Media & Publishing systems built around real operations.

Use server-side validation, scoped credentials, idempotency keys, signed webhooks, bounded timeouts, retries and a reconciliation job that catches whatever slipped through. The interface needs matching states: pending, failed and recovered, shown honestly. An order must never depend on the customer's browser returning successfully from a payment page.

Include environment documentation, schema and migration history, build and deployment steps, test coverage, operating responsibilities, known limitations and access owned by the client—not only a repository archive.

Cost follows workflow depth, integration count, data sensitivity and how much of the surrounding operation has to change, not screen count. A bounded first release solving one journey is far more estimable than an open brief. Expect a discovery step to produce a defensible range; anyone quoting a fixed figure before understanding the data model is guessing.

Configure wherever a supported product already matches the process closely. Custom work earns its cost when the process is a genuine competitive difference, when licence costs scale badly against users, or when integration and data ownership requirements cannot be met by the product. The honest answer is often a hybrid.

You do. The repository, environments, credentials and deployment accounts should be in your organisation's name from the start rather than transferred at the end. Contracts should state IP assignment explicitly, and any third-party or open-source components should be listed with their licences.

Software needs an owner. That means monitoring, dependency and security updates, defect response, and a route for changes as the operation evolves. Scope should name who does each of those, what response time applies, and what happens if you later move the work in-house or to another supplier.

Start by trying to buy it. If a supported product covers the process closely and the objection is cosmetic, buy it — configuration is always cheaper than construction. Custom work earns its cost in three situations: the process is genuinely part of how you compete, licence costs scale badly against your user count, or integration and data ownership requirements exceed what the product permits. A fourth situation is worth naming because it is common and it is a trap: the process itself is undefined, and software is being asked to settle an argument nobody has had. That never works, and it is expensive to discover late.

A prototype tests one assumption and is meant to be thrown away — it can be ugly, hardcoded and unsafe because nobody real will use it. An MVP is a minimal operable product for actual users, which means it still needs authentication, authorisation, error handling, a way to see what people are doing and a route to fix bad data. Version one is what you build once the MVP has told you which assumptions were wrong. Conflating the first two is the most expensive mistake in this sequence: building an MVP to prototype standards produces something that cannot be operated, and building a prototype to MVP standards wastes months proving something you could have learned in a week.