Groww SIP Shield: Fixing the Correction-Cancellation Trap.
Groww is India's largest retail broker by active NSE clients (26.27% market share, DRHP FY25; current share may vary - NSE publishes updated data quarterly) and dominates the direct mutual fund distribution channel. Direct-plan SIPs are placed without distributor intermediation; Groww's 39% direct-plan share reflects its positioning as a self-directed platform - a separate universe from the distributor-led traditional MF market. But size creates a specific vulnerability: when markets correct, Groww absorbs the largest share of retail SIP cancellations in the industry.
The problem is not that users cancel SIPs.The problem is that Groww's cancellation flow is SEBI-compliant but behaviorally silent - it confirms cancellation in 30 seconds without surfacing a single piece of context that might change the decision. Users are cancelling at exactly the wrong market moment, and the product is not in the room when it happens.
This memo proposes SIP Shield- a three-part intervention: extending Groww's existing (but capped) SIP pause into a true multi-month pause, a contextual intervention screen at the point of cancellation during market corrections, and a one-tap “Dip Boost” that lets a user top up at the lower NAV instead of leaving. All three are SEBI-compliant. None blocks cancellation. Together they address the three distinct cancellers who drive Groww's correction-period SIP churn.
The Correction-Cancellation Trap
India's SIP ecosystem has a structural behavioral problem that worsens every time markets correct. AMFI data shows the pattern clearly - and Groww, holding 18.5% of industry SIP accounts, absorbs more of it than any other platform.
SIP Stoppage Ratio - AMFI data (discontinuations ÷ new registrations)
| FY22 (bull market) | 41.74% | Post-COVID recovery. Nifty at all-time highs. Healthy baseline - most new investors stay invested. |
| 2022 (bear year) | 51% | Rate hike cycle begins. FPI selling. New retail investors face their first bear market - 10pp jump in exits. |
| FY24 | 52.41% | Stable market. Ratio plateaus with no major correction to trigger panic exits. |
| January 2025 | 109% | First month in AMFI history: SIP cancellations exceeded new registrations. Nifty ~11% below September 2024 peak. |
| April 2025 [*] | 296% | [*] Not directly comparable: SEBI-mandated cleanup of dormant folios inflated registrations-to-cancellations ratio. Included for completeness; should not be cited as a behavioural cancellation peak. Underlying voluntary cancellation trend remained elevated. |
Sources: AMFI monthly data (amfiindia.com); Business Standard reporting on AMFI FY21–22 data. A ratio above 100% means cancellations exceeded new registrations that month.
Why this is Groww's problem specifically
Groww holds 18.5% of active SIP accounts and processed 39% of all direct-plan SIP inflows in FY25 (DRHP). At industry scale, that means Groww absorbs roughly one in five SIP cancellations during every market correction. With 36% of new FY25 users starting with SIP as their first product (DRHP, Q2 FY26 data), SIP retention is the entry point to the entire platform relationship - not just a mutual fund metric.
A user who cancels their SIP during a correction is not just a lost SIP. They are the highest-risk cohort for full platform churn - closing the demat account, stopping stock trades, uninstalling the app. The lifetime value loss is an order of magnitude larger than the cancelled SIP installment.
The product gap
Current Groww SIP cancellation flow
User sees −14% on portfolio dashboard
Market is 11% below its 52-week high. The red number is the only signal the app provides.
Taps "My SIPs" → taps the fund → taps "Cancel SIP"
Three taps. No friction. No context.
Standard confirmation: "Are you sure?"
A generic modal. No RCA data, no goal impact, no pause alternative.
SIP cancelled
Done in under 30 seconds. The user locked in their exit at the bottom of the correction.
Groww
SIP pause exists but is capped - skip up to 2 consecutive installments; the 3rd auto-cancels the SIP. No true open-ended pause, no guaranteed auto-resume, no contextual screen at cancellation.
Kuvera
SIP pause: 1–6 months with automatic resumption. Users facing temporary cash-flow issues can pause without cancelling.
Why Users Cancel at the Worst Possible Moment
Three users, three different problems - but the same product failure. Groww's cancellation flow treats all three identically.
Aarav Kumar
Software Engineer · First-time investor, SIP started mid-2024
"My portfolio was down 14%. I cancelled - felt like I was throwing ₹3,000 into a hole every month."
✗Entered markets at the 2024 all-time highs; first correction arrived within 6 months
✗No in-app context for what a 14% drawdown means across a 10-year SIP horizon
✗Cancellation took 30 seconds - the app confirmed it without a word about rupee-cost averaging
Priya Menon
Freelance Designer · Irregular monthly income
"I needed to stop the SIP for 2 months. I couldn't find the pause option, so I just cancelled it."
✗Genuine cash-flow constraint - not loss aversion, just temporary liquidity pressure
✗Groww's skip-installment caps at 2 consecutive skips - the 3rd auto-cancels the SIP; she needed a true multi-month pause with guaranteed resume
✗Cancelled a 14-month-old SIP because the pause UX was unclear - not because she wanted to stop investing
Rajan Sharma
Mid-level Manager · 3 years of investing experience
"I know I shouldn't cancel. But watching ₹3,000 leave my account every month when I'm already down ₹45,000 is psychologically brutal."
✗Understands the theory - still feels loss aversion acutely at each installment
✗No dashboard element shows how rupee-cost averaging is working in his favour during the correction
✗Would stay invested if the app showed "this month's SIP bought 18% more units than September"
Behavioral foundation
Loss Aversion - Kahneman & Tversky, Prospect Theory (1979)
Losses feel 2–2.5× more painful than equivalent gains feel pleasurable. A ₹45,000 paper loss feels far worse than a ₹45,000 gain feels good - even though neither has been realised. This asymmetry drives Aarav's cancellation: the monthly ₹3,000 SIP installment feels like compounding a loss, not buying more units at a discount.
The Rupee-Cost Averaging Paradox
A market correction is precisely when SIPs work best. Lower NAV = more units per rupee = lower average cost basis. The investor who cancels at −14% has not yet lost anything - but they exit before the recovery accrues to them. The product's failure is that it shows users the portfolio loss and nothing else. The RCA gains happening simultaneously are invisible.
Nudge Theory - Thaler & Benartzi, Save More Tomorrow (2004)
Contextual information and well-designed defaults materially improve long-term savings adherence without restricting choice. In the SIP context: showing RCA progress and offering a pause option does not force anyone to stay invested - it ensures the decision is made with full information rather than only the red number on the dashboard.
SIP Shield: Three Components
SIP Shield is not a single feature - it is three independent product decisions that together address all three canceller personas.
SIP Pause - extend the existing cap into a true 1–6 month pause
Groww already lets users skip installments - but it caps at 2 consecutive skips, auto-cancels on the 3rd, and offers no guaranteed auto-resume. Extend it into a genuine 1–6 month pause: the user picks a duration, the SIP auto-resumes with no action required, and a pre-resume notification fires 5 days before restart. This is what removes Priya's reason to cancel instead of pause.
Solves for
Priya - genuine cash-flow pressure. She does not want to cancel; she wants to stop temporarily. The pause converts a cancellation into a pause.
Ship condition
Low risk - this extends an existing feature, it doesn't add friction. Ship to 100% in Phase 1, independent of the Shield screen test.
SIP Shield Screen - contextual intervention at cancellation
When a user initiates SIP cancellation and the Nifty is ≥7% below its trailing 52-week high, intercept with one screen of factual context before the final confirmation. The screen shows three elements:
CTAs on the Shield screen
Three exits, none mandatory: lean in (Boost), step back (Pause), or leave (Cancel). Cancellation still completes normally if the user proceeds - it is never blocked.
Dip Boost - the original idea: turn the cancel moment into an opportunity
Every retention screen in fintech plays defense - “are you sure you want to leave?” Dip Boost inverts it. On the same Shield screen, during a correction, offer an optional one-tap top-up at today's lower NAV: “Add a one-time ₹3,000 this month - at today's price that's ~18% more units than your September installment.” The user who opened the flow to stop investing is offered a way to lean in instead.
Why it's new
No mainstream Indian broker turns the cancel moment into a buy-the-dip prompt today. It reframes loss aversion into agency - and it is the only component that can grow AUM from a churn moment rather than just defend it.
Compliance guardrail
Strictly opt-in, never a default, never pre-checked. Copy stays factual (units at current NAV) - never advisory (“you should invest more”). Same SEBI line as the Shield screen; legal review before shipping.
After SIP Shield - proposed flow
User sees −14% on dashboard
Same red number. Nothing has changed here.
Taps "My SIPs" → fund → "Cancel SIP"
Same three taps.
SIP Shield screen fires (correction trigger active)
Market context, RCA gains, goal impact, and pause CTA shown for the first time.
User chooses: Boost, Pause, or continue cancelling
Three respected paths. No mandatory step added to the cancel flow.
Outcome
Priya pauses 2 months - auto-resumes. Aarav taps Dip Boost and buys ~18% more units than September. Rajan still cancels - but with full context. Three different right answers.
Experiment: SIP Shield Screen
The pause feature ships without an experiment. The Shield screen requires one - because the risk of introducing friction into a legally mandated easy-cancellation flow must be measured before full rollout.
Hypothesis
“Users who see the SIP Shield screen - showing RCA gains, goal impact, and a pause option - before the final cancellation confirmation will cancel at a 20% lower rate during market correction periods (Nifty ≥7% below its 52-week high) compared to users who see only the standard confirmation screen. The guardrail - support tickets citing blocked cancellation - will remain below 0.1% of treatment users.”
🟢 Treatment
SIP cancellation attempt → SIP Shield screen (RCA data + pause option + goal impact) → user chooses pause or proceeds to standard confirmation → outcome recorded.
⬜ Control
SIP cancellation attempt → standard "Are you sure?" confirmation screen → SIP cancelled. No Shield screen shown.
Test parameters
| Duration | 8 weeks minimum - must span a ≥7% correction period to generate sufficient treatment events |
| Split | 50/50 user-level randomization. Sticky assignment across sessions. |
| Trigger | Nifty ≥7% below its trailing 52-week high at time of cancellation attempt |
| Scope | All SIP cancellation attempts during the test window |
| Control | Standard cancellation confirmation - no intervention screen |
| Treatment | SIP Shield screen (RCA gains + pause option + goal impact) before final confirmation |
Experiment metrics
| SIP cancellation rate (correction periods) | Primary | −20% vs control |
| Pause adoption rate | Primary | >25% of users who see Shield screen |
| Dip Boost adoption rate | Secondary | >8% of Shield-screen users during a correction |
| Support tickets: "cancellation was blocked" | Guardrail | Near zero - <0.1% of treatment users |
| SEBI / AMFI regulatory complaints | Guardrail | Zero |
| 90-day SIP retention (treatment vs control) | Secondary | +15pp |
| Time-to-cancel (counter-metric) | Counter | +~30s - expected; this is the mechanism, not a side effect |
Edge cases handled
User wants to cancel for a legitimate reason (switching platforms, financial emergency)
Shield screen does not block cancellation. User can dismiss and proceed to cancel in one additional tap. The intervention respects user intent - it only ensures the decision is made with full context.
Market recovers mid-test, correction threshold drops below 7%
Trigger logic checks real-time Nifty price vs 52-week high at the moment of each cancellation attempt. If the market has recovered, the Shield screen does not fire. Experiment continues; correction-period data accumulates separately.
User force-closes app during Shield screen
Cancellation does not process until the user explicitly confirms on the standard confirmation screen. App restart shows the Shield screen again if the correction is still active.
User in the control group sees paused SIP behavior from treatment group (social contamination)
Randomisation is user-level (not session-level) and fully sticky. Control users never see the Shield screen. Cross-contamination risk is low - SIP cancellation is a private, individual action.
Success Metrics
🎯 North Star
SIP 12-month survival rate
The percentage of SIPs started in any given month that are still active 12 months later. This is the product's core promise - that a user who starts a SIP on Groww stays invested through at least one full market cycle. Everything else is a leading indicator of this number. Target: >60% of SIPs survive to month 12.
Full metrics stack
| SIP 12-month survival rate | Percentage of SIPs started in any month that are still active 12 months later. The long-term health signal. | >60% |
| Cancellation rate during correction windows | North Star for this intervention. Measures if Shield screen is working when it matters most. | −20% vs pre-intervention baseline |
| Pause feature adoption rate | Whether users are choosing pause over cancel. High rate = Priya persona problem is real and solved. | >20% of SIP management interactions |
| Shield screen bypass rate | Users who dismiss the screen and cancel anyway. >60% bypass = wrong copy or wrong targeting threshold. | <50% during correction periods |
| SIP-to-demat cross-sell rate (12 months) | SIP is the acquisition hook; equity trading is the revenue engine. Measures platform retention flywheel. | >28% of 12-month SIP holders open a demat |
| Goal progress card engagement | Whether users interact with the goal-impact element on Shield screen. Informs copy and personalisation decisions. | >35% of Shield screen impressions |
Rollout Plan
Phase 1 - Extend the Pause Feature
Weeks 1–3- - Extend Groww's existing skip-installment (capped at 2 consecutive skips, auto-cancels on the 3rd) into a true 1–6 month pause with guaranteed auto-resume - matching Kuvera. Low risk, no experiment needed.
- - Monitor pause adoption rate and support ticket volume for the first 2 weeks
- - This phase is independent of the Shield screen - ship it regardless of experiment outcome
Phase 2 - SIP Shield A/B Test
Weeks 4–12- - Implement real-time correction trigger: Nifty price vs trailing 52-week high, refreshed every 15 minutes
- - Run Shield screen experiment: 50/50 split on all SIP cancellation attempts
- - The Shield screen carries all three options - Dip Boost, Pause, Cancel; track tap-share across them to learn whether users lean in, step back, or leave
- - Monitor guardrail metrics daily - kill criterion: if "blocked cancellation" tickets exceed 0.3%, pause test and revert
- - Do not declare success until the test has run through at least one full correction window (Nifty ≥7% below 52w high)
Phase 3 - Personalisation + Full Rollout
Month 4+- - If test succeeds: ship Shield screen to 100% of SIP cancellation attempts, not just during corrections
- - Personalise by SIP age: new investors (<6 months) see more RCA education; experienced investors see less text, more data
- - A/B test Shield screen copy variants: "Your units this month cost 18% less" vs goal-framing vs neutral RCA summary
- - Proactive notification variant: push alert when Nifty crosses −10%, before users even open the cancellation flow
What I'd Push Back On
Is this manipulation?
The line between persuasion and manipulation in fintech is real and worth taking seriously. SIP Shield does not block cancellation, add mandatory steps, or create artificial delays. It adds one screen of factual, verifiable data - how many units you have accumulated, what your goal timeline looks like - and offers a genuine product alternative (pause). A user who reads the screen and still cancels made an informed decision. That is the right outcome. The product respects user autonomy; it just ensures they have information they demonstrably did not have before.
SEBI compliance is non-negotiable - and this respects it
SEBI's investor protection framework requires easy SIP cancellation. SIP Shield does not compromise this: cancellation is not blocked, the screen adds zero mandatory steps, and the copy must be strictly factual ("your average cost this month is ₹X") - not advisory ("you should not cancel"). It cannot qualify as investment advice, which requires a registered investment adviser licence. Legal review before shipping is mandatory. The copy needs SEBI-safe language from day one.
The pause gap is narrower than it looks - but it is real
I want to be precise: Groww does have a pause - users can skip up to 2 consecutive installments, but the 3rd auto-cancels the SIP and there is no guaranteed auto-resume. Kuvera offers a true 1–6 month pause with auto-resume (verified May 2026). So the gap is not "pause doesn't exist" - it is duration, the auto-cancel cliff, and discoverability. Extending the existing pause is a low-risk Phase 1 ship, not a net-new feature, and its success is measured separately from the Shield screen A/B test.
What I'd validate before engineering starts
Before the Shield screen goes into a sprint: interview 15–20 users who cancelled a Groww SIP in the last 6 months. Ask specifically: would seeing your RCA gains have changed your decision? Would a pause option have been enough? The intervention assumes most cancellations during corrections are driven by loss aversion or friction - if exit interviews reveal most users cancelled because they were switching platforms or had a genuine financial emergency, the product design changes significantly. Discovery first.
What I set out to show
I approached a retention problem rather than an activation one - a less common angle in fintech PM work. I applied behavioral economics to a real, AMFI-backed issue and stayed within SEBI's compliance constraints throughout, because they're the constraint that makes or breaks any intervention in a regulated product.
What I haven't addressed
Long-term cohort analysis: do paused SIPs resume and stay active, or just delay cancellation? Whether Dip Boost genuinely helps users or simply increases exposure at a volatile moment - that is an ethics-and-outcomes question I would test, not assume. And cold-start: users with no goal set, where the goal-impact card falls back to a generic projection.
What I'd validate first
Before engineering: interview 20 users who cancelled a Groww SIP in the last 6 months. Ask: what would have changed your decision? The whole intervention assumes loss aversion and missing context are the primary drivers. If most cancellations are due to platform switching or genuine financial hardship, the design changes significantly.
Sources & References
- [1]Groww (Billionbrains Garage Ventures Ltd). Draft Red Herring Prospectus (DRHP) filed with SEBI, September 2025. sebi.gov.in. Primary source for all Groww-specific financial and user metrics: revenue (₹3,901 Cr operations revenue, ₹4,061 Cr total income FY25), PAT (₹1,819 Cr FY25), SIP inflows (₹34,028 Cr FY25), SIP market share (18.5% of active SIP accounts, 39% of direct-plan SIP inflows), new demat account share (25.8%), and 3-year customer retention rate (77%).
- [2]AMFI (Association of Mutual Funds in India). Monthly SIP Data. amfiindia.com/research-information/amfi-monthly. Source for SIP stoppage ratio data, monthly SIP volumes, and SIP folio counts. January 2025 monthly note (AMFIMonthlyNote_January2025.pdf) confirms stoppage ratio reached 109% - first time in AMFI history that monthly SIP cancellations exceeded new registrations.
- [3]Business Standard (December 2022). "SIP closure ratio surges to 51% in 2022 from 41% in 2021." business-standard.com. AMFI data reported by Business Standard; used for FY21–22 stoppage ratio benchmarks confirming the correction-cancellation correlation.
- [4]Business Standard (March 2025). "SIP account openings surge, but premature closures also on the rise." business-standard.com. AMFI data showing SIP cancellation trends through the October 2024–March 2025 correction window.
- [5]Business Standard (February 2025). "Don't lose rupee-cost averaging by stopping SIPs; stay focused on goals." business-standard.com. Context on behavioral drivers of SIP cancellations during the 2025 correction; quotes fund managers on RCA math.
- [6]SEBI (September 2024). "Updated SEBI Study Reveals 93% of Individual Traders Incurred Losses in Equity F&O Between FY22 and FY24; Aggregate Losses Exceed ₹1.8 Lakh Crores Over Three Years." sebi.gov.in/media-and-notifications/press-releases/sep-2024. Demonstrates scale of behavioral loss in Indian retail investing; context for anxiety-driven SIP exits.
- [7]Groww IPO. NSE/BSE Listing, November 2025. IPO details: ₹6,632 Cr raised, price band ₹95–100 per share (DRHP), listed at 29% premium; market cap ~$9B at debut (derived: issue price × total shares post-IPO). TechCrunch (November 2025) cited for news framing; financial figures sourced from DRHP and exchange listing data.
- [8]Kahneman, D., & Tversky, A. (1979). "Prospect Theory: An Analysis of Decision under Risk." Econometrica, 47(2), 263–292. Foundational behavioral research establishing that losses feel 2–2.5× more painful than equivalent gains feel pleasurable - the mechanism behind panic SIP cancellations during corrections.
- [9]Thaler, R. H., & Benartzi, S. (2004). "Save More Tomorrow: Using Behavioral Economics to Increase Employee Saving." Journal of Political Economy, 112(S1), S164–S187. Nudge-based framework for savings persistence; establishes that contextual information and opt-out defaults materially improve long-term savings adherence.