A subscription business is not a shop that charges a card every month. It is an operating model: a plan catalogue, a renewal engine, a way to suspend access when payment fails, a reconciliation routine and a support queue that expects cancellation requests. When those pieces are assembled from different tools, the gaps show up as failed renewals, confused customers and a month-end that never closes cleanly.
This comparison sets out four common ways to run subscription commerce operations in Bangladesh, the trade-offs between them, and the questions that reveal which model fits your business.
What subscription operations must handle
Before choosing a model, separate the work into functions. A subscription operation has to do all of the following, even if one tool covers several:
- Plan catalogue: monthly, quarterly, annual, usage-based, seat-based, bundles, trials and grandfathered prices.
- Sign-up and mandates: card authorisation, wallet or bank mandate where available, and a record of what the customer agreed to.
- Renewals: scheduling, retries, grace periods, suspension and reactivation.
- Changes: upgrades, downgrades, pauses, address changes and proration.
- Entitlements: what the customer can use — seats, features, content, delivery quantities or service hours.
- Fulfilment: physical delivery, digital access or service scheduling, with subscription state as the source of truth.
- Billing and revenue: invoices, receipts, taxes, deferred revenue and settlement reconciliation.
- Support and retention: cancellation reasons, save offers, refunds and win-back.
No single category is best at all eight; the comparison is about which functions one system should own.
The four operating models
Most SMEs end up with one of these patterns, sometimes with a hybrid layer added later.
1. Storefront subscription app
If the business already sells through Shopify or WooCommerce, a subscription app adds recurring plans, a customer portal and automatic orders inside the storefront. It is the shortest path when the catalogue, checkout and delivery already live there.
Strengths: fast launch, familiar administration, and the subscription order looks like any other order.
Limits: payment support is constrained by the storefront's integrations. Local wallet or bank mandates may need manual handling, and a stack of apps can create conflicting renewal, tax and discount rules. Deep entitlement control and complex proration are usually outside what a storefront app is designed to do.
2. Billing platform plus a light catalogue
A dedicated billing platform owns plans, invoices, retries, proration and revenue reporting, while the storefront or application owns the customer experience. This model suits digital services and card-first businesses with simple fulfilment.
Strengths: strong renewal and dunning logic, clean invoice history, and useful APIs and webhooks.
Limits: you now run two systems. The product must ask the billing platform whether access is allowed, and someone must reconcile settlement reports against the local ledger. Payment coverage is the deciding question in Bangladesh: not every wallet supports automatic recurring mandates, so confirm the arrangement with each provider before designing around it.
3. ERP or accounting-led recurring invoicing
For business-to-business subscriptions sold on purchase orders and bank transfers, the finance system often leads. Recurring invoices are generated from a contract, the customer pays by transfer or cheque, and access is managed by the delivery team.
Strengths: fits existing credit terms, procurement and tax processes; finance already trusts the ledger; there is no new billing platform to explain to auditors.
Limits: collections remain manual, renewals depend on a person following up, and there is no reliable link between payment status and entitlement. Self-service changes are usually absent. This model works at low volume with high contract value and breaks when the customer count grows.
4. Custom subscription operations layer
A custom layer models your plans, mandates, entitlement rules and reconciliation exactly, then connects the storefront, payment providers, delivery and finance systems. It is justified when subscriptions are central to revenue and the standard tools force a workaround on every third order.
Strengths: exact fit for local payment methods, complex plan rules, multi-system entitlement, audit requirements and distributor or branch structures.
Limits: higher upfront cost and ongoing ownership. The build must include dunning, retries, proration, tax and reporting, or you have simply recreated a worse billing platform. A custom layer is a product, not a one-off script.
Comparison criteria
Score each option against the same criteria, using your own volumes and exceptions:
- Payment coverage: cards, wallets, bank transfer, cash collection and the mandate type each requires.
- Renewal control: retry schedules, grace periods, suspension rules and customer communication.
- Plan changes: upgrades, downgrades, pauses and proration, including mid-cycle changes.
- Entitlements: seats, features, usage limits and delivery quantities, enforced at the point of use.
- Reconciliation: gateway settlement, bank deposit, invoice and ledger matched without spreadsheets.
- Integration: storefront, CRM, delivery, accounting and support desk.
- Cost model: licence, transaction fees, app subscriptions, implementation and internal administration.
Where the models usually break
Payment methods. A model that assumes automatic card renewal will struggle if most customers pay by wallet or bank transfer. Design for the payment mix you actually have.
Dunning. A failed payment is not an error; it is a process. The system needs a retry schedule, a grace period, a clear suspension point and messages that give the customer a way to fix the problem.
Entitlements. Billing says paid; the product says expired. That gap is where customers lose access after a successful renewal or keep access after cancelling. One system should own the decision and the other should consume it.
Reconciliation. Subscription revenue is repetitive, which makes small mismatches easy to ignore until quarter-end. Match settlement to invoice to ledger on a fixed schedule and treat unmatched items as a queue with an owner.
Decision guidance
- Storefront app: a standard plan set, a catalogue already on Shopify or WooCommerce, card or prepaid collection, and a small operations team.
- Billing platform: digital delivery, card-first payments, a simple catalogue, and a product team that can consume webhooks.
- ERP-led invoicing: a small number of high-value B2B contracts with purchase orders, credit terms and manual collection.
- Custom layer: multiple payment methods, complex plans, branch or distributor structures, or entitlement across several products.
- Hybrid: keep the storefront for commerce, use a billing platform for renewals, and build only the step that is genuinely unique.
Vendor questions
- Which payment methods can be charged automatically, and which require manual follow-up?
- What happens on the third failed retry, and who can override a suspension?
- Which reports reconcile settlement, invoices and revenue, and who owns the unmatched queue?
- Can we export every customer, mandate and invoice if we leave?
Operating metrics
- Renewal rate by plan and cohort, using a written definition.
- Involuntary churn: subscriptions suspended or cancelled because payment failed.
- Days to reconcile each settlement batch, and value left unmatched.
- Support contacts per hundred active subscriptions, split by billing and access issues.
Where this fits in our work
If subscriptions are becoming a material part of revenue, start with Commerce Systems. The pricing page explains how discovery, build and run work are scoped. For the finance routine behind the model, read how daily payment reconciliation changes as you grow.
Sources
- Stripe, Billing documentation, accessed 17 November 2026.
- Shopify, Subscriptions and purchase options, accessed 17 November 2026.
- WooCommerce, Subscriptions documentation, accessed 17 November 2026.
- Bangladesh Bank, Mobile Financial Services data, accessed 17 November 2026.
- PCI Security Standards Council, accessed 17 November 2026.
Responses (0)
No comments yet. Add the first one below.
Leave a response
Your email address will not be published. Required fields are marked *.