivered a solution integrated into Instagram's ads ecosystem. including new archiving logic, visibility rules, and creation flows

The pivotal contribution was reframing the problem at the platform level, shifting the team's investment from optimizing a broken flow to building a new model entirely.

The Problem

Boosting posts had come with a hidden cost

Instagram Ad Tools screen showing options to create a new ad with your top post, boost content from profile, and manage ads with no current ads.
Boosting was restricted to content from profile only
91% of users who opened the boost flow didn’t complete it.
96% of first-time advertisers abandoned the boost flow, despite strong intent.

This wasn't a typical drop-off pattern. Three behavioral signals I identified through behavioral analysis and dogfooding pointed to something deeper:

Signal 01
First-time advertisers rarely returned to saved drafts (2%), suggesting unresolved hesitation, not task interruption
Signal 02
Drop-off increased when previewing how content would appear publicly as an ad, identity anxiety, not flow confusion
Instagram profile page for doughboy, a doughnut brand with 1,134 posts, 2K followers, and 1,756 following; shows profile picture of a pink doughnut cartoon, bio stating 'Your local doughnut spot' with website link, professional dashboard info, and story highlights labeled Season, Coffee, Glaze, and Stuffed with respective doughnut and dessert images.
Signal 03
Upsell prompts post drop-off had ~2% CVR despite strong prior intent, users had already made up their mind

The pattern was consistent: users weren't confused about the flow, they were worried about what the ad would say about them.

The solution couldn't live inside the existing boost flow. It required rethinking how ads were created from the ground up.

The flow existed, users found it, but the majority abandoned before publishing.

Alternatives Explored

Before arriving at the right solution, I explored three directions, and ruled each out with a clear rationale.

Mobile screen shows goal selection for ad with pop-up about boosting posts and sponsored post visibility.
Upsell at beginning of ad creation 
Education at the start
Provide reassurance as users enter the flow
  1. + Reduced uncertainty
  2. - Introduced friction in a high-intent moment
  3. - Risked negatively impacting the Instagram boost ecosystem
Mobile screen showing Ad Center menu with options to Boost content, Create new ad, and Get started with Automated Ads.
Facebook's Create Ad model
New ad creation flow
Adopt a structured flow similar to Facebook Ads
  1. + Clearer mental model
  2. - Increased complexity and learning curve

Mobile app screen titled Audience showing options for selecting who should see your ad, with Suggested audience selected and Next button at the bottom.
Audience selection
Guidance during audience creation
Clarify who sees the ad
  1. + Increased transparency
  2. - Addressed only a narrow part of the journey
  3. - Did not impact core hesitation

Option one was too complex. The others didn't go deep enough. The solution had to change what ads are on Instagram, not just how they're set up.

The Insight

Screen showing Posts archive with clay pottery pieces and jars of pale seagreen and black paint.
A group of handmade terracotta clay sculptures of various characters displayed on a brick surface.

Answer already existed

The infrastructure existed, it just hadn't been framed as an ad product.

Instagram's archive feature already hid posts from public profiles.

Solution

Decoupling identity from promotion

I introduced a new ad model: ad-only posts, content created specifically for advertising that is not tied to a user's public profile.

This fundamentally shifted Instagram ads from "promoting what you posted" to "creating content specifically for advertising."

Core principles

Identity: Ads should not alter a user's organic identity, what they post for followers stays separate from what they run as ads
Control: Promotion should feel safe, reversible, and controlled, users needed to believe they could advertise without consequence
Simplicity: The mental model should remain simple, ads are just posts that don't show on your profile

How it works

Create
Users create ads as separate posts, new photos, videos, or reels made specifically for promotion
Archive
Content is automatically archived and not visible on the public profile, visible only to ad audiences
Run
Ads run normally across Instagram surfaces, same delivery, same targeting, zero profile impact

System Design

Absorbing complexity so users don't have to

Ad-only posts required redefining how content, identity, and ads interact across Instagram's systems.

I partnered with Engineering and Product on three interconnected changes, with a single constraint: the user-facing model had to stay simple.

Creation flow
New ad-only post creation flow with clear education, users understand upfront that this content won't appear on their profile
Adoption path
Upsells designed to guide users from boosting toward ad-only posts,  meeting them where existing intent already existed
Mobile screen showing archived posts with a ceramic cup, paint containers labeled Pale Seagreen and Black, and clay figurines with expressive faces.
Archiving logic
Updated archiving and visibility rules in the backend, to keep the user experience lightweight

Design principle:

Clarity belongs in the experience. Complexity belongs in the system.

Tradeoffs & Risks

Decisions made against pressure

Each tradeoff had a stakeholder pushing the other direction.

Trust vs. short-term conversion
There was pressure to stay within the existing boost flow, smaller bets, faster metric movement.
Decision: Prioritize long term trust over incremental conversion gain, which enabled more sustainable adoption.
System complexity vs. user clarity
Ad-only posts introduced significant backend changes, archiving logic, decoupled delivery, new visibility rules.
Decision: Absorb complexity in the backend to preserve a simple user experience.
Transparency vs. simplicity
Separating ads from profiles risked confusing users about where their content lived and who could see it.
Decision: Add one clear moment of education, "this won't show on your profile", rather than repeated friction throughout the flow.

The Impact

  1. ~$450K in daily revenue generated in experiment
  2. 3.5x increase in upsell CVR (5% vs. ~1.5% baseline)
  3. Increased boost completion by 8% for first time advertisers
  4. New monetization pathway for Instagram
The work changed how the platform thought about ad creation: separating "what do you want to promote?" from "what is your identity on this platform?"
That separation has compounding value beyond the experiment window.

Reflection

What worked:

Reframing as identity
Reframing the problem as identity, not usability, opened a solution space that flow-level iteration never would have reached
Ruling out alternatives
Ruling out alternatives with clear rationale built stakeholder confidence and accelerated the pivot to the right direction
Absorbing complexity
Absorbing system complexity to protect UX simplicity, can create tremendous user value  

What I'd do differently:

Bring UXR earlier
Bring UXR earlier, the identity insight came from behavioral data and dogfooding, but direct interviews would have sharpened it faster
Define success metrics upfront
Define success metrics upfront with Data Science to avoid ambiguity on what "adoption" meant across teams during the experiment

Last note

How one project became another

Ad setup options showing create ad or use existing post, multi-advertiser ads, and ad creative selection UI.Ad creative panel with buttons to select or create a post and a list of three suggested social media posts.
Reducing advertiser friction in Ads Manager

Previously shipping an ML-suggested posts experience on Ads Manager surfaced an unexpected behavioral pattern:

90% of advertisers used "create ad”
10% of advertisers used "existing posts"

Even small businesses, the most resource constrained advertisers, predominantly created new ads.

Mapped across platforms, the gap was clear, Instagram had no equivalent creation experience.

That insight is what led me to pursue this project.

WorkMeta Profile+Meta New IG AdMeta Ad FormatsMonamies