Oppam — UX Case Study by a Product Designer

This is a long one, so here is its shape. Why people in Kerala do not come. What the apps in this category already do, and where they stop. What Oppam has today, which is a website. Then the app I designed for it, screen by screen, and the reasons behind each part of it.

ഒപ്പം. It means beside you.

Kerala, and the walk to the door

Kerala has more psychiatrists per person than any other state in India. About 1.2 for every 100,000 people, where the country as a whole manages 0.75.

It also has one of the highest suicide rates in the country. 30.6 for every 100,000 people in 2023, against a national average of 12.3. Third highest of any state or union territory.

And the National Mental Health Survey still puts the treatment gap at 84.5 percent. The best served state in the country, and most people who need help are not getting any.

Kerala in three numbers: 1.2 psychiatrists per 100,000, an 84.5 percent national treatment gap, and 30.6 suicides per 100,000 in 2023 against a national 12.3 — The state with the best coverage in the country.

Psychiatrists per 100,000 people: Kerala 1.2, India 0.75, a workable minimum 3.0 — Leading India, and still under half of a workable minimum.

So the wall is not a shortage of therapists. The wall is the walk to the door.

What stops people here is quieter than a waiting list. Therapy is whispered about. A clinic is a building other people can watch you walk into. Phones get shared with a sibling or a parent. And nobody knows what actually happens in a first session, so the imagination fills it in with something worse than the truth.

I should say plainly what this is built on. There are no user interviews behind it. It is the National Mental Health Survey and the Kerala paper drawn from it, published work on shared phones in Indian homes, a read of the live oppam.me service, and a walk through every competing app from first open to booking, marking each place I hesitated and why. The people I kept in mind while deciding are archetypes from that reading, not people I sat with. If there were a budget I would replace them in the first week.

What the apps in this category already do

The category is neither empty nor bad. Wysa has reached eleven million lives across a hundred and five countries and sits at 4.8 stars. Amaha has served four and a half million users. YourDOST puts counsellors on text, audio and video. These are real products with real reach.

They are also open about their hardest problem, which is that people leave. Retention is the known weak point of the whole category. Language is the other one. An app that only speaks English is asking somebody to describe their worst week in their second language.

But look at where they all begin. Open each one and the first screen is a chatbot introducing itself, a pitch for your wellbeing, or a logo followed by a sign-up.

The first App Store screenshot of Wysa, Amaha and YourDOST — The first thing each app shows you.

The ten minutes before a person decides to ask for help is nobody’s screen.

Price is the other thing worth looking at, because most of this category was built for a different economy. BetterHelp runs 70 to 100 dollars a week, billed monthly. A subscription is a commitment, and a commitment is the exact thing somebody unsure is not ready to make.

BetterHelp at 70 to 100 dollars a week beside Oppam at 1,000 rupees a session — A subscription, against a price you pay once.

Wysa, Amaha and YourDOST with their reach: 11 million lives across 105 countries, 4.5 million users, counsellors on text, audio and video — Three of the biggest, and where each of them starts.

That is the gap this app is built into.

What Oppam has today

Oppam is Kerala’s first 24/7 online counselling platform, run out of Kozhikode. The site counts more than 50,000 therapy hours and a 4.8 rating on Google. Individual, couple and follow-up sessions, plus corporate and campus programmes, across the things people actually arrive with: anxiety, stress, depression, trauma, grief, anger, work burnout, parenting, queer-affirmative concerns.

The oppam.me homepage: a sunrise illustration, 24x7 online therapy, and a consult a therapist button — oppam.me as it stands today.

Oppam has raised 1.5 crore in seed funding led by Phoenix Angels and 90 lakh in an extended round, delivered over 50,000 therapy hours, and rates 4.8 on Google — Money, clients and hours behind it.

It raised 1.5 crore in a seed round led by Phoenix Angels, then 90 lakh more in an extension. It runs Oppam Academy, training psychologists out of the Calicut office. This is a working company rather than a hypothetical one, and that decides what a redesign is allowed to be: something the people already using it would recognise on the first open.

And the website is already good. That is worth saying plainly, because it changes what the app is for.

Therapist cards on the website showing photo, credential, concern tags, a voice introduction player, languages, therapy hours, price per session and a book now button — One card on the website, and everything it already carries.

Look at what a card already holds. Her photo and her credential. The concerns she takes. A voice introduction you can play before you decide. The languages she works in. The hours she has done. The price per session in rupees. When she is next free. One button to book her.

Nothing important is hidden, and none of it needs a phone call first. So the app is not here to rescue a broken website. Most of the good ideas in it started here, and it keeps them: a price on every card, a voice before the money, the language named up front.

What an app can do that a page cannot

Three things, and they are the whole reason this project exists.

Privacy. A browser keeps history. A tab can be read over a shoulder. A shared phone has no Face ID door, no cover in the app switcher, and no way to make a notification say nothing. That is the barrier the research keeps pointing at, and it is the one a website structurally cannot answer.

It is not a theoretical worry in this category either. In 2023 BetterHelp, the biggest name in online therapy, paid 7.8 million dollars to settle with the US Federal Trade Commission over sharing email addresses, IP addresses and the answers people gave to health questions with Facebook, Snapchat, Pinterest and Criteo, for advertising. Around 800,000 people were eligible for a refund.

BetterHelp paid 7.8 million dollars to settle with the US Federal Trade Commission over sharing health answers with advertisers — The exact thing this design is built to make impossible, already done at scale.

So privacy here is not a policy page. It is the reason the app holds everything on the device and asks for no account at all.

The part before choosing. The website still opens onto a grid of therapists, which asks a person to evaluate strangers while anxious. An app can start from how somebody feels and bring three people forward with the reason for each.

A reason to open it on a day you are not booking. A website’s job ends at the booking. An app can hold a two minute breath, a page nobody else reads, a small win from last week. That is what keeps a person near help long enough to eventually ask for it.

So the app is not the website with a tab bar bolted on. It is the three parts the website was never going to hold.

The sentence everything was tested against

Oppam is not a therapy app. It is the feeling of someone quietly coming to sit beside you, and its real job is the moment before somebody asks for help, not the transaction afterwards.
The one line every decision had to survive

Four rules came out of that, and between them they settled most of the arguments.

Four rules: trust before identity, one loud action, visible not promised, and presence not performance — Pinned above the desk for the rest of it.

The first minutes

A sign-up wall feels like registering an illness, so there is not one. The app asks what to call you, and accepts any answer, including none.

The opening screens: the welcome, the what should I call you question, and the language chooser — A name, and a language.

The language question is not a settings toggle here, and it is not only about labels. Feelings do not translate well. Malayalam, English or Manglish switches the lines that carry weight, not the chrome.

Then it asks, gently, whether anything is sitting with you lately. Every answer is optional and one of them is nothing right now. It ends on a promise rather than a permission request.

Two onboarding screens: is anything sitting with you lately, and before we begin, one promise — Nothing is compulsory, including answering.

Home is a veranda, not a dashboard

A dashboard asks you to manage something. A veranda is a place you can sit without a reason, which is closer to what this app is for.

The Oppam home screen with a sky, the companion, a greeting and the next session card — Home, in the morning.

The sky and the companion follow the hour, so the app feels like it is in the same day you are. After nine at night the companion sleeps and holds a lantern.

The same home screen at morning, afternoon, evening and night, the sky and companion changing with each — One screen, four times of day.

The check-in asks for weather, not symptoms. Is my problem big enough is one of the questions that keeps people out, and a symptom checklist answers it badly. Weather has no threshold to clear.

Small wins collect underneath, and they only ever grow. Nothing here can go down, because a number that falls is a reason to stop opening an app.

The quiet moment card and the small wins list on the home screen — A quiet moment, and things that went right.

The quiet moment is a two minute breath on a 4-2-6 count. It is the one thing in the app you can finish.

The breathing screen with the companion and the words breathe in with me — Two minutes, and no account needed.

Most of this part only exists in motion, so this one is the app running rather than a screenshot of it.

One full breathing cycle animating, captured frame by frame from the simulator — One cycle, in and out.

Finding a person

A directory asks the hardest question first: evaluate a screen full of strangers and choose correctly, while anxious. So Care starts from a feeling instead.

The Care screen listing feelings such as anxiety and worry, relationship issues, grief and loss — Start with what is going on, not with a list of people.

Then it brings three people forward and says why each of them. Three is a number you can actually decide between. The full list is one tap deeper for anyone who wants it.

Introductions, not listings.

Every therapist answers the same question in their own recorded voice: what your first session is honestly like. The unknown first session is the tallest wall in this whole thing, and a voice gets over it in a way a paragraph does not.

Her voice, before your money.

The price is on the card. Per session, always, no packages. Cost opacity is its own barrier, and the answer to it is a number you can see before you are invested.

The booking screens: choosing a rhythm with prices per session, and picking a time — A rhythm, a time, a price you already know.

The booking review screen and the confirmation — Four steps, no forms.

The confirmation ends on the thing people are actually anxious about, which is not the calendar entry.

The you are booked screen reading nothing to prepare, come exactly as you are — “Nothing to prepare. Come exactly as you are.”

The room itself says what it does with the session, which is nothing. Video, voice or chat, and camera off is a normal choice rather than a failure.

The live session screen with the therapist and an end call button — The session.

Pages, which nobody else sees

Writing is the cheapest first step there is, and it does not need a therapist or a payment. Pages is a journal with the prompt turned off by default, because a prompt is a small exam.

The Pages screen with the garden of small wins and the note that these pages are only yours — Pages.

Letters can be sealed until a date you pick, which turns writing into something you do for a version of yourself that does not exist yet.

Writing a page, and a letter being sealed for a future date — A page, and a letter to later.

Privacy you can watch working

A privacy policy is a promise. A shared phone needs proof. So every protection in here is something you can see doing its job.

The You tab with the language of the heart switcher and the privacy room — The You tab is deliberately small.

Face ID on the journal door. A sunrise that covers the app in the switcher. Notifications that show you the exact text somebody else would see on your lock screen. One tap turns Pages into a grocery list.

The privacy room with a lock on the journal, neutral notifications with a live preview, and quiet hours — Each one shows what it actually does.

The companion

The mark in the logo is a smile, and the companion is that smile alive. It is drawn as animated geometry rather than an image, so it can breathe, blink, and sit one register softer than whatever you are feeling.

It never mirrors distress and never celebrates louder than you do. It never speaks first. When it does speak, the words are set in a serif, like a letter rather than a chat bubble.

The five companion registers: peaceful, listening, happy, comforting and sleeping — Five registers, one face.

The other half, which is the therapist

A marketplace that designs only the buyer’s side quietly decides that the supply side does not matter. Therapists are the supply here, and burnt out therapists are how a service like this fails. So the console is built too.

The therapist console today screen with the live session and the client who has arrived — Today, for the therapist.

People rather than cases. A file shows a first name and only what that person chose to share, including the words they typed on day one.

The console today, people and calendar screens — The week, and the people in it.

Hours are opened by the therapist, not assigned to them. Earnings show the split written out rather than a net figure that needs trusting.

The practice screen with earnings and the hours the therapist has opened — The practice.

Notes are written the way you would say it to a colleague you trust, and an unfinished note is not a backlog with a badge on it.

Session notes and the unfinished notes list — Notes.

The earnings screen with the split written out and the payout history — Earnings, with the split shown.

And the part nobody puts in a marketplace: an on-call clinical lead, peer supervision, and an escalation protocol that says what to do when somebody is not safe.

The support screen with an on-call clinical lead, peer supervision and the escalation protocol — Nobody holds this alone.

The bad days are designed too

Anything that only works when the network does is a prototype. These are the screens the app spends its worst minutes on.

The offline ribbon sliding in under the status bar with an explanation screen — Offline says so, quietly.

Three error states: cannot load right now, the payment did not go through, the connection dropped — Reassurance first, one calm retry second.

An empty screen is never a dead end. The companion waits in the blank space with a line and a way forward.

The nothing booked yet empty state with a find someone button — Empty, with a way out of it.

More states: offline, no sessions yet, and cannot reach the call — And the rest of them.

The something went sideways error screen — Even this one is written like a person.

Lamplight

Dark mode here is not a grey inversion. It is lamplight: warm browns and amber, because this app gets opened late at night by people who are not having a good one, and cold grey at 2 AM is a worse room to sit in.

The home screen in dark mode with a warm night sky — The night veranda.

Care, after dark.

Pages, You and the privacy room in dark mode — Pages and the privacy room.

The breathing screen and the morning home screen in dark mode — A design system is only real once the quiet screens hold.

What this is, and what it is not

It is a running SwiftUI app on iOS 18, with the data on the device and nothing in a cloud. Every figure above is a capture of it, taken by launch argument so the same screen comes back the same way each time.

The Oppam app running on a phone held in one hand — On a real phone.

What it is not is validated. The archetypes are mine, the priorities are mine, and no part of this has met a person who came looking for help. That is the honest limit of a self-initiated project, and it is the first thing I would fix.

Deliberately left for later: the home screen widget, a live activity for the session countdown, Apple sign in, and a night companion that knows how badly you slept. The first of those is a weekend. The last one needs a conversation about health data that this prototype has not earned yet.