Appearance
Go to market, positioning, roadmap and risks
Part of
car_sales— the dealership operating layer. Evidence-class legend there.
1 · Pune first, and the channel that does not exist
🌐 Pune is a genuinely good choice and the registration data says so plainly: Pune RTO booked 3,04,368 new registrations in FY25 — more than all four Greater Mumbai RTOs combined (2,94,940) — and Pune plus PCMC is 5,00,650, or 17.3% of Maharashtra's 28,90,497.
A software vendor cannot join FADA
🌐 Membership requires "Proof of Dealership - LOI/Dealership Agreement / OEM Certificate", and there are exactly two dealer-only membership categories. There is no associate, vendor or technology-partner tier. And FADA's regional Vyapar events run in Delhi, Rajasthan, Gujarat, Kerala, Tamil Nadu and Bengal — none in Maharashtra.
🌐 The one open national surface is the Finance & Insurance Summit, which is a travel-and-sponsorship expense rather than a membership. ❓ Note that a large used-car marketplace appeared there as a partner, not a sponsor, so this is weak evidence that a neutral software vendor can buy a stage.
🔎 If an association channel was in the plan, it is not there. What remains: founder-led selling into a named list, then group-to-group referral — which in a city where twelve groups hold sixty Maruti outlets is a small, well-connected community where one visible success travels fast.
The Pune denominator, with its disagreements intact
| Unit | Count | Note |
|---|---|---|
| Dealer entities of every kind, Pune metro | 🌐 ~215-403 | Two independent methods, wide band |
| Franchised PV dealers | 🌐 45-70 | The ceiling on any franchise-only plan |
| Outlets, bottom-up | 🌐 ~639 | 224 PV + 244 2W + 171 used-car. ❓ The 2W component varies 206-244 across sources |
| Maruti outlets / groups | 🌐 60 / 12 | 5.0 outlets per group — the ratio that justifies selling to groups, and it survived fact-checking |
| Skoda · Mahindra | 🌐 9 outlets / 3 entities · 6 / 3 | Consolidation is normal here, not exceptional |
| Maharashtra dealerships | 🌐 ~1,520-1,586 | 10.1-10.6% of national, corrected for Vahan-4's exclusion of Telangana |
🌐 Named first-call list — multi-brand groups verified to exist in Pune: Sai Service (Maruti + Bajaj
- KTM, 72 workshops across six states), B.U. Bhandari (VW + MG + Triumph), plus three Skoda entities and three Mahindra entities. ❓ Per-group outlet counts from aggregators disagree; verify each by phone rather than trusting a locator.
Expansion sequence
- Pune + PCMC. Prove the wedge and get one measured reference.
- Maharashtra — Mumbai, Nashik, Nagpur, Aurangabad. ⚠ Nagpur is Excellon's registered office.
- Gujarat or Tamil Nadu next, for a channel reason rather than a market one. 🔎 Both have FADA Vyapar events that Maharashtra lacks, so there is an association surface to stand on.
- National only behind Path B or Path C, never as a remote self-serve motion.
2 · Positioning
🔎 The open position is dealer-chosen, multi-brand, group-level operations. Nobody occupies it, and the emptiness is structural rather than accidental: an OEM cannot build a cross-brand product, and a vendor whose customer is the OEM cannot sell one without a conflict.
| Quadrant | Occupant | Note |
|---|---|---|
| OEM-chosen, single-brand | 🌐 Maruti, Hyundai GDMS, Tata Siebel, Mahindra DMS 2.0 | Dead scope |
| OEM-chosen, network platform | 🌐 Zoho SKYLine, DigiForce, Salesforce Cerebro | Path C lives here |
| Dealer-chosen, single-brand back office | 🌐 INFOMAN, Orbitsys, AutoBooom, eMsys | ₹29 k-₹3 L, crowded |
| Dealer-chosen, per-seat CRM | 🌐 AUTOSherpa, LeadSquared, Zoho | Breaks at 10-40 users |
| Multi-brand but tiny ACV | 🌐 GaragePlug, Autorox, RAMP | ₹175-3,499/month |
| Dealer-chosen, multi-brand, group-level | Empty | The position |
The sentence to test in E1:
"Your manufacturer's system reports to the manufacturer. This one reports to you, across every brand you carry, and your numbers never leave your group."
⚠ 🔎 Note that Path C sits in the opposite corner of the same map, which is why the two cannot both be pursued. See commercial-model.
3 · Roadmap, sequenced by behaviour change required
⚠ REVISED 2026-08-22 - this page previously carried a SECOND, conflicting roadmap
An earlier version led Wave 1 with a "Retention wedge" (service annuity) while product map and operating model led with the attribution spine. Two roadmaps coexisted in one section. They are reconciled here, and the reconciliation is argued in value proposition §6: they were never actually in conflict, because both depend on parties.
The resolution: attribution is the PRODUCT THESIS; zero-behaviour-change capabilities are the ADOPTION PATH. Build the spine, lead the sale with the surface.
| Wave | When | Contents | Why here | Becomes sellable | Target |
|---|---|---|---|---|---|
| Validation | Weeks 1-6 | E1, E2, E3. Zero engineering | The price has never been tested | Hand-delivered pilot at Rs 25,000/3 months | A measured price and a read franchise agreement |
| Wave 1 | Nov-Dec 2026 | Fix the card-view beacon (QRS-734) · parties · visitor register · card attribution + auto-lead | 🔎 The beacon makes the product's first claim true. parties unblocks four registered features. The visitor register needs zero behaviour change from anyone and is the adoption wedge | Foundation, Rs 75,000/yr | 3-6 paying outlets |
| Wave 2 | Jan-Apr 2027 | Service and renewal reminders (the asset table plus the existing recurrence engine) · leads · targets · follow-up SLA · RBAC, its own wave and the largest single dependency | 🔎 Reminders are the first capability that produces revenue with no staff behaviour change. RBAC gates everything above Team Leader | Growth, Rs 1,50,000/yr | 8-15 logos cumulative |
| Wave 3 | May-Aug 2027 | organizations write path · seats · billing · Group Intelligence · org-admin surface. ⚠ Needs an ADR-0011 amendment and a design round, neither of which exists | The moat and the premium SKU. Longest path to a customer | Complete Rs 3,00,000/yr · Group Intelligence Rs 80,000/outlet | First 2-4 groups |
Why Wave 1 contains what it does
🔎 Three of its four items need none of RBAC, seats, billing or the org-admin portal - which is what makes them reachable inside 2026. And the visitor register is chosen over the employee card deliberately: reception must log walk-ins anyway, so we only have to be faster than a pen, whereas the card asks every rep to change a habit permanently against 🌐 29.53% annual attrition.
⚠ The employee card still ships in Wave 1. It is the product thesis and the moat's foundation. What changed is that it is no longer what the sale leads with.
4 · Build versus integrate versus refuse
| Capability | Decision | Why |
|---|---|---|
| Tenancy tree, oversight, sharing | 🧮 Built already | Proven by pgTAP. Do not touch it |
| Vehicle assets, service recurrence | Build first | Recurrence engine exists; asset is ~0.5 day of schema |
parties, leads, targets, RBAC | Build — Wave 2 | 📘 Four registered features declare customers as parent. Build the primitive once (QRS-823) |
| Integrate direct | 📘 Meta Cloud API per ADR-0029. Dealer-owned account via Embedded Signup. No reseller | |
| Telephony | Integrate | 🌐 Exotel / MyOperator / Acefone. Never own a dialler |
| Reviews / Google Business Profile | Integrate | ⚠ 🌐 The GBP API needs application and approval; an unapproved project gets 0 queries per minute. Apply early |
| Payments | Integrate | Razorpay is already in the stack |
| DMS ingestion | CSV / batch extract only | 🌐 No Indian OEM publishes an API. Design for a dealer-mediated export and never promise live sync |
| DMS, accounting, parts, lending, insurance | Refuse | product-scope |
Architecture consequences worth deciding now
- Price on outlets, license on seats internally. 📘 ADR-0023 D2 says seats license users; 🌐 the market says per-user pricing is unsellable. Both survive: keep seats as the internal mechanism (accountability requires individual logins, so it is self-enforcing) and sell a flat per-outlet bundle with a generous ceiling using
feature_grants.limit_valueandon_exceed, 🧮 which already exist. - Never build an OEM-facing view of discount or margin data. A legal safeguard and the strongest trust position available.
- 🌐 QRSETU is the DPDP processor; the dealer is the fiduciary. The Rules make one contract term mandatory: the processor must be obliged to implement reasonable security safeguards. Put it in the MSA template before the first customer. Substantive duties commence around mid-May 2027, i.e. inside this window; penalties reach ₹250 crore for safeguard failure, ₹200 crore for breach non-notification. First breach intimation is "without delay" with detail within 72 hours; data-principal requests answered in 90 days.
- 🌐 Retention is a ceiling, not a floor. The guidance is to retain logs and personal data for one year only unless required by law. Design deletion, not accumulation.
- ❓ Does pooled cross-dealer benchmarking make QRSETU a Data Fiduciary rather than a processor? This bears directly on the one genuinely defensible product and needs counsel, not more research.QRS-837.
5 · Risks and assumptions
| # | Risk | Severity | Cheapest test or mitigation |
|---|---|---|---|
| R1 | ❓ OEM dealer agreements may forbid exporting DMS data to a third party. Six of eleven research tracks named this their top unknown; none could answer it | Existential | E1 — read five agreements. QRS-835 |
| R2 | OEM displeasure. 🌐 The CCI record shows manufacturers policing dealer conduct with fines and supply threats | High | Never expose dealer data to an OEM |
| R3 | 🧮 Build consumes the selling window. Five of seven primitives, RBAC from zero, seats, billing, and an org-admin surface in a stack with no auth | High | Wave 1 deliberately needs none of it |
| R4 | 🌐 Excellon: 466 people, ₹74 Cr, 26 years, bootstrapped, 8 km away | High | The OEM-conflict position is the only defence that does not depend on speed |
| R5 | 🌐 The buyer is loss-making — PBT minus ₹13.30 Cr at a ₹6,381 Cr group | Medium | Price against the advertising line; expect slow payment |
| R6 | 🌐 Cannot out-price free in used-car: Cars24, CarTrade and Mahindra First Choice give dealer tooling away | Medium | Do not lead with used-car |
| R7 | ❓ Support minutes per customer unmeasured. 📘 For two people this is the binding constraint, not sales capacity (QRS-748) | Medium | E3 measures it directly |
| R8 | WhatsApp repricing, and possible TRAI OTT regulation | Medium | Pass-through pricing; never bundle a fixed message allowance |
| R9 | Two people, no sales function, no association channel, no reference customer | High | Hire from Q1 2027 or accept Path B's timeline |
| R10 | ❓ Churn completely unmeasured. At 30%, netting 200 needs ~260 gross adds | Medium | Annual prepay; onboarding fee as commitment |
| R11 | 🌐 Platform dependency is resolved politically in India, not contractually — Google Play delisted several large Indian apps in March 2024, reinstated after ministerial intervention | Low | Do not build a single-channel product |
6 · The three experiments that should happen before any code
🔎 Total cost roughly six weeks, about 12% of remaining runway. The alternative is spending 100% of it building an org-admin portal against a price nobody has tested.
E1 · Twelve dealer-principal interviews — two weeks
Converts R1 and five missing-research items at once. Ask for:
- The IT and data-ownership clauses of the actual franchise agreement.
- Every software line item with its rupee amount, and who invoices it — the OEM or the dealer.
- Who signs, at what approval threshold.
- Monthly spend with CarDekho, CarWale, Cars24.
- "If I handed you cross-outlet advisor performance tomorrow, what would you pay for it?"
Record verbatim. Do not soften a no. 🔎 Pune is the one city where this is physically possible for two people in a fortnight.
Gate: build nothing until 8 of 12 are done.
E2 · The Pune dealer census — one week, one spreadsheet
Scrape 12-15 OEM dealer locators for Pune and PCMC, cluster on legal entity name, phone each for its current DMS or CRM. Output: entity, brands, outlet count, service centres, incumbent software.
🔎 This replaces three mutually inconsistent inferred counts with the measured denominator every calculation divides by, and it produces E1's call list. No permission needed, no engineering.
Near-zero marginal add: while on each dealer's page, log the Google review count and last-review date. That settles the reputation-product question in an afternoon.
E3 · A ₹25,000 paid pilot with five dealers — four weeks
Not a survey. Stated pricing intent is worthless and this study is already full of unvalidated list prices. Offer five Pune dealers the single best-evidenced pitch — cross-outlet advisor accountability plus service-due reminders, both of which the OEM spine demonstrably does not do — at ₹25,000 for a three-month pilot, invoiced and collected. Deliver it by hand, with spreadsheets and a WhatsApp number.
🔎 Zero engineering is the point: it measures willingness to pay for the outcome before a line of code commits the roadmap. Three signals: does anyone pay at all; the annualised rate they accepted; and support minutes per dealer per week, which is the actual binding constraint on two people.
Pre-committed decision rule
If 0 of 5 pay ₹25,000 for three months, ₹1 lakh × 200 is dead — and you learned it for the cost of one month instead of twelve.
⚠ This experiment was already testing the CORRECTED price while the price book listed ₹30,000
🧮 ₹25,000 for three months annualises to ₹1,00,000/yr — above the ₹79,999 Foundation floor and below the ₹1,49,999 Growth tier of the corrected book, and 3.3x the ₹30,000 list price this section published until 2026-08-23. 🔎 The company's own validation instrument and its own price list disagreed by a factor of three, and neither page cited the other. QRS-844.
🔎 And note E3's third signal — support minutes per dealer per week — is the measurement that cost to serve had to estimate at ₹1,500/h because nobody has it yet. It is the single most valuable number this pilot can produce, because it sets the floor under every tier.
What all three deliberately avoid
🔎 None requires the org-admin portal, seats, billing, RBAC or any of the five missing primitives. That sequencing is the whole point: 🧮 the repo's measured state says that build consumes most of the remaining twelve months, and on current evidence it would be committed before a single price had ever been measured.
7 · What to do first
- Do not start the org-admin portal. It is the most expensive item on the roadmap and it is gated on a price nobody has tested.
- Run E2 this week. One spreadsheet. It produces both the denominator and the call list.
- Book E1. Get a franchise agreement's data clause in hand. Nothing else here matters more.
- Sell E3 by hand for ₹25,000 before writing a line of dealership code.
- 📘 Finish Ganesh Chaturthi first. It is weeks away, twelve vendors are named, and it is already paid for. Fork the roadmap in October. 🔎 Note that three of the four things that season needs — measurement, WhatsApp transport, a collection path — are shared with this plan, so the fork is cheaper than it looks.
- Decide Path B or Path C deliberately. Dealer-owned and OEM-sold are mutually exclusive positions.
Related
- Commercial model — the pricing and the ₹2 Cr arithmetic
- Discovery brief — what E1 must produce for the G-D gate
- 📘 Strategy: 12-month priorities — where this sits against the platform-level plan