← All projects

Case study · Research and design thinking · 2026

Why do young people fall for Instagram ad scams?

A six-week study that tested hypotheses with 14 young adults and led to a proposal: ALINHA Sphere, an anti-phishing design guideline applied to ads before they go live.

In short

Overview

Problem

Young people decide in seconds on the feed, and scams exploit that rush.

Proposal

A six-pillar guideline, applied by an API before the ad is published.

Finding

Many didn’t notice the warning. Informing isn’t enough.

The problem

The people who fall for scams most aren’t who you’d expect: it’s the young.

From a broad theme to a six-week scope

Three findings shaped the solution

A CSD matrix, desk research, interviews, personas and journeys led to three conclusions.

1. 01

   The decision is automatic

   Brand and price replace verification.

   Research and interviews

2. 02

   Doubt loses to urgency

   The discount wins, and the scam only shows up after paying.

   User journey

3. 03

   Every ad goes through review

   A point to act before the ad reaches the user.

   Advertiser journey

Where the decision becomes a risk

See the research artifacts: CSD matrix, personas and full journeys

CSD matrix (Certainties, Assumptions, Doubts)

Certainties

Users are usually in a hurry: in lines, mid-conversation.

Assumptions

If young people stopped to read carefully, they could avoid many attacks.

Doubts

Which young people are most susceptible?

See the full matrix

Certainties

Scams look more and more like official platforms.

Phishing exploits decisions under pressure; mobile is a context of distraction.

Attacks mainly target human error.

When distracted, the ability to notice visual inconsistencies drops sharply.

Young people tend not to stop and read, used to digital immediacy.

Assumptions

A strong warning would be enough to pull users out of autopilot.

Deliberate friction on risky actions would reduce clicks without causing drop-off.

Users are conditioned to interact impatiently.

On social media, younger users browse on autopilot.

Someone is more likely to have Instagram and not TikTok than the other way around.

Doubts

At what point does friction stop protecting and start being ignored?

How do scammers attack users?

Do attacks happen more on phones?

Can the interface reduce warning overload?

Personas

Audience

Mariana Dantas

19 · Student · Lower-middle income

“The digital window shopper”

Pain point

Hard to check new profiles; ads mixed in with personal conversations

Empathy map and personality

Personality

• IntrovertExtrovertmore extrovert
• AnalyticalCreativemore creative
• BusyFreebalanced
• DisorganizedOrganizedbalanced
• IndependentTeam playermore of a team player
• PassiveActivebalanced
• ConservativeLiberalmore liberal

Interests

• Deals on clothes, beauty and home goods
• Local influencers and budget stores
• Short, free courses

Influences

• Friends and family
• Instagram influencers
• Reviews and comments on profiles

Goals

• Land her first formal job
• Shop without breaking her budget
• Become more independent

Needs

• Affordable prices and transparent purchases
• Feel safe before paying
• Simple guidance to compare offers

Motivations

• Autonomy
• Keeping up with her friends’ trends
• Building her adult identity

Pain points

• Hard to check new profiles
• Fear of wasting money
• Ads mixed in with personal conversations

Audience

Lucas Silveira

22 · Delivery rider · Low income

“The opportunity chaser”

Pain point

Irregular income and little time to weigh information

Empathy map and personality

Personality

• IntrovertExtrovertbalanced
• AnalyticalCreativebalanced
• BusyFreebusier
• DisorganizedOrganizedmore organized
• IndependentTeam playermore of a team player
• PassiveActivemore active
• ConservativeLiberalbalanced

Interests

• Side-income opportunities
• Practical, vocational courses
• Affordable investing

Influences

• Coworkers and friends
• WhatsApp and Telegram groups
• Word of mouth

Goals

• Increase his monthly income
• Find steadier work
• Plan for big purchases

Needs

• Straightforward ways to earn money
• Simple information on his phone
• Trustworthy sources on social media

Motivations

• Getting ahead in life
• Personal recognition
• Better working conditions

Pain points

• Irregular income
• Little accessible guidance
• Little time to weigh information

Designer

Sofia Martins

34 · Trust & Safety product designer

“The guardian of healthy friction”

Pain point

Seeing important warnings ignored; safety seen as friction

Empathy map and personality

Personality

• IntrovertExtrovertbalanced
• AnalyticalCreativemore analytical
• BusyFreebusier
• DisorganizedOrganizedmore organized
• IndependentTeam playermore of a team player
• PassiveActivemore active
• ConservativeLiberalmore liberal

Interests

• Behavioral design and cognitive psychology
• User-centered security
• Ethics in digital products

Influences

• Research on heuristic and systematic thinking
• Studies on FoMO and impulsive decisions
• Guides on ethical design and dark patterns

Needs

• Protect without frustrating
• Validate interventions with controlled tests
• Decide based on risk evidence

Motivations

• Protect people in vulnerable moments
• Make safety a natural part of the experience
• Reduce the impact of scams

Pain points

• Seeing important warnings ignored
• Safety seen as friction
• Teams focused only on conversion

Journeys

User journey: Mariana sees an ad

Synthesis · Mariana’s journey

Doubt shows up, but doesn’t last

Risk grows at every stage

The scam is noticed late

1. 01Discovery

   Excited

   Sees a product she wants on sale in her stories

   RiskActs on impulse

   ImprovementStore authenticity signal

2. 02Checking

   Suspicious

   Does a quick search on price and store

   RiskIn a hurry, skips details

   Improvement“Verified store for 5 years” badge

3. 03Engagement

   Thrilled

   The discount doesn’t look like a scam: she taps to buy

   RiskExposes data and device

   ImprovementTransaction focus mode

4. 04Payment

   Anxious

   Types her card or copies the Pix code

   RiskHands over sensitive data

   ImprovementA pause to check where the money goes

5. 05Aftermath

   Uneasy

   Tracks the order and waits

   RiskSpots the scam too late

   ImprovementSafety summary and a reporting channel

See the full journey

Persona Mariana’s journey with an Instagram ad

Row	Discovery	Checking	Engagement	Payment	Aftermath
Trigger	A product she really wants shows up on sale in her stories	A moment of doubt about the offer and the store	The discount doesn’t seem absurd enough to be a scam	Reaches the payment screen	Gets the confirmation email
Action	Sees the ad and is happy to find what she wanted at a price she can afford	Does a quick search to check the price and the store	Taps to buy	Enters her card details or copies the Pix code	Follows the order through tracking
Feeling	Excited about the chance to save	Confused and suspicious about the lack of clarity	Thrilled: suspicious, but decided to buy anyway	Anxious from time pressure and fear of missing the deal	Unsure whether the order will arrive
Risk	Excitement makes her feel she has to act fast	Multitasking makes her skip details	Data exposure and even device infection	Exposure of sensitive data	The order never arrives, and she’s slow to notice the scam
Improvement	Signal the store’s authenticity before the content opens	A badge that translates technical data: “verified store for 5 years”	A “transaction focus” mode, with fewer distractions	Positive friction: a pause to check where the money goes	Post-purchase safety summary and a reporting channel

Advertiser journey: boosting a post

Synthesis · advertiser journey

Seven steps, few explanations

Every ad goes through review

The point of intervention

1. 1

   Post

   Unsure

   “Is this post good for an ad?”

2. 2

   Goal

   Confused

   “What’s the difference between the options?”

3. 3

   Destination

   Wary

   “What if I get the link wrong?”

4. 4

   Audience

   Confused

   “What does ‘automatic’ mean?”

5. 5

   Budget

   Afraid of overspending

   “How much is too little or too much?”

6. 6

   Payment

   Wary

   “Will I be charged again?”

7. 7

   Review

   Anxious

   “How long does it take? What gets rejected?”

   ALINHA Sphere steps in here

See the full advertiser journey

Advertiser journey when boosting an Instagram post

Row	Post	Goal	Destination	Audience	Budget and duration	Payment	Review
Action	Picks a post and taps “Boost”	Chooses a goal: profile visits, website or messages	Adds a link or chooses to receive messages	Lets Instagram choose or defines the audience	Sets how much to spend and for how many days	Confirms card or balance	Taps “Promote” and waits: “Your ad is in review”
Question	“Is this post good for an ad?”	Doesn’t quite get the difference between the options	Afraid of getting the link wrong; unsure which channel converts best	Doesn’t understand what “automatic” means	Doesn’t know how much is too little or too much; afraid of spending with no return	Afraid of future automatic charges	Doesn’t know how long it takes or what might get rejected
Opportunity	Suggest posts better suited to ads	Explain each option with practical examples	Tips by type of business	A clear comparison: “automatic = similar to your followers”	Estimated reach simulator	Explain billing and how to pause	A review timeline and common rejection reasons

From findings to decisions

The scam depends on the architecture of the decision. So the question changed.

Every decision started from evidence

HypothesisIf ad design includes healthy friction and transparency, users will make less impulsive decisions and be less susceptible to phishing.

See more: comparing paths, vision board, ideas and scope

Comparing the three paths

The three paths compared against the team’s criteria

Criterion	Voluntary guideline	In-app interventions	API before publishing
Reaches scammers’ ads	no	yes	yes
Acts before the risk reaches the user	partly	no	yes
Doesn’t depend on the platform	yes	no	no
Keeps the decision with the user	yes	yes	yes

yes partly no

Product vision board

Vision

Enable people to browse safely on their phones, reducing the impulsive decisions that lead to clicks and data sharing, through UX micro-interventions.

Audience

Young adults aged 18 to 25

Lower-middle and low income

Instagram users, exposed to stimuli all the time

Needs

Spot warning signs and resist urgency triggers

Reduce vulnerability to phishing

Product

Redesign of mobile interactions

Neutralize triggers and impulsiveness

Focus on human behavior

Business

Left open in this phase; it came back as a feasibility question

Ideas from the ideation session

Discovery

Tone down terms like “LAST CHANCE”

Show when the advertiser’s account was created

Checking

Preview of the landing page

Warn when a logo imitates a brand

Engagement

A transition screen when leaving the platform

Show how long the advertiser has run ads without reports

Payment

A checklist before card or Pix payment

Warning for prices far below average

Aftermath

Reporting that explains how the scam works

Some of the ideas, mapped onto Mariana’s journey. This was the first version of the solution.

Scope: is, isn’t, does, doesn’t

Is

• A preventive cybersecurity solution
• An accessible product
• A safe product

Isn’t

• A platform
• A traditional antivirus
• A security consultancy
• A rigid blocker

Does

• Keeps phishing ads from reaching users
• Mitigates the effects of phishing
• Educates users

Doesn’t

• Stop users from deciding
• Invade privacy
• Replace moderation
• Guarantee zero risk

The proposal: ALINHA Sphere

A conceptual API would assess every ad against the guideline’s six pillars. A questionable ad goes live with friction, without being blocked.

1. InputAdvertiser submits
2. ReviewAPI checks the 6 pillars
3. Outputs
   • Sent back for changes
   • Published with friction
   • Published as usual

Six pillars, three tested

• Visual authenticity

  Not tested directly

• Destination transparency

  Tested

• Recent domain

  Tested

• Safety friction

  Tested

• Neutral language

  Not tested

• Contextual education

  Not tested

Tests and results

14 young adults, a simulated Instagram feed, and ads with and without the guideline.

How we tested

About the brands: the team created these ads for tests in a closed environment, in Portuguese. Real brands have been hidden here.

Feeling safe isn’t being safe.

What we’d change

• A warning that informsA warning that says what to do
• Hidden link destinationA truly visible destination
• Friction as one more tapFriction that prompts reflection
• Relying on the warning aloneActing before publication

See all results, validation by area and next steps

Described the guideline as an “extra layer of safety”

3 of 14

Reported feeling safer

2 of 14

• Automatic decisionValidated
• Visual heuristicsPartial
• WarningsPartial
• LanguageNot tested
• Data protectionNot tested

With 14 people, the numbers show trends; they don’t prove an effect.

Next steps

1. 01

   Test one guideline at a time

2. 02

   Measure perception, understanding and action

3. 03

   Simulate the sensitive-data step

4. 04

   Expand the sample

My role

A team project. I took part in every stage, and was most involved in four areas.

Closing

If design shapes decisions, those who design interfaces are also accountable for them.

Keep exploring02 / 05Web platform · Public sectorSefAIzTax incentives explained with learning paths, a simulator and AI.