Custom wholesale CRM for dealer accounts, credit visibility, and WhatsApp-driven reorders

Project overview

The client — a wholesale distribution business we refer to here only as a confidential regional distributor — moves fast-moving consumer goods and industrial consumables through a network of dealers, sub-dealers, and occasional direct B2B accounts. Sales happen on phone calls, WhatsApp voice notes, and occasional in-person visits. Orders land in a desktop accounting package that finance trusts for invoicing and GST, but that system was never designed to be the daily workspace for twelve field representatives and four inside-sales staff who need to know credit headroom, last order mix, and whether a dealer already asked for a reorder yesterday on WhatsApp.

Before this engagement, “CRM” meant a shared Google Sheet, Word quotation PDFs, and WhatsApp groups where reorder requests disappeared into scrollback. Credit limits lived in finance’s spreadsheet; reps discovered holds only when a dealer called angry.

Web6 was engaged from Surat to build a purpose-fit wholesale CRM — not a generic SaaS seat license, not a full ERP replacement, but an operational layer that sales and account management could live in while finance kept the ledger system they already audited. The stack was Laravel for API and business rules, React for the staff-facing UI, PostgreSQL for relational integrity around accounts and orders, Redis for session and queue work, Docker for reproducible environments, Cloudflare for edge protection and TLS, and AWS for production hosting. We exposed a documented REST API so a later ERP integration — discussed openly with leadership and scoped on our ERP development service page — could sync master data and posted invoices without re-platforming the CRM.

The project ran fifteen weeks with a team of five. We scoped dealer-centric B2B CRM with credit and order context at the point of sale — not ecommerce, WMS, and ERP cutover in one release — so the client could go live before peak festival season. Outcomes were measured in operations, not fabricated revenue percentages: faster quote-to-order cycles, fewer credit disputes, and ninety-four percent field rep adoption within eight weeks.

Engagements like this are delivered through Web6 wholesale CRM development .

Business and technical challenges

Dealer identity was fragmented. The same shop might appear under slightly different names in WhatsApp contacts, accounting exports, and the owner’s visiting card. Some sub-dealers ordered through a parent dealer’s account; others had their own GSTIN but shared delivery addresses. Without a canonical dealer profile linking contacts, territories, credit tier, and order history, every conversation started with twenty minutes of “let me check and call you back.”

Credit visibility was a trust problem, not just a data problem. Finance maintained credit limits and tracked credit notes in the accounting package, but sales had no read-only view that updated within a business day. Reps routinely promised SKUs on net terms that finance would later reject, damaging relationships that took years to build. Leadership wanted credit limits and open credit note balances displayed prominently on dealer accounts — read-only for reps, editable only by finance roles — without giving sales a back door to change limits.

WhatsApp was the real order channel, but it left no auditable trail inside any system. A dealer would voice-note “send the same as last month plus two cartons of the blue pack”; the rep would interpret, call back office, and someone would type a sales order from memory. When disputes arose — wrong SKU, wrong quantity, “I never asked for that” — there was no linked record of what was said and when. The client asked explicitly for WhatsApp reorder conversations to be loggable against dealer accounts: not full WhatsApp Business API automation in phase one, but structured capture of forwarded chats, notes, and attachments reps already received on their phones.

Order history needed beside CRM context, not in a separate reporting tool. We designed REST import endpoints and nightly sync jobs, plus manual refresh, so screens stayed useful before full ERP realtime sync.

• ✓ Canonical dealer and sub-dealer account model with territories
• ✓ Read-only credit limits and credit note visibility for sales
• ✓ Order history adjacent to pipeline and quotation workflows
• ✓ WhatsApp reorder context logged per dealer account
• ✓ Role-based access for reps, inside sales, finance, and managers
• ✓ ERP-ready REST hooks without mandatory phase-one integration

Goals and success criteria

Product goals centered on a dealer account page that answered the questions reps actually ask in the field: Who is this dealer? What is their credit headroom? Are there open credit notes? What did they last order? Is there an active quotation? Has anyone logged a WhatsApp reorder request this week? Secondary flows included creating tiered quotations with SKU search from the client’s product catalog, moving deals through a simple pipeline (lead, quoted, negotiation, won, lost), and converting won quotations into order requests that finance could pick up in the accounting system.

Technical goals required a Laravel API with explicit versioning, Sanctum-based authentication for staff users, PostgreSQL schemas normalized enough for reporting but pragmatic enough for fast iteration, Redis-backed queues for order import jobs and email notifications, and a React SPA with role-aware navigation. All sensitive mutations — credit limit changes, quotation approvals above threshold, user role edits — would audit to an append-only activity log with actor, timestamp, and before/after snapshots.

Integration goals were phased: REST endpoints and webhooks for ERP — dealer export, quotation events, order requests — with docs linked to our ERP development service. Nightly order imports and manual refresh were honest phase-one boundaries.

We set explicit non-goals: no dealer-facing ecommerce portal in phase one (though we referenced B2B ecommerce development for a future self-serve catalog), no warehouse barcode scanning, no payroll, no marketing automation beyond basic quotation follow-up reminders. WhatsApp automation via official Business API was flagged as phase two if message volume justified template costs.

Discovery and research

Discovery opened in the client’s Surat office with two half-day workshops. We mapped the current quote-to-cash story on a whiteboard: dealer call, informal price check, Word quotation, WhatsApp confirmation, back-office order entry, accounting invoice, dispatch, payment follow-up. We timed three real quotations reconstructed from email threads — median elapsed time from first dealer ask to order keyed was four and a half business days, with two handoffs where information was retyped.

We interviewed four field reps on their phones — not to surveil WhatsApp, but to understand how they search for “what did Bhai order last Diwali?” today. Every rep had a personal Notes app snippet or photo of a ledger page. Inside sales showed us their pipeline spreadsheet columns: dealer name, estimated value, “hot / warm / cold,” and a free-text cell that meant different things to different people. Finance walked us through credit note issuance and how limits were recomputed monthly; we captured the exact fields that had to appear read-only on CRM dealer headers.

Technical discovery included anonymized exports from the accounting package: dealer master, last twelve months order headers and lines, open credit notes. We profiled row quality — duplicate GSTINs, trailing spaces in city names, legacy dealer codes no one used anymore. Data cleansing rules became a shared spreadsheet reviewed with finance before import scripts ran on staging.

We reviewed three off-the-shelf CRMs the client trialled — none showed credit notes beside order history or parent–sub-dealer hierarchies without painful workarounds. Custom build won once WhatsApp logging and credit visibility were non-negotiable.

System architecture

The system follows a classic modular monolith API with a separate React frontend — appropriate for team size and integration surface, without microservice overhead the client could not operate. Laravel 11 serves versioned JSON under /api/v1, enforces Sanctum token authentication, and organizes domain logic into modules: Accounts (dealers, contacts, territories), Credit (limits, notes, read-only snapshots), Catalog (SKU search synced from master import), Quotations, Pipeline, Orders (imported history and order requests), Communications (WhatsApp log entries), Reports, and Admin (users, roles, audit).

PostgreSQL holds dealers, credit limits and notes, quotations, pipeline stages, imported orders keyed by accounting IDs, communication logs with attachments on AWS S3, and append-only audit events. Redis backs sessions, catalog cache, and Horizon queues for imports, PDF generation, and approval notifications. The React SPA uses Vite and TanStack Query with dense B2B tables; Laravel policies remain the authorization source of truth.

Docker Compose defines local and staging parity. Production runs on AWS EC2 behind a load balancer; Cloudflare terminates TLS, WAF-protects login routes, and caches static assets. Nightly PostgreSQL snapshots with a pre-launch restore test.

REST integration layer exposes documented endpoints and signed webhooks for ERP phase two: GET /dealers delta since timestamp, POST /order-requests, webhook quotation.won with payload schema. Internal runbooks reference our ERP development approach so the client’s future integrator — Web6 or otherwise — inherits OpenAPI specs and sandbox credentials.

• ✓ Laravel modular monolith — API versioned at /api/v1
• ✓ React SPA with role-aware routes and dense B2B tables
• ✓ PostgreSQL for dealers, credit, quotations, orders, audit
• ✓ Redis queues via Horizon for imports and notifications
• ✓ AWS hosting with Cloudflare edge TLS and WAF
• ✓ Documented REST hooks for optional ERP sync

User experience for sales, finance, and management

UX priority was the dealer account detail page — the screen reps open fifty times a day. Header shows legal and trade name, territory badge, assigned rep, GSTIN, primary phone, and credit summary: limit, utilized, available headroom, count of open credit notes with expandable list (number, date, amount, reason code). Below, tabbed panels: Order history (last ten with “view all”), Active quotations, Pipeline status, WhatsApp and call log (manual entries plus pasted transcript notes), and Contacts. Finance users see an additional Credit tab with edit controls and audit history; reps see read-only.

Quotation builder: search SKU, pick tier price, adjust quantity, auto-sum GST. PDF export uses client logo and terms. Pipeline kanban for managers; won deals create order requests finance picks up in accounting.

WhatsApp logging UX respects field reality. Reps tap “Log reorder chat,” select dealer, paste text or upload screenshot, tag intent (reorder, price check, complaint, other), and save. Full WhatsApp Business API sync was out of scope; the client wanted accountability without forcing reps into a new chat app. Managers filter communication logs by rep and date during dispute reviews.

Reporting stays operational — quotation funnel, cycle time by rep, credit holds — with CSV export for finance. Core flows work on mid-range Android over 4G with skeleton loading on order history panels.

Delivery process

Weeks 1–2: discovery and wireframes. Weeks 3–5: dealer import, credit screens, catalog. Weeks 6–8: quotations, pipeline, approvals. Weeks 9–10: order imports, WhatsApp logs. Weeks 11–12: reports, REST docs, mobile polish. Weeks 13–15: UAT, finance parallel run, cutover, hypercare. Two-week sprints with Friday staging demos; finance signed off credit display before reps saw production data.

Discovery deliverables included a dealer account wireframe printed A3 for field reps to mark up — more useful than Figma alone because reps wrote " where is last Diwali order " in margins. Those annotations became acceptance tests for the order history panel placement above pipeline, not below contacts as originally designed.

Data migration ran in three waves: legacy accounting dealer master week three, twelve-month order history week nine, open credit notes week eleven. Each wave had finance reconciliation gate — zero tolerance on credit mismatch, one percent row quarantine tolerance on historical orders with known bad GSTIN rows. Rep training started only after wave two so practice quotations referenced real SKU mix dealers recognized.

UAT split by role: reps walked fifty dealer updates and twenty quotations; finance replayed ten historical credit disputes using only CRM exports; managers tested pipeline rollups against spreadsheet totals. Parallel run week fourteen allowed reps to CRM-first while finance still accepted phone order requests — order-request volume through CRM crossed eighty percent by day five of parallel run without mandating cutover.

SEO and public web presence

This engagement was internal CRM software, not a marketing site rebuild. We coordinated with the client’s existing public website — built years earlier on a separate stack — so staff bookmarked the CRM subdomain correctly and robots.txt blocked indexing of authenticated app routes. No SEO deliverable was in scope; our custom website development team documented a future phase if they want dealer-facing catalog pages tied to CRM catalog sync.

For Web6’s own content, this case study documents the CRM practice alongside our custom CRM development service page. Keywords such as wholesale CRM and distributor CRM reflect how the client searched for help — operational software, not consumer Google traffic.

Performance and reliability

Staff-facing targets: dealer search p95 under 300ms, quotation save under 500ms, paginated order history for large accounts. Nightly imports run in background queues; failed rows quarantine for finance reconciliation. Load tested at forty concurrent users; mobile verified on Jio and Airtel 4G.

Catalog search uses PostgreSQL trigram indexes on SKU code and description prefix — reps type partial codes from memory faster than scrolling category trees. Quotation PDF generation runs async with polling UI; reps continue editing lines while PDF job queues — eliminating the frozen-screen complaints from UAT week one when PDF was synchronous.

Import jobs idempotent by accounting invoice reference — rerunning a failed nightly batch does not duplicate order rows. Admin dashboard shows last successful import timestamp per data source so reps trust the " as of " label on order history when finance delays export by a day during month-end close.

Security and compliance

Sanctum auth with role policies on every mutation. Credit edits audit append-only. Attachments on private S3 with signed URLs. Cloudflare WAF on login; secrets in AWS Parameter Store. NDA anonymization applies to this case study.

Dealer PII and credit data classified internal-confidential — export to CSV restricted to manager and finance roles with watermark and logged user id. Rep offboarding triggers forced session revoke and territory reassignment workflow so orphaned dealers do not linger invisible to new hires.

WhatsApp log attachments scanned on upload in staging and production; executable MIME types rejected. Password policy aligned with client IT standard — minimum length, breach list check via Have I Been Pwned API k-anonymity flow — without storing plaintext passwords beyond bcrypt hashes.

• ✓ Sanctum auth with role policies on every mutation
• ✓ Append-only audit log for credit and approval changes
• ✓ Private object storage for WhatsApp log attachments
• ✓ Cloudflare WAF and TLS at edge
• ✓ Encrypted backups and tested restore
• ✓ CRM subdomain excluded from search indexing

Testing and quality assurance

Unit tests cover quotation discount caps, credit headroom calculations, pipeline stage transitions, and RBAC policy matrix. Feature tests exercise API flows: create dealer, attach credit snapshot, build quotation, approve, convert to order request, log WhatsApp entry.

Fixtures built from anonymized real exports caught data quirks — duplicate GSTIN merge, null territory on legacy dealers — before production import. UAT script walked four reps through fifty dealer updates, twenty quotations, and ten dispute scenarios replayed from historical emails.

Finance parallel-run compared credit displays on fifty accounts against accounting exports; zero mismatches required before go-live gate opened. Regression suite runs on CI for every merge to main.

Dispute replay UAT used ten anonymized real email threads from the prior year — reps logged WhatsApp context, managers pulled communication search exports, finance verified order-request matched quotation lines. That scenario pack became regression fixtures so quotation-to-order regressions cannot slip back silently.

Measured outcomes

Within sixty days of launch, median quote-to-order cycle time for standard catalog SKUs fell from four and a half business days to two and a half — a forty-one percent reduction measured from quotation created timestamp to order request accepted by finance. Gains came from reorder shortcuts, fewer back-and-forth phone clarifications, and quotations that did not need retyping into the accounting system.

Credit hold disputes — defined as dealer complaints logged to management when finance blocked an order reps had promised — dropped from thirteen per month in the baseline quarter to five in the second month post-launch, a sixty-three percent reduction. Reps cited visible credit notes and headroom on mobile as the primary behavior change.

Field rep weekly active usage reached ninety-four percent by week eight; inside sales hit one hundred percent earlier because their workflow was entirely desk-based. Two reps needed extra training on WhatsApp logging habits; adoption was not perfect but far above prior CRM attempts that stalled under forty percent.

We did not publish revenue growth percentages. The distributor’s seasonality and product mix drove top-line more than software; leadership explicitly asked for honest operations metrics. Inventory turns and margin were discussed internally but are not reproduced here.

Lessons learned

Wholesale CRM wins when the dealer account page answers field questions in one screen — credit, orders, quotations, communications — not when it copies generic SaaS pipeline jargon without distributor semantics.

Read-only credit visibility earned finance trust faster than giving sales edit access “just in case.” Audit logs on the few finance mutations sufficed for control.

WhatsApp logging as manual structured capture beat waiting for Business API budget and template approvals — reps already worked in WhatsApp; we met them with logging, not migration.

Phased ERP honesty prevented disappointment. Nightly order imports plus documented REST hooks set correct expectations; leadership could budget ERP integration as phase two using our ERP development service without blaming CRM for not being magic realtime sync.

“Our reps lived on WhatsApp and our finance team lived in the accounting software — nothing connected. Web6 built a CRM that shows credit and last orders where we actually sell, and we can finally see what was promised on WhatsApp when a dealer disputes a delivery. We did not get a fantasy revenue number; we got fewer arguments and faster quotes.”

Suggested project visuals

• Dealer account hub: credit headroom, open credit notes, and recent orders in one view for field reps.

  File: wholesale-crm-dealer-account-dashboard.webp

  ALT: Wholesale CRM dealer account screen showing credit limit, credit notes, and order history

• Quotation workflow with catalog search, discount caps, and PDF export replacing Word templates.

  File: distributor-crm-quotation-builder.webp

  ALT: B2B quotation builder with SKU search and GST breakdown in a React CRM interface

• Manager pipeline view with stage notes — quoted, negotiation, won — rolled up by territory.

  File: wholesale-pipeline-kanban-manager-view.webp

  ALT: Sales pipeline kanban board for wholesale distributor opportunities by stage

• Structured logging of WhatsApp reorder chats and attachments for dispute resolution.

  File: whatsapp-reorder-log-dealer-crm.webp

  ALT: WhatsApp reorder communication log attached to a dealer record in CRM