PRODUCT & STRATEGY MEMO · APRIL 2026

Most great trips
are group trips.

Every Indian friend group has a Splitwise tab and a UPI chain. Both end in the same place: arguments and unsettled balances three months later. The travel card sees every transaction. It's the obvious place to make group spending feel effortless. This memo proposes Scapia Splits.

By Gaurav Mankele IIT Bombay '27 Product & Strategy
scapia splits.
SCAPIA / 23B0435
FLIGHT GMA-027
SOLO
FROM · PAY & ASK
Direct 12 Weeks
SPLIT
TO · AUTO & SETTLED
Audience
Group travelers
Lever
Card-native
Network
Viral
CO-DESIGNED
WITH GROUPS
SCAPIA001
Honest Note

I built this without internal Scapia data and without access to a specific industry report on Indian group travel spending. The framings below are directional truths I can defend from first principles or universal user experience (most leisure travel is social, Splitwise is widely used, primary cardholders front group expenses, unsettled balances are common). Where I cite a number, I label it as an estimate. As an intern, my first job is replacing every assumption with your real numbers from Scapia's own transaction logs and cohort analysis.

The Setup

Group travel is the biggest unsolved UX problem in Indian fintech.

Every Scapia user is, on most trips, also paying for two or three other people. The card sees the transaction. The app doesn't know the trip exists. There's a Splitwise app for that. And a UPI chain. And a screenshot at midnight. This is the gap.

Group
Indian leisure travel is overwhelmingly social, not solo. Couples, families, friend groups. The card today assumes a single user, which is a mismatch with how the customer actually travels.
Directional. Would size precisely with Scapia's own trip-cohort data on day one.
Splitwise
A separate, popular app where Indian users already track shared expenses. The pain has a clear name. Nobody has solved it natively inside a payment product.
Splitwise app store presence in India.
Front
On most group trips, one person pays for hotels, flights, and shared dinners. That person's card spend rises sharply versus a solo trip. The economic asymmetry is invisible to the card today.
Universal user experience. Would quantify with Scapia's primary-cardholder transaction logs.
Decay
Group debts that don't get written down at the moment of payment commonly sit unsettled for weeks. Some quietly get forgiven. The friction tax is paid in trust and accuracy, not just rupees.
Universal user experience. Would size with internal data on repeat group trips.
A Real Trip

Here's a Goa weekend, transaction by transaction.

Four friends, three days, one Scapia card doing all the heavy lifting. The card knows every detail. The split conversation lives somewhere else, in screenshots and reminders.

Goa, March 2026
4 friends, 3 nights

Primary cardholder: Aanya
Group: Aanya, Riya, Vikram, Karan
Total trip spend: Rs 84,200
Day 1, 2:14 PM
Vivanta Panaji, 2 rooms, 3 nights

Aanya pays the full hotel bill on her Scapia card. The other 3 say "we'll split it later." Splitwise tab opened.

Rs 36,400
Aanya paid · 4 owe
Day 1, 9:08 PM
Thalassa, dinner for 4

Vikram pays this one on his card. Aanya texts: "split this 4-way pls." Group chat now has 2 different "to split" lists.

Rs 8,400
Vikram paid · 4 share
Day 2, 11:30 AM
Scuba diving, Grand Island

Karan and Riya only. Aanya pays for everyone (group rate cheaper). Two people owe 1 person. Different math.

Rs 11,000
Aanya paid · 2 owe
Day 3, throughout
Cabs, beachside lunches, drinks at Curlies

17 small transactions across 3 cards. Some on UPI, some on cards. Nobody has a clean record at end of day.

Rs 28,400
Mixed payers · chaos
Days after trip ends
23
Time it takes for the group to fully settle the trip on Splitwise + UPI.
Outstanding 30 days later
Rs 4,200
Vikram still owes Aanya. Vikram has forgotten. Aanya is too polite to ask.
Scapia rewards earned by group
Rs 0 extra
All splits go through UPI, not Scapia. The brand sees one transaction; the platform sees one user.
From First Principles

Seven irreducible truths that force Splits to exist.

If you accept these seven atoms, the solution is not a feature, it's a logical consequence. None depends on a specific market-size number. Each is either logical or anchored in universal user experience.

i

A group payment is four events, not one.

When Aanya pays Rs 8,400 at dinner with three friends, the card captures three events (authorization, transfer, receivable). The fourth event, the social debt to Riya, Vikram, and Karan, lives only in human memory.

ii

The cardholder is an unpaid intermediary.

On every group trip, the cardholder fronts credit, takes risk, settles debts, and keeps records. Banks charge for this work. The cardholder does it for free, badly.

iii

Debts not written down at the moment of payment decay.

Memory is not a settlement system. The longer the gap between event and settlement, the more friction accumulates. Some debts quietly get forgiven.

iv

Network effects come from peer-cluster awareness.

The value of a platform to user N rises with how many of N's relevant peers are on it. Scapia today has zero awareness of its users' travel networks. Splits captures that as a side effect of normal usage.

v

The cheapest acquisition moment is at an unsolved problem you solve.

A non-Scapia friend just got Splitwised by a Scapia friend. They have a real debt to settle, a working example of the product, and zero friction. That moment is not bought; it's a byproduct of an existing user paying for dinner.

vi

Switching costs are social, not technical.

The strongest moat in retail finance is the user's coordination problem with their peers. Once a friend group's settlement history lives on Scapia, leaving means rebuilding the social graph elsewhere.

vii

Card margins scale with volume per active user.

Splits drives more transactions through the primary cardholder, because they now front group spend instead of just their own. Same interchange margin, larger gross profit, zero new acquisition cost.

What this means. The proposal is not "build a feature." It's "capture the fourth event that already exists in the world but isn't yet in the database." Once that data exists, network effects, lower acquisition costs, and group-level switching costs follow as logical consequences. The Solution section below is just the user-facing surface of these seven atoms.

The Solution

Scapia Splits. Native to the card. Built for real trips.

Four mechanics that turn every Scapia transaction into a group-aware event. The app sees the trip, the group, and the math. The user just taps "split."

Every Scapia transaction asks one simple question:
"Was this just for you, or for the group?"

When the answer is "group", everything else (rewards, settlement, currency conversion, photo memory) follows automatically.

i

Tap-to-split at point of sale

Each transaction comes with a "split" prompt. One tap, pick your group, done. The split posts in the app instantly, before you've even left the table.

ii

Trip auto-detection

Geo + spending pattern triggers "Looks like you're on a trip." App pre-creates the trip, suggests the group based on past travel, lets you confirm in one tap.

iii

In-app settlement, not UPI

Friends settle balances inside the Scapia app via a single tap. Money moves over UPI rails behind the scenes, but the user never leaves the brand. Scapia stays the visible layer.

iv

Group-shared rewards

Scapia coins earned on shared transactions split proportionally to the group. Now everyone in your group has a reason to be on Scapia. Group trips become a group rewards engine.

In The App

Three screens. The split prompt, the group picker, the trip dashboard.

The whole feature lives in moments: at the table, at the gate, at the reckoning. Each screen earns its place by saving 10 minutes of friction.

scapia×
PAYMENT CONFIRMED · JUST NOW
🍽
Thalassa, Vagator Day 1 · 9:08 PM · Goa Trip
Rs 8,400
on your Scapia Card · XX-4521
Was this just for you, or for the group?
Just for me
Split it
Smart guess: Goa Trip · 4 people. Confirm in 1 tap.
At the table
scapia
Split Rs 8,400
Pick who shared this. Equal split unless you tap to adjust.
A
Aanya (you)
Pays Rs 2,100
R
Riya Sharma
Owes Rs 2,100
V
Vikram Mehta
Owes Rs 2,100
K
Karan Iyer
Owes Rs 2,100
Confirm Split
3 friends owe Rs 2,100 each
Group picker
scapia
Goa Trip '26
Mar 14-17 · 4 of you
Total trip spend
Rs 84,200
Coins earned
8,420
Group balance
RRiya owes you
+Rs 12,200
VVikram owes you
+Rs 9,800
KKaran owes you
+Rs 14,100
Settle now · via UPI
Trip dashboard
Feasibility

Can Scapia actually ship this?

Four lenses, honest answers. The conclusion is yes, with one regulatory pre-check, one behavioural hypothesis to validate in Phase 1, and zero changes to the existing card-side payment flow.

Technical

Can engineering build this in 16 weeks?

Tap-to-split is a UI pattern, not a payment innovation. Trip auto-detection uses geo and spend patterns the card already collects. Settlement runs on existing UPI rails. The hardest piece is the social-graph layer (who's in your travel group), which is one entity table and a relationship lookup.

Standard mobile-app engineering. No new payment infrastructure required.
Regulatory

Will RBI and Federal Bank allow it?

Settlement is just UPI under the hood, not a new payment instrument or a deposit product. Scapia's role is the orchestration UI above existing rails. The card-side flow does not change; we are adding a record-and-settlement layer above it.

Compliance review on day one. Most-likely path is no structural change. Worst case is repositioning Splits as a "shared-expense tracker with UPI integration."
Operational

Can Scapia run this without breaking existing flows?

Yes, in phases. Phase 1 ships to 1,000 power users in two cities with manual group setup. Phases 2 to 4 add intelligence, group rewards, and forex. Each phase is independently shippable and measurable. No phase blocks any other.

Phased rollout, no big-bang dependency. Each phase shippable in ~4 weeks.
Behavioural

Will users trust Scapia for splits instead of staying on Splitwise?

Not initially. The wedge is timing. When the prompt fires at the moment of payment, Scapia wins by default; Splitwise wins only if the user remembers to open it later. Users do not need to "switch tools." They need to stop opening a second tool because the first one already covered it.

Hypothesis to validate in Phase 1, measured by 30-day adoption (see Metrics section). Validated by data, not declared by founders.
The Network Effect

Each Scapia user pulls their group onto Scapia.

Splits is a viral install loop disguised as a feature. Aanya invites her group to settle a trip. Riya, Vikram, and Karan each download the Scapia app to receive their share. Two of them apply for the card within 30 days, because they've now seen the rewards math live.

This is acquisition economics that no paid channel competes with: CAC trends to near-zero on group-driven signups, while LTV climbs because each new user has 3+ social hooks already on the platform.

3-4
Estimated invitees per primary cardholder per trip. Based on typical group size, would validate against Scapia's data.
?
Conversion of invitees to cardholders. The metric to instrument from day one. Even a single-digit rate makes the math work.
A PRIMARY R V K D 1 trip = 4 invites = 12 second-order reaches
The Math

Why this is worth more than just one feature.

Splits doesn't just retain users. It cuts CAC. Here's the simplified napkin math, with placeholder numbers I'd refine on Day 1.

Step 01
Average primary cardholder takes some number of group trips per year
Ntrips
N is whatever Scapia's data says it is. The argument works at N as low as 2.
Step 02
Each trip exposes ~3 non-Scapia friends to the platform
3Nreaches/year
Group size is the cleanest assumption in this chain. Three other group members is typical for a couple-plus-friends or family group.
Step 03
Some fraction of those reaches convert to Scapia cardholders
C%
C is the metric to instrument from day one. Viral-loop fintechs typically see double-digit single-touch conversion when friction is low; the right reference set for Scapia would be Niyo's referral-driven funnel.
Net Result
Effective CAC for users acquired via Splits
low
A paid credit-card acquisition typically costs more than zero. The Splits-acquired user costs near-zero in marketing spend, because the touchpoint is generated as a byproduct of an existing user paying for dinner. Even a low C makes the wedge meaningful.
Metrics to Watch

What we would measure to know if this is working.

One north star plus three supporting metrics. Each maps to one of the seven atomic truths above, so we know what's being tested. Each has a fail condition, so we'd know if the bet didn't land.

North Star

Share of trip transactions that flow through Splits.

For a Scapia user on a group trip, what fraction of their card transactions during that trip get tagged via Splits? This is the single metric that proves Splits is the way Scapia users handle group trips. Below 10 percent at six months would mean the feature did not land.

Adoption · tests truth v
30-day Splits activation rate.

What fraction of active Scapia users have used Splits at least once in their first 30 days post-launch? Tests whether the in-the-moment prompt is reachable.

Fail. If under 15 percent activate within 30 days, the trigger UX is buried. Re-evaluate the prompt timing.
Network · tests truth iv
Non-Scapia friend conversion via settlement.

Of friends invited to settle a Split, what fraction download Scapia within 30 days? Of those, what fraction apply for the card within 60? Tests whether the social loop actually compounds.

Fail. If app-download rate is under 25 percent, the settlement-link UX is too heavy. If card-apply rate is under 5 percent of downloads, the social proof is not converting.
Business · tests truth vii
Spend lift on Splits-active cards.

For users who become primary cardholders on group trips, what is the lift in monthly card spend versus their pre-Splits baseline? Tests the revenue thesis directly.

Fail. If lift is under 1.5x on group-trip months, the revenue thesis weakens. Below 1.2x, the feature does not justify its engineering cost on a financial basis alone.

Why these three. Together they triangulate the strategic thesis. Adoption tests whether the in-the-moment prompt actually works as a behaviour change. Network tests whether the viral loop compounds at all. Business tests whether the revenue case is real. Each is independently falsifiable, each is a kill condition for one piece of the argument, and each can be measured cleanly without the counterfactual-modelling problems that retention-style cohort analysis introduces at this stage.

The Itinerary

Four stops. Sixteen weeks. Each one a real launch.

Don't bet the company on Phase 4. Each stop ships something users feel, in a small enough cohort that course-correction is cheap.

Departure · Wk 1-4

Splits Beta

Tap-to-split at POS. Manual group setup. Limited to 1,000 power users in Mumbai/Bangalore.

  • Split prompt UI
  • UPI settlement plumbing
  • Manual group lists
Layover · Wk 5-8

Trip Auto-Detect

Geo + spend pattern triggers trip mode. Suggested groups based on past co-travelers. Open to all card users.

  • Trip detection model
  • Past-trip group suggestions
  • In-app trip dashboard
Connection · Wk 9-12

Group Rewards

Scapia coins earned on shared transactions distribute to the whole group. Group leaderboard within Scapia community.

  • Coin split logic
  • Group rewards screen
  • Notification triggers
Arrival · Wk 13-16

Forex on Splits

International trips: shared expenses split in INR after live FX conversion. Single global flow, no fee guesswork.

  • FX-aware split engine
  • Multi-currency display
  • Open to all markets
Honest Risks

What can go wrong, and what to do about it.

Six likely failure modes. Each paired with a real mitigation, not a hand-wave.

High Severity

Splitwise inertia is real

Splitwise has trained a generation of urban Indians to expect a separate app for shared expenses. Muscle memory is a real switching cost, even when Scapia would be a better tool in the same moment.

MitigationSplits opens automatically on the card-holder's tap. The friend just clicks a link, no install required for the first interaction. Lower friction than asking them to switch tools.

High Severity

The group can't all be on Scapia

If 2 of 4 friends don't have Scapia, splits feel half-finished. Friction doesn't disappear; it just shifts.

MitigationSplits work without requiring all members to be cardholders. Settlement via UPI to non-Scapia members. The non-cardholder gets a smooth payment experience and a soft pitch for the card. Low pressure, high reach.

Medium

Split disputes get ugly

"I didn't have the dessert." "I wasn't there for that round." Splits in real life have edge cases.

MitigationAdjust-share UI at the group picker. Per-item assignment for restaurants. Quick "I wasn't there" toggle that re-balances the rest. Solve 90% of edge cases, accept the rest.

Medium

RBI / banking partner constraints

Federal Bank and RBI have rules around what apps can do with payment data and pseudo-settlement. Some of this might require careful regulatory mapping.

MitigationSettlement is just UPI under the hood. No new regulatory creature. Scapia's role is the orchestration UI. Worth confirming with compliance early, but no fundamental block.

Low

Niyo or Federal Bank's own app copies the model

Other Indian travel-fintech could ship a similar feature within months.

MitigationThe moat is the social graph, not the feature. First-mover captures the group networks. After 50K+ groups are on Scapia, switching means rebuilding your travel social graph elsewhere. Not impossible, but real.

Low

Adoption is slow if travel volumes dip

Splits scales with trips. A travel slowdown stalls the network effect.

MitigationSplits also work for non-trip group spend (dinners, weekend plans, recurring groups). The trip case is the wedge, but the use case is universal. Falls back gracefully to "everyday splits" if trip volumes dip.

Let me build
the first split.

I'd love to intern at Scapia on product/strategy. First task: replace every assumption in this memo with your real numbers, then ship Splits MVP in 4 weeks. Talk to me.

gauravmankele.iitb@gmail.com