A budgeting app that helps independent earners aged 25–30 stop relying on willpower and start saving by default — designed for the moments when motivation runs out.
01/ Overview
Why Vault exists.
42% of young adults in India earn enough to save, but don't — because saving requires daily decisions willpower can't keep up with. Vault makes saving the default, not the discipline.
7+
Money-related decisions a typical 25-year-old makes every day.
DAILY DECISIONS
3 mo
Average runway before "I'll save next month" becomes a habit of not saving.
PROCRASTINATION CYCLE
5x
More likely to save when the choice is removed and saving becomes automatic.
BEHAVIORAL LIFT
14B
Underserved market of independent earners aged 25–30 across top tier cities.
OPPORTUNITY
💡
The Thesis: if young earners are losing money to inertia automatically, savings can win the same way — by removing the daily decision entirely.
02/ The Problem
Money leaves silently. There's no moment of decision.
Users without auto-save
Months since salary
73%
of users have NEVER set up an automated saving
Decision fatigue.
When saving requires a monthly choice, 73% never make it. Willpower is not a UX pattern.
Anxiety level
Day 1Day 12Day 25
Day 12
the median moment users feel broke despite full balance
Mid-month dip.
82% felt broke before payday in the last 6 months — even those with ₹15K+ in their account.
Subscription bleed
Months since signup
₹1,800/mo
average ghost-spend on subscriptions left untouched 30+ days
Invisible bleed.
66% pay for ≥3 subscriptions they haven't opened in a month. Cancellation is saving.
03/ Research
Twenty-three conversations across four cities.
I ran a mixed-methods research study from Jan 8 to Feb 14, 2026: 23 semi-structured interviews (45 minutes each) with salaried professionals in Bangalore, Mumbai, Delhi, and Chennai, plus a quantitative survey of 142 respondents.
The hypothesis was that tier-2 / tier-3 first-time UPI users were the audience. Within six interviews it became clear I was wrong. The people most hurt were established metro earners with 3–5 years of income who had less in savings than at 23.
23
in-depth interviews
142
survey respondents
4
metros covered
6 wks
research window
Who I interviewed
DEMOGRAPHICS
N = 23 in-depth interviews · 4 metros · Jan–Feb 2026
N=23
USERS
25–27 years · early career
39%
27–29 years · senior IC
35%
29–30 years · early manager
17%
Outside age band (control)
9%
Behavioural findings
FROM N=23 + N=142
Check balance within 2 hours of payday
87%
Have NEVER set up an automated saving
73%
Currently have an active SIP they remember
31%
Felt 'broke' before payday in last 6 months
82%
Pay for ≥3 subscriptions untouched 30 days
66%
Saved successfully when money was auto-deducted
94%
04/ Insights
Five patterns I couldn't un-see.
Every pattern below was confirmed across at least 15 of 23 interviews and cross-checked with the 142-respondent survey.
What kept coming up
After 23 interviews and three rounds of synthesis, the same frictions kept surfacing. People had different vocabularies — but they were describing the same product gap. These pills are the exact phrases interviewees used.
Salary disappears in 5 days
I forget to save
Too many apps to open
I have an SIP, scared to check
Don't want a gamified app
Round-up isn't enough
Money locked = scary
Treats me like a child
Where did it all go?
Saving requires willpower
Auto-debit just works
Need it liquid
Mid-month panic
INSIGHT 01
73 → 94%
Saving rate
The decision is the failure point.
When saving requires a monthly choice, 73% never make it. When the choice is removed — auto-deduct, tax, employer SIP — 94% save by default.
INSIGHT 02
₹15K+
Median balance when 'broke'
People feel poor at full balance.
Mid-month anxiety isn't about being broke. Most users had ₹15K+ when they reported feeling broke — the feeling came from watching the number shrink, not from being short.
INSIGHT 03
4x
Story vs amount retention
Goals beat amounts.
Asked 'how much,' users gave numbers they forgot. Asked 'what for,' they gave specific stories: Berlin in October, mom's anniversary trip. Stories retained 4× better.
INSIGHT 04
₹1,800/mo
Average ghost-spend
Subscriptions are invisible bleed.
Average user had 7 active subscriptions, could name 3. Untouched-30+-days subscriptions averaged ₹1,800/month — more than most said they could afford to save.
INSIGHT 05
0
Apps for short-term liquid savings
SIPs solve a different problem.
The market view is 'people don't save.' The truer view: people don't have liquid short-term savings. SIPs are locked, long-term, emotionally inaccessible. Vault fills that gap.
INSIGHT 06
9–12am
Window of opportunity
The 1st of the month is sacred.
Salary landed on the 1st for 18 of 23 interviewees. By the 3rd, ~40% of monthly discretionary spend had already happened. Any nudge after noon on payday misses.
05/ Competitive analysis
The market has saving apps. None of them save for you.
I audited five Indian competitors against four criteria that mattered for the research findings: does the product make saving the default? Does it work without a decision per month? Does it surface invisible spend? And does it speak the language of now, not retirement?
PRODUCT
AUTO-SAVES BY DEFAULT
PRE-SALARY INTERCEPT
SUBSCRIPTION VISIBILITY
LIQUID (NOT LOCKED)
VOICE FITS 25–30
Jar
Round-up only
No
No
Locked in gold
Gen-Z heavy
Jupiter
Manual
No
Insights only
Yes
Yes
Fi Money
FIT rules
No
No
Yes
Yes
Niyo
Manual
No
No
Yes
Travel-led
Walnut / Cred
Tracking only
No
Yes
—
Premium-coded
Vault
Yes
Yes
Yes
Yes
Yes
Five competitors. None doing the one thing the research said worked.
ROUND-UP APPS
Jar
✕
Save change.
Meaningful amounts (₹6,500/mo at 10%) require intercept, not loose-coin sweeps.
BANKING APPS
Jupiter · Fi · Niyo
✕
Manual still.
FIT-rules and dashboards moved goalposts; the saving decision still falls on the user.
TRACKING APPS
Walnut · Cred
✕
Watch the bleed.
Visibility without action. Knowing where money goes ≠ stopping it from going.
Vault's wedge: moving meaningful money out of the user's reach before they spend it.
06/ Personas
Three users I keep designing for.
Personas are synthesised from interview clusters — not single users, but clear patterns. Each one represents a real combination of behaviour, income, and pain I heard repeatedly. Priya is the primary persona; Arjun and Neha are the design check.
PERSONA 01 · PRIMARY
Priya Ramanathan
26 · UX RESEARCHER · BANGALORE · ₹65K/MO
"I'm doing fine. I just don't want to think about money once a week."
GOAL
Save ₹2L for a Berlin trip in October. Keep a 3-month buffer if she switches jobs.
PAIN
Has tried Groww, deleted it. Has an SIP from 2023 she's scared to open.
CONTEXT
Salary on the 1st. By the 5th, ~₹35k spent (rent, EMIs, groceries, weekend).
WANTS
Money she can't see. Goal-shaped, not amount-shaped.
PERSONA 02 · CHECK
Arjun Mehta
28 · SOFTWARE ENGINEER · MUMBAI · ₹1.4L/MO
"I make a lot. I save like a college kid. I know it's broken. I just don't fix it."
GOAL
Down payment on an Enfield in 12 months. Stop feeling guilty.
PAIN
High income, high spend. 11 active subscriptions. Hates being "managed."
CONTEXT
Splits rent + EMIs across cards; salary day is for clearing dues.
WANTS
Set-and-forget. Will leave any product that asks too often.
PERSONA 03 · CHECK
Neha Krishnan
29 · PRODUCT MANAGER · CHENNAI · ₹95K/MO
"I want to feel like a grown-up about money. I don't want a gamified dog telling me to save."
GOAL
Build emergency fund. Reduce mid-month anxiety.
PAIN
Tried Jar, found it patronising. Mid-month checks her balance "3–4 times a week."
CONTEXT
Splits salary into 4 buckets manually. Wants automation, not infantilisation.
WANTS
A calm, adult product. Numbers, not characters.
07/ User Journey Map
From silent leak to silent save.
Priya's month, mapped from interview transcripts. The before-state is composite from interviews; the after-state is what Vault sets out to achieve. Emotion ratings are derived from think-aloud sessions.
The interface, not the intent, is broken.
PR
Priya
Age26
OccupationUX Researcher
Income₹65K/mo
LocationBangalore
GoalBerlin trip · ₹2L
Objective
As Priya, I want saving to happen automatically, so that I can build a 3-month buffer without thinking about it every month.
Task analysis
Receive salary on the 1st
Auto-intercept 10% to Vault
View progress on Berlin goal
Mid-month buffer reframe
Cancel ghost subscriptions
Withdraw if needed (2 taps)
User Flow
Day 1 → Day 28 · Salary intercept loop
Entry point
Salary lands
Vault intercept
₹6,500 → Vault
Push: "You won't miss it."
Home · 42%
Berlin goal
Day 1 · 8 sec interaction
DAY 12
Mid-month check
Feeling broke?
Show buffer
₹18,400 in Vault. Reframe lands.
DAY 18
Ghost spend
Cancel sub
₹299 saved
Hotstar untouched 47 days.
DAY 28 · LOOP RESETS
No reflection
Saving happened
Loop continues next month
Success criteria met
What changed?
The same five days, before and after Vault — measured by emotional state and friction count.
BEFORE
Without Vault
Day 1
Salary lands. Brief relief.
Calm
Day 5
₹38k spent. Unaware.
Unaware
Day 12
Surprise dip. "Where did it go?"
Anxious
Day 18–25
Restraint. Resentful.
Resentful
Day 28
Notion note: "Save next month." Doesn't.
Defeated
AFTER
With Vault
Day 1
Intercept push. 8 sec to confirm.
Satisfied
Day 5
Same spend cycle on ₹58.5k.
Unchanged
Day 12
Mid-month nudge. Reframe lands.
Reassured
Day 18
Ghost spend reveal. ₹299 saved.
Wins quietly
Day 28
Saving happened. Loop continues.
Default
08/ Decisions
Five forks in the road, with reasons.
Every design decision below has a rejected alternative. The shape of the product is the shape of the choices that didn't make it.
DECISION 01
The intercept moment
When salary lands, what does Vault do?
✓ CHOSEN
Modal sheet, 8 sec, dismissible
6/6 testers wanted acknowledgement; 5/6 said full screen felt "too much" for a passive event. Sheet hits the middle.
✕ REJECTED
Silent + push notification only
Tested as "suspicious — did it actually move?" (4/6). Trust requires day-1 visibility.
DECISION 02
The default amount
How much should Vault save by default?
✓ CHOSEN
10% of salary, slider 5–30%
Survey median was 12%. 10% is below the median (no regret) but above round-up threshold (meaningful).
✕ REJECTED
5% — "safer" default
Users increased the slider when default was 5% (felt unimpressive); decreased when >15%. 10% had highest stickiness.
DECISION 03
The brand voice
How should Vault talk to the user?
✓ CHOSEN
Calm · numerical · no exclamation marks
Neha's line — "don't want a gamified dog telling me to save" — repeated by 9/23. Calm voice scored 4.6/5 on "feels mature."
✕ REJECTED
Friendly mascot, encouragement, streaks
The Jar/Cred-coin direction. Tested as "treats me like a child" (6/6). Worked for users under 24, who weren't the audience.
DECISION 04
Withdrawal speed
How fast should the user pull money back?
✓ CHOSEN
2 taps, ≤2 hours, no questions
The "SIP fear" finding (Insight 05). 5/6 testers tested withdrawal within 5 minutes of onboarding. Liquid access = trust.
✕ REJECTED
24-hour cooling period before withdrawal
Intent was "protect users from themselves." Reality: emergency users would feel betrayed. Trust > paternalism.
DECISION 05
Goal framing
How should users describe what they're saving for?
✓ CHOSEN
Free-text name + optional amount + date
"Berlin in October" outperformed "₹2L by Oct 2026" 4× on recall (15 vs 4 a week later). Some goals resist exact targets.
✕ REJECTED
Required amount + dropdown category
Standard SIP-app pattern. People don't know exact amounts ("a trip… ₹1L?") and categories felt corporate.
09/ Usability Testing
What worked, what broke, what I changed.
I ran three rounds of moderated testing in Bangalore in February 2026 — six participants matched to the primary persona profile. Sessions were 60 minutes, think-aloud, recorded with consent. Tasks were drawn from the most-cited research pain points.
82.5
SUS Score
vs 68 benchmark
5/6
Task completion
P-3 KYC bug
6/6
Comprehension
intercept moment
3m 12s
Time to first save
median
Findings & fixes
Eight key findings from three testing rounds, with the design change made for each.
STATUS
FINDING
EVIDENCE
DESIGN CHANGE
✓ KEPT
Intercept moment
6/6 understood instantly
Kept the modal sheet pattern
✓ KEPT
Ghost Spend reveal
4/6 cancelled a sub mid-test (unprompted)
Promoted to home dashboard
✓ KEPT
Mid-month buffer reframe
Biggest emotional reaction (P-2)
Anchored to Day 12 nudge
✓ KEPT
Free-text goal naming
6/6 entered a real goal, not a placeholder
Made amount field optional
↻ FIXED
Skip-this-month flow hidden
3/6 couldn't find it
Moved to intercept sheet itself
↻ FIXED
"3-month buffer" preset misread
P-1 read as "3 months until full"
Renamed to "Emergency cushion"
↻ FIXED
Withdrawal screen too easy
2/6 worried about mistaps
Added single confirmation modal
○ OPEN
Slider default too low for high earners
P-Arjun fed back on 10% default
Investigating: scale default with income
Verbatims that mattered
"Wait. It just… did it? That's actually nice."
P-02 · 27 · intercept
"Oh I've been paying for Notion AI? I genuinely thought I cancelled that."
P-05 · 26 · ghost spend
"The whole thing feels like it's for an adult."
P-04 · 29 · closing comment
10/ Screens
The product, rendered.
Below: the three flagship screens. The full system has 14 screens across onboarding, core, goals, and management — built as a Figma file with auto-layout components and bound design variables.
01
CORE MECHANICSalary Intercept
A pre-set percentage of your salary moves to Vault automatically the moment it lands — before you ever see it in your spending account. No reminders, no decisions.
JUPITER POTSUser must remember to move money manually each month. Most stop after week 3.
VAULTSet once. Runs forever. No memory, no willpower, no chance to skip.
02
ONBOARDING MOMENTGhost Spend Reveal
During onboarding, Vault scans for recurring debits — Netflix, Spotify, Hotstar, gym memberships you've forgotten — and shows you in one screen what's leaving silently each month. High emotional impact, shown once, never again.
WALNUTConstantly notifies about every transaction. Notification fatigue kills it.
VAULTShows ghost spend once, with emotional weight. Then trusts you to act.
03
MOTIVATIONNamed, photo-anchored goals
Not "Emergency Fund." Not "Savings." Specific, emotional, photo-attached goals like "Berlin in October" or "3 months of rent if I quit." Goals work when they're tangible, not when they're abstract.
MOST SAVING APPSGeneric goal categories. "Travel," "Emergency," "Other." Cold, impersonal.
VAULTYour goal, your photo, your dates. Saves are deposits to a memory.
04
MID MONTH MOMENTMid-month nudge
Once a month — around Day 19–22, when paycheck-anxiety peaks — Vault sends a single, calm notification showing how much you've saved. Not to celebrate. To remind you that you're okay.
CRED, JARDaily streaks, push notifications, gamified urgency. Anxious by design.
VAULTOne mid-month message. Calm, confident. "You have 18,400 if you need it."
05
TRUST MECHANICLiquid by default
Money in Vault is always withdrawable. No 30-day lock, no penalty, no "are you sure?" friction. The trust comes from knowing you can take it out — not from being prevented.
JAR (GOLD)Locked into gold. Selling has friction, fees, 1–3 day settlement.
VAULTInstant withdraw to UPI. The choice not to spend is yours, every day.
