“How much is an OA?” is as useful as “how much is a workshop?” Until headcount, flow count, a contract ledger, and finance integration are set, any quote is a guess. Use checkable dimensions below to put budget and cycle in a sane range.
| Phase-one scope | Typical cycle | Where money actually goes |
|---|---|---|
| 3–5 high-frequency approvals + basic org permissions | 6–10 weeks | Flow inventory, form routing, chase, trial close |
| Plus contract ledger / seals / knowledge base | 10–16 weeks | Ledger model, renewal to-dos, document rights, training |
| Connect HR/finance or replace a legacy OA | 3–6 months | APIs, read-only archive, dual-run, permission isolation |
The numbers are experience ranges, not a quote. Turning on Feishu, DingTalk, or WeCom apps can be faster, but when conditional routing, seals, and your amount splits do not match, the rewrite cost comes back. Compare: Feishu, DingTalk, and a custom back office: what actually differs. Packaged OA is cheap and still unusable; see A packaged OA nobody uses is almost as bad as having none.
Why can two “just build an OA” quotes differ by several times?
The gap is flow count, amount/department splits, how many role and data-scope layers, contracts and seals, who writes policy and keys historical contracts, and whether attendance or vouchers connect. One hidden line: who is the system admin. Without an owner, build stalls on “change one more node.”
Can the cycle compress to two weeks?
What can ship in two weeks is usually a packaged flow with your org name swapped in. A permission matrix, chase rules, and a real-ticket trial will not be solid. Ship leave, expenses, and seals first rather than a shell nobody opens. Whether an unused system was worth building: see If nobody opens the system after launch, was it worth building?.
Will old data and a legacy OA slip the date?
Yes. Whether historical contracts move, and whether old tickets are read-only or re-run, must be written into the plan. How to migrate: see How to migrate data from a legacy OA. A vendor-abandoned swap needs its own risk estimate; see Replacing a vendor-abandoned OA: what to watch.
How do you hold budget without shipping a prop?
Name the three–five approvals that must leave WeChat. Cover salary, contracts, and export first. Contract and ERP modules can wait for phase two. Do you need a full ERP? See Do you need a full ERP?. The full path: see From brief to launch: how a management system is built.
Price hidden cost in, or a cheap quote doubles later
Beyond the quote: interview hours from each department, policy cleanup, scanning historical contracts, dual-run, and admin training. If those are not in the plan, they become “why isn’t it done.” Cutting twenty unused nodes in phase one usually saves a build cycle without hurting adoption. Maintenance, servers, and certificate renewal should be line items.
After the brief is clear, DaXi returns scope, phases, and an acceptance list—not an unverifiable lump sum. Bring headcount, current tools, and the flows that must ship first to the Management systems service page. If you are still deciding whether to build at all, start at Why companies need a management system.
When you compare vendors, make them quote the same approval list and write whether trial, training, and dual-run close are included. Comparing only the total rewards whoever dared to omit lines. Phasing lets you prove chase and permissions before you buy a contract ledger and ERP modules—less risk than signing everything. If Feishu or DingTalk is already in use, phase one can customize only what the suite cannot change; see Feishu, DingTalk, and a custom back office: what actually differs. Which flows a growing company should ship first: see Growing companies: which approvals to ship first.