Subscription billing RFP questions to ask before you pick a partner
If you are issuing an RFP for a self-hosted billing build on Kill Bill or Lago, this page gives you a neutral question set with good-answer criteria and red flags for each area. Start by sending every question in writing, so answers can be compared side by side and kept on file. Ask each bidder for named references that have shipped the exact platform you are choosing, not a general claim of billing experience. Score the written answers against a fixed scoring table agreed before responses arrive, and weight the areas that carry your risk most heavily: catalog migration, payment gateway and tax plugin work, high availability, and upgrade handling. Reviewers should record the evidence behind each score, such as a reference call note or a sample runbook, so the decision can be defended later.
Once you shortlist a partner, tie the scope to a written statement of work that names the platform version, the components in scope, the plugins, the environments, the data migration volumes, the acceptance tests and the support model. Ask for the ownership of each deliverable in writing, including who maintains the catalog after go-live and who owns the upgrade calendar. Confirm which questions depend on your own environment, such as your cloud account, data location and existing gateway contracts, and make those dependencies explicit. Ask whether any enterprise edition or vendor subscription is required, and who pays for it, instead of assuming it is included. DeployOpen does not publish prices. A quote follows from the written scope, so use the same scope document to compare bids on equal terms.
Expect a partner to answer these questions with specifics: named versions, the migration method they used, the test evidence they can show, and the runbooks they hand over. Treat vague answers, promises of a standard timeline without a scope, or refusal to name references as red flags to resolve before award.
Which questions reveal whether a partner can run subscription billing in production
What production experience does a partner have with Kill Bill or Lago?
Ask for named deployments, not general familiarity: which billing engine, which version, how many live subscriptions, and which payment gateways they connected. Kill Bill publishes releases such as killbill-0.24.22, released on September 29, 2026, so ask which release the partner would target and how they tracked changes since their last project. Ask them to describe one failure they resolved in production, such as a stuck invoice run or a duplicate charge, and what they changed afterward. Vague answers about "billing expertise" are a warning sign. Request references you can call, and confirm whether those references ran the platform themselves or only built it.
How will a partner migrate our catalog and active subscriptions?
A credible partner starts with a written mapping of existing plans, add-ons, price versions, currencies and proration rules to the target catalog, then reconciles it against live subscriptions before any cutover. Ask for the artifacts they produce: a catalog mapping document, a migration dry run report, a reconciliation of subscription counts and balances, and a rollback plan. Ask how they handle subscriptions mid-term, customers on legacy prices, and payment method tokens that must move between gateways. Migration effort depends mainly on the number of plans, the volume of active subscriptions and how much history must be carried over, so ask the partner which of these drive their estimate.
How does the partner test payment gateway and tax plugins before go-live?
Ask which gateway connectors and tax integrations they have built or deployed, and whether each one is a vendor-supported plugin, a community plugin or custom code. For each, the answer should cover sandbox testing with declined, partially refunded and disputed transactions, webhook replay tests, and tax calculations checked against a set of reference invoices. Ask who maintains the connector after go-live and how a gateway API change would be handled. Tax rules often change outside the release cycle, so ask how the partner keeps rate tables current and who approves changes. Confirm the exact plugin names and versions in the partner's proposal rather than accepting general claims of support.
Which components fall in PCI DSS scope, and how is card data kept out of them?
A capable partner will draw a data flow diagram showing where card data enters, which components touch it, and where only gateway tokens are stored. Expect them to explain that the billing engine should hold tokens and never raw card numbers, and to name which services are in scope if a vault or orchestrator such as Hyperswitch is added. Ask which Self-Assessment Questionnaire type they expect to apply and why, and who makes that determination (typically your acquirer or assessor, not the implementer). They should also describe how logs, backups and webhook payloads are checked for accidental card data. Ask for the evidence they will deliver for each in-scope component.
How does the partner design high availability and disaster recovery for billing?
Billing outages delay invoices and payment collection, so ask for the target architecture: how the application tier, database and job scheduler are redundant, and where they run across availability zones or data centers. Ask how they test failover, how often they restore backups into a separate environment, and which recovery point and recovery time objectives they propose and why. Ask how scheduled invoice runs are protected from double execution after a failover. Any figures in the answer should come from your agreed requirements, not from the partner's marketing, and you should confirm them in the written scope.
How will the partner handle upgrades, patches and customizations over time?
Ask how the partner tracks upstream releases, which environments they use to test them, and how they keep customizations separate from core code so upgrades stay manageable. A good answer names the pipeline: a staging copy seeded with anonymized production data, automated regression tests for invoice generation and proration, a reconciliation check on test invoices, and a documented rollback. Ask how long they expect a major upgrade to take in their experience and what drives that time, such as schema migrations or custom plugins. Ask who approves each change and how those approvals are recorded for change management.
What security testing and audit evidence will the partner deliver?
Ask what testing they run against the billing stack and how often: application and API testing, infrastructure review, and access control checks on catalog and price changes. Under the FTC Safeguards Rule at 16 CFR 314.4, covered financial institutions must test their information systems, and absent continuous monitoring, the rule calls for penetration testing annually and vulnerability assessments at least every six months, so confirm which of those your partner will perform and which you must schedule yourself. Ask for sample evidence: change tickets, access review records, audit logs of price changes and invoice credits, and test reports. Note that DeployOpen prepares systems, controls and evidence for assessments but does not issue audit reports or certifications.
What does the support model cover, and who pays for what after go-live?
Ask for a written split of responsibilities: who monitors the platform, who handles incident response and on what hours, who applies security patches and who owns gateway and tax connector updates. Clarify whether support is a fixed-scope handover, a retainer for ongoing operations and upgrades, or your own team with partner backup. DeployOpen offers ongoing support, upgrades and continuous testing as retainers, and most engagements are fixed-scope with scope, deliverables and timelines agreed in writing before work starts. Ask each bidder for their response process, escalation path and what is included in a retainer versus billed separately, then compare those answers in the quote.
Do we need a vendor enterprise edition, and what does it add?
Ask the partner to state, in writing, which components run on open-source releases alone and which would benefit from a commercial support agreement, and to explain the reasoning for each. For Kill Bill, ask about the commercial enterprise offering the partner would recommend and what it includes, such as vendor support or additional modules, and confirm the current terms directly with the vendor before deciding. Ask how the partner would operate without that agreement, including how they would fix a critical defect in the core. The cost of any license should be compared with the in-house effort to maintain the same capability, so ask for both figures in the bid.
Which partner answers should make you stop the evaluation
Partner answers reveal how a bid will go after signature. The cards below cover the answers that should pause or end a self-hosted billing evaluation, and the follow-up question that separates a credible partner from a risky one.
No hands-on work with the named platform
A partner who has only read the Kill Bill or Lago documentation will learn the platform on your budget. Ask for named engagements where the team configured the catalog, ran invoice cycles, handled retries and reconciled payments on the same platform version you plan to run. Ask which components they have operated in production and which they have only demonstrated. A credible answer names the modules, the data volumes and the problems they had to fix, without naming clients you cannot verify.
Fixed timeline before catalog mapping
A firm go-live date offered before the team has inventoried plans, add-ons, price books, currencies and active subscriptions is a guess. Catalog design and data mapping decide most of the schedule, so a responsible partner will hold the date until those are scoped. Ask what they need to see before committing, which artifacts they produce in discovery, and how a written scope changes the timeline if the source billing data is messier than expected.
Vague answers on the upgrade path
Self-hosted billing stays in service for years, so the upgrade path matters as much as the first install. If a partner cannot explain how they test a new release against your catalog, custom plugins and invoice templates, or how they roll back a failed upgrade, you will carry that risk alone. Ask for their upgrade runbook, the staging environment they use, how they handle database migrations and the release cadence they will follow for your platform.
Card data treated as in scope by default
A partner who assumes the billing platform must store card numbers, or who does not ask about your gateway and tokenization design, will widen your PCI DSS scope without telling you. Ask how they keep payment data in the gateway or vault, which billing components they expect to touch cardholder data, and which self-assessment questionnaire type they expect your design to produce. The answer should start with scope reduction, then describe the evidence they will produce for your assessor.
No plan for enterprise edition needs
Some features that finance teams expect may sit outside the open-source core or depend on commercial support, so the licensing question should appear in the first call. Ask the partner which capabilities they would need from an enterprise option for your requirements, how they would confirm those features with the vendor, and what changes in the architecture or support model if you stay on the community edition. A partner who cannot separate open-source features from commercial offerings is not ready to scope your budget.
Delivery depends on one engineer
When one person holds the catalog logic, the gateway integration and the deployment scripts, a holiday or departure stops the project and the run book. Ask who is named on the team for architecture, billing configuration, integration, security and operations, and how knowledge transfer is documented. Request the support model after go-live, including who answers a failed invoice run at the end of a month and how response is measured in the written scope.
How Kill Bill and Lago change the questions you ask
Start by naming the platform in the RFP. Kill Bill is an open source subscription billing and payments platform licensed under Apache-2.0, and the latest tagged release named in our research is killbill-0.24.22, published on September 29, 2026. That date matters for the questions you ask: a partner should be able to explain which release line they would deploy, how they track upstream changes, and how they would test a plugin or catalog change before it reaches production. Ask each bidder to describe its experience with Kill Bill specifically, including the number of Kill Bill deployments it has run in regulated environments, the plugins it has configured, and how it handled the catalog and invoice data in those projects. Ask them to confirm current licensing terms and release status directly with the vendor and in the source repository, since both can change between the RFP and contract signature.
Lago is a separate open source billing option, and it should be evaluated on the same criteria rather than assumed to be interchangeable with Kill Bill. Ask partners to walk you through the data model, the event ingestion path, the invoice and payment flows, and the components they would run in your environment. A partner who has deep knowledge of one platform may only have surface knowledge of the other, so require each bidder to answer the same catalog, gateway, tax and upgrade questions for the platform it recommends and to explain why it fits your pricing model. Ask for the names of the components, the database and message infrastructure, and the monitoring each platform would need in your cloud account or data center.
How to weigh answers for SaaS, usage-based and regulated buyers
For a SaaS company, the core risk is the catalog and the revenue cycle. Plans, seats, trials, upgrades and proration are rules that change often, and every price change touches live subscriptions. Score partners highest on how they have handled catalog design and catalog migration on Kill Bill or Lago, because a mistake there reaches every invoice. Ask them to describe how they tested proration and mid-term plan changes before cutover.
- Catalog migration evidence: Ask for a worked example of moving an existing plan set onto the new catalog without re-pricing active subscriptions.
- Gateway and tax plugins: Ask which payment gateway and tax engine plugins they have deployed, and how they validated invoice totals against the previous system.
- Upgrade handling: Ask how they apply Kill Bill or Lago version upgrades to a live invoicing run and what rollback looks like.
- Support model: Ask who answers a billing defect at 2 a.m. and what response commitments are written into the statement of work.
- Enterprise edition needs: Ask which features the partner would require from a commercial edition or paid support tier, and why.
Confirm every answer against a sandbox run using your own plan list before you score it.
Usage-based and AI API billing shifts weight toward the event pipeline. Invoices depend on metered events arriving complete, once, and on time, so the partner's answers about ingestion, storage and reconciliation matter more than catalog depth. Ask how they design event volume capacity, how they handle late or duplicate events, and how usage is reconciled to invoices before any customer sees a bill.
- Event ingestion and storage: Ask how they sized event volume and which event store they chose for aggregation, and what they measured to justify it.
- Prepaid credits: Ask how credit balances, drawdowns and top-ups are modeled and how they are reversed on refunds.
- Real-time usage visibility: Ask what customers and internal teams can see during a billing period and how far that view lags the raw events.
- High availability: Ask how the ingestion path behaves during a node failure, and whether events are buffered or lost.
- Reconciliation: Ask for the report they would produce to show that every metered unit ties to an invoice line.
Ask for a load test plan with your expected event profile, and treat any answer without measured figures as unverified.
Regulated buyers, including broker-dealers, lenders and card-accepting institutions, need answers that produce evidence. The FTC Safeguards Rule at 16 CFR 314.4 describes controls such as encryption, multi-factor authentication, change management, activity logging and periodic penetration testing and vulnerability assessments. SEC Rule 17a-4(b) describes retention of billing records and business communications for a set period. Whether either rule applies to your billing records is a question for counsel, so treat these as evidence drivers to confirm, not as settled obligations.
- Access and change control: Ask how catalog and price changes are approved, logged and restricted to named roles.
- Logging and retention: Ask where activity logs and invoices are stored, how long they are kept, and how they are retrieved for an examiner.
- Encryption and MFA: Ask how data is encrypted in transit and at rest and how administrator access requires multi-factor authentication.
- Testing cadence: Ask who performs penetration tests and vulnerability assessments on the billing stack and how findings are tracked.
- Card data scope: Ask which components would hold card data and how gateway tokens keep the billing system out of direct card handling.
Have counsel and your compliance lead confirm which obligations apply before you weight these answers.
How to run the evaluation from shortlist to signed scope
This sequence turns a shortlist of implementation partners into a comparable set of written answers, a tested view of migration risk and a fixed-scope statement of work that names the platform and version. Each step produces a record that the buying team can score and defend.
Define requirements before sending any questions
The billing lead, finance controller, security lead and engineering owner agree on the pricing models, invoice volumes, currencies, tax jurisdictions, payment gateways, dunning rules and general ledger interface the system must support. Write these down as a short requirements document with a named owner for each item. The step is done when every partner receives the same requirements and no one can add scope later without a change record.
Send written questions to each shortlisted partner
Send the same written question set to every bidder, covering experience with Kill Bill or Lago, catalog migration, gateway and tax plugins, high availability, upgrade handling, the support model and any enterprise edition needs. Ask each partner to name the platform and the exact version they would deploy, and to state which parts of the work they would subcontract. The step is done when every answer is in writing and the answers can be compared line by line.
Request a sample catalog migration plan
Ask each partner to take a representative slice of your current catalog, such as three plans with different pricing models and one active subscription with a mid-term change, and show how it would map to the target platform. The plan should list data mappings, invoice numbering decisions, payment token handling and the reconciliation checks used before cutover. Reviewers should look for specific handling of edge cases rather than generic migration language. The step is done when each partner has a written plan that finance and engineering have both reviewed.
Check references for the named platform and version
Speak with references that ran the same platform at a similar version, and ask about the upgrade path, incident handling and how the team handled a gateway or tax plugin change. Confirm the reference used the partner's hands-on delivery team, not a different group. Record what each reference confirmed and what they declined to discuss. The step is done when each finalist has at least one reference check recorded against the platform and version it proposed.
Score responses against the requirements
Build a scoring sheet with weighted criteria taken from the requirements document, such as migration approach, security and PCI scope handling, high availability design, upgrade and support model, and evidence delivered for audit. Each reviewer scores independently, then the group reconciles large differences in a recorded session. The step is done when the scoring sheet has a rationale for every score and the top finalists are agreed in writing.
Agree a fixed-scope statement of work before work starts
Convert the chosen partner's answers into a written statement of work that lists deliverables, the platform and version, the migration scope, acceptance criteria, responsibilities for infrastructure and gateway accounts, and any retainer for ongoing support. Ask for a quote that separates one-time implementation fees from recurring hosting, support and plugin costs, and confirm which items are excluded. The step is done when both parties sign the statement of work and no implementation work begins before signature.
How to score partner responses consistently
This table gives each scoring criterion for a self-hosted billing bid, what a full-mark answer contains, and what should score zero. Weights are relative guidance you set for your own priorities, so adjust them before issuing the RFP. Use the zero triggers to disqualify vague answers early and keep scoring comparable across bidders.
| Criterion | Relative weight (set by buyer) | What full marks look like | What triggers a zero |
|---|---|---|---|
| Platform experience | High | Named Kill Bill or Lago deployments of comparable scale, with the partner's own engineers described and reference contacts offered | Cannot name any live self-hosted billing deployment or relies only on generic open-source experience |
| Migration method | High | Written catalog mapping approach, plan for active subscriptions mid-term, invoice numbering continuity and a reconciliation test before cutover | Proposes a bulk import with no reconciliation step or no plan for in-flight subscriptions |
| Plugin and gateway depth | Medium to high | Lists the gateway and tax plugins it has built or configured, states which are community and which are custom, and names what it will test | Assumes a plugin exists without confirming its maintenance status or the gateway versions it supports |
| High availability and upgrades | High | Describes the target topology, database and queue resilience, a tested upgrade runbook and a rollback path tied to the release notes of each version | No upgrade procedure, or upgrades described as a one-time task after go-live |
| Security and audit evidence | High | Maps controls to the buyer's requirements, including card data scope, access logs and change records, and offers evidence samples | Refuses to provide evidence samples or treats audit readiness as the buyer's problem only |
| Support model | Medium | Defined response channels, escalation path, on-call coverage, and clear separation between partner support and vendor support | No stated support hours, escalation path or named owner after go-live |
| Licensing clarity | Medium | States which components are open source, which features need an enterprise edition, and confirms licensing terms in writing with the vendor | Assumes enterprise features are included or cannot say which edition each feature needs |
What buyers ask before hiring a billing implementation partner
The first question buyers ask is what drives cost and effort, and the answer starts with scope. A self-hosted Kill Bill or Lago deployment is priced and timed by the number of product lines, the depth of the catalog (plans, add-ons, usage meters, price overrides and trial rules), the number of payment gateways and tax engines that must be connected, and the volume of active subscriptions to migrate. Each gateway connector and each tax plugin is a separate integration with its own credentials, test cases and failure handling. Enterprise edition needs add another decision: whether the buyer requires vendor-backed support or features that sit outside the open-source core, such as the Kill Bill Aviate offering named in the cost comparison. Ask every bidder to price these items separately so the quote shows which driver moves the number.
The second question is what is one-time and what recurs. Build work (catalog modeling, integration, data migration, reconciliation, testing and cutover) is one-time and billed as a fixed-scope project. Recurring work covers hosting, database backups, security patching, version upgrades, plugin maintenance when a gateway or tax API changes, and on-call response. Ask bidders to state whether upgrades and plugin updates are included in the retainer, how many upgrade cycles a retainer covers, and who owns the test environment used for them. Confirm the Kill Bill and Lago release each bidder proposes against the current entries on the project's GitHub repositories before accepting a version plan. The third question is what the work needs from the buyer, and whether the internal team can do it. Finance must sign off on the revenue rules, invoice numbering, credit approval and the general ledger mapping, because the partner cannot decide those for you. Catalog owners must confirm every plan and price, and an engineering owner must grant access to staging, production, payment sandboxes and the tax engine account. If the internal team cannot staff these roles for the project duration, price that gap into the decision. A useful next step is to compare the outcome with the requirements checklist and the cost comparison before issuing the RFP.
How DeployOpen scopes, delivers and supports billing evaluations and builds
DeployOpen plans, implements, integrates, hardens, documents and supports self-hosted open-source platforms inside the client's own infrastructure, including subscription billing built on Kill Bill or Lago. Each evaluation or build starts with a written scope agreed before work begins, so the partner's answers to the questions on this page can be checked against defined deliverables.
How we scope
- Written scope that names the billing platform, catalog size, gateways, tax plugins and general ledger interfaces in play
- Agreed list of deliverables, acceptance criteria and responsibilities before any work starts
- Confirmation of which component versions and enterprise edition features the scope depends on, verified at the vendor
How we build
- Catalog migration plan with test cases comparing source and target prices, plans and add-ons
- Gateway and tax plugin configuration tested in a non-production environment before cutover
- High availability and upgrade procedures tested and documented in the client's own cloud or data center
How we hand over
- Runbooks for deployment, backup, restore, upgrades and rollback, written for the client's operations team
- Architecture, configuration and access control documentation kept as evidence for finance and audit reviews
- Knowledge transfer sessions with named client staff, recorded against the agreed scope
How we support
- Ongoing support, upgrades and continuous security testing available as a retainer after handover
- Patch and version upgrade planning aligned with the vendor's published releases and the client's change process
- Security testing of billing APIs, integrations and payment flows, scoped and agreed in writing

