Quick Commerce: from revenue growth to store-level profitability
A leading Indian quick-commerce operator is growing fast, yet 22 of its 60 dark stores lose money. This study finds out why, and what to do about it.
60 dark stores · 6 cities12 months · 720 store-month recordsRun-rate quarter Jul–Sep 202610 min read
01 · Summary
Executive summary
The situation. A leading Indian quick-commerce operator was growing fast, yet 22 of its 60 dark stores lost money and the network earned only 1.8% store EBITDA on net order value. The average was improving; the dispersion was not.
The diagnosis. The profit gap is a volume-density problem. Bottom-quartile stores run 430 orders a day against 1,313, spreading ₹104 of fixed cost per order against ₹38. Overlapping catchments, discount leakage, fixed costs sized ahead of demand, and wastage and delivery cost that rise as volume falls compound the gap.
The recommendation. Segment, don’t cut. Forty-two stores to scale or optimise carry 86% of net order value; eighteen need a turn-around, resizing or exit decision. In the Base case the programme adds ₹14.9 crore a year to store EBITDA (1.8% to 3.0%) and pays back in month 7, counting no ramp-up growth. These are forecast outcomes, to be validated through a 90-day pilot.
02 · Context
Business context
The CEO’s brief was direct: growth is accelerating, but dark-store profitability is inconsistent; identify the underlying causes and develop a financially viable transformation plan.
The perimeter is 60 dark stores across Delhi NCR, Mumbai MMR, Bengaluru, Hyderabad, Pune and Jaipur. Over twelve months they sold ₹1,103 crore of net order value, with 56,195 orders a day at an average basket of ₹649 in the latest quarter. Store EBITDA excludes central overheads. Across the industry, leading operators are moving their emphasis from adding stores to improving what each store earns.
₹8.5 cr12-month store EBITDA, 0.8% of net order value
22 of 60stores with negative EBITDA in the run-rate quarter
1.8%store EBITDA as a share of net order value, run-rate quarter
937orders per store per day, run-rate quarter
₹649average order value
03 · Challenge
The average is improving. The dispersion is the problem.
Network store EBITDA moved from a ₹74 lakh loss in October to ₹210 lakh in September, and the count of loss-making stores fell from 35 to 19. Yet 22 stores still lost money in the latest quarter. A network average hides exactly the stores that need a decision.
Store EBITDA improves through the year, but loss-making stores persist
Network store EBITDA, ₹ lakh per month (bars) and number of loss-making stores (line), Oct 2025 – Sep 2026
04 · Investigation
Profit follows volume: where the gap comes from
The 720 store-month records were reconciled to the network P&L through twenty checks before any diagnosis began. Five dimensions were examined: revenue and demand, customer and basket, store costs, inventory and fulfilment, and network and catchment. The pattern is consistent: profit follows order density, and the weakest stores carry ₹104 of fixed cost per order against ₹38 in the strongest.
Each dot is one store: EBITDA rises with orders per day
Store EBITDA as % of net order value against orders per day, Jul–Sep 2026
Bottom-quartile stores cost ₹88 more per order
Gap by cost line, top versus bottom quartile, ₹ per order
Small baskets barely cover variable cost
Contribution ₹ per order by basket size; average fixed cost per order ≈ ₹49
05 · Findings
Five findings, and what to do about each
FINDING 1
Volume is below breakeven in a minority of stores
What the numbers reveal
22 of 60 stores lose a combined ₹129 lakh a month, and 14 of them are under 12 months old. Top-quartile stores average 1,313 orders a day; the bottom quartile, 430. Fixed cost is spread at ₹104 an order against ₹38.
What it means
Rent and rosters were sized ahead of demand in immature or low-potential catchments. The issue is concentrated, so it needs a store-level answer.
What management should do
Segment store by store. No blanket cuts.
FINDING 2
Overlapping catchments cap volume
What the numbers reveal
Mature stores with another own store within 1.5 km sell 22.0 orders per thousand households, against 30.5 for those without (28% lower; 11 stores versus 35). Because only mature stores are compared, age does not explain it.
What it means
Openings inside an existing store’s radius split the same households. About ₹32 lakh a month of loss sits in overlapped stores.
What management should do
Stop in-radius openings. Where two stores overlap, close only the weaker one, and only if demand stays covered.
FINDING 3
Discounts leak on new and small-basket orders
What the numbers reveal
New-customer discount runs at 13.3% of gross order value against 3.5% for loyal customers. Orders under ₹250 are 15% of volume and earn ₹2 of contribution each, against ₹106 for baskets of ₹650–1,000.
What it means
Blanket offers with no basket threshold subsidise the least valuable orders. Closing the gap to the peer median is worth about ₹53 lakh a month after allowing for lost orders.
What management should do
Move to targeted offers and basket thresholds, and protect repeat customers.
FINDING 4
Fixed costs are mismatched to demand
What the numbers reveal
The network carries 139 FTE more than peer productivity of 38.2 orders per FTE per day would need, and 15,655 sq ft of excess space on leases that can be acted on within 12 months.
What it means
Rosters were fixed to design volume, and some leases are oversized or above market. The labour and rent gap is worth about ₹66 lakh a month.
What management should do
Roster to the demand curve; resize and renegotiate space where lock-ins allow.
FINDING 5
Wastage and delivery cost rise as volume falls
What the numbers reveal
Wastage is 3.4% of net order value in low-volume stores against 2.4% in high-volume ones. Delivery cost has a median of ₹38 an order; the bottom quartile pays ₹44 against ₹36 for the top.
What it means
Fixed assortment depth and thin rider batching make small stores structurally costlier. The gap is worth about ₹79 lakh a month.
What management should do
Introduce assortment tiers by store size, rider batching and delivery-radius rules.
Findings overlap and are not additive. The business case counts each lever once per store, with an interaction haircut. EBITDA correlates strongly with orders per day and orders per FTE, but young stores are both small and discount-heavy, so the overlap finding is tested on mature stores only.
06 · Financials
What the programme is worth, and how much of it you can bank
In the Base case the programme adds ₹14.9 crore a year to store EBITDA, lifting the margin from 1.8% to 3.0%, and pays back in month 7. Revenue falls slightly, by about ₹101 lakh a month, because discount cuts and two closures lower volume: profit rises while revenue does not. The simulator in section 09 runs the same store-by-store logic as the underlying financial model, so every lever can be tested for yourself.
07 · Actions
Segment, don’t cut
Forty-two stores to scale or optimise carry 86% of net order value. Eighteen need a turn-around, resizing or exit decision. Closure is the last resort: it is allowed only where the loss is structural, a neighbouring store can absorb the demand, and the exit pays back within the gate.
Class
Stores
% of NOV
EBITDA ₹L/mo
EBITDA %
Orders/day
Decision
Scale
14
29%
124
4.1%
1,212
Grow demand, protect availability; no cost-out
Optimise
28
57%
183
3.1%
1,108
Close discount, wastage, roster and delivery gaps
Turn around
7
6%
−30
−4.7%
495
90-day stage-gated plan; mostly ramping stores
Relocate or resize
8
5%
−70
−13.7%
345
Resize or renegotiate space; relocation study where catchment is weak
Close
3
4%
−15
−4.0%
658
Exit only if the loss is structural, a neighbour can absorb demand and payback is within the gate
Total
60
100%
192
1.8%
937
Five interventions, Base case
Intervention
Benefit ₹ cr/yr
One-time ₹ L
Owner · timing
Key risk → mitigation
1. Targeted discount and basket economics
4.10
7
Chief Commercial Officer · M2–M4
Order loss above plan → stop-loss trigger; keep loyalty offers
Service slips → on-time floor with automatic rollback
4. Rent and footprint: resize and renegotiate
2.06
70
Head of Real Estate / CFO · M5–M10
Lock-ins → only leases with ≤ 12-month lock-in counted
5. Network rationalisation, only where customers are covered
2.86
187
CEO / COO with CFO sign-off · M4–M8
Demand transfer below plan → stage-gate and reversal option
Total before programme cost
15.57
296
Programme office ₹13 L one-time and ₹5.8 L a month recurring; net benefit ₹14.9 cr a year
Considered, not prioritised. Blanket cost cutting (gaps are store-specific and would hurt the 14 Scale stores). Closing every loss-making store (most are ramping or fixable). Price or fee increases (no elasticity evidence yet). More ad income (it already scales with volume). New openings and central overheads (outside the mandate and the store dataset).
08 · Roadmap
Pilot in 90 days, then scale what is proven
Days 1–30
Mobilise
Stand up the PMO and weekly store P&L close, freeze KPI definitions, specify the Power BI/DAX cockpit, design discount rules, assortment tiers and roster templates, and open lease talks. Gate: baseline signed off by the CFO.
Days 31–60
Pilot
Run each lever in the highest-gap stores with the next-highest as controls. Test demand transfer for closure candidates without closing anything. Gate: weekly scorecard against controls.
Days 61–90
Validate and decide
Compare pilots with thresholds, re-run the model with measured elasticity and transfer rates, and decide go, adjust or stop per lever. Closures proceed only if all three tests pass. Gate: steering committee.
12-month rollout by lever
Ramp from start month to full run-rate; responds to the implementation-speed control in the simulator
KPI
Baseline
Target / floor
Owner
Escalate when
Store EBITDA, 60 stores (₹ L/month)
192
316
CFO
Below 80% of ramped target for 2 months
Loss-making stores
22
17
COO
No reduction after the 90-day review
Network orders per day (guardrail floor)
56,195
55,071
COO
Below floor for 2 weeks
On-time delivery
86.2%
≥ 85.2%
COO
Breach for 2 consecutive weeks pauses the lever
Repeat order share
78.2%
≥ 76.2%
Chief Commercial Officer
Down 2 points or more
Rollback rule: a breach for two consecutive weeks in a pilot store pauses the lever and restores the previous setting. A benefit counts as realised only when Finance reconciles it to store P&L against control stores.
09 · Simulator
Run the numbers yourself
Adjust the levers, closure decisions and timing. Every output below recalculates from the same store-level logic as the supporting workbook.
10 · Supporting analysis
Supporting analysis
Explore the supporting financial analysis and underlying store-level model.
Methodology & Disclosure: This case study is an independent analytical exercise developed using a constructed business dataset. The analysis, financial scenarios and recommendations are illustrative and are not representations of a commissioned client engagement or verified business outcomes.
