Hospitality Technology

Shiji Outlines PMS Migration Plan Built on 20-30 Tested Workflows

Shiji's software development chief sets a 20-30 workflow test standard for PMS migration, with a six-stage timeline and Day 1 prioritization by revenue risk for a 100-room hotel.

PMS Migration: How To Get Connectivity Right
PMS Migration: How To Get Connectivity Right — The Metropolitan Museum of Art / The Met · CC0

A property management system migration should be validated against 20 to 30 real daily hotel workflows before go-live, according to Vitaly Gamarnik, Director of Software Development at Shiji Group, who published a step-by-step connectivity framework this week using a 100-room hotel as a worked example.

Shiji, whose cloud portfolio spans PMS, point-of-sale, guest engagement, distribution, payments and data intelligence for more than 91,000 hotels worldwide, framed the guidance around one central claim: most migration failures come from treating connectivity as a systems list rather than a workflow problem. Gamarnik's advice is vendor-specific in its examples but general in its method. Whether independent operators can resource it is a question the framework does not address.

What does the framework require first?

Hotels should start with workflows, not systems. Knowing the channel manager or payment gateway by name is only the starting point. Front office, reservations, revenue management, housekeeping, F&B and finance teams must map how departments exchange information before any integration decisions are made.

From there, operators build a connectivity inventory covering four items:

  • Vendor, product and version — different versions may need different integration methods.
  • Connection type — cloud or on-premise, plus any additional infrastructure.
  • Data flows — room status, postings, rates and inventory, and their direction.
  • Vendor ownership — technical contacts, contracts, support terms and sandbox availability.

Gamarnik's timing recommendation is blunt: "Start building your connectivity inventory during the sales and scoping phase, not after implementation kickoff." That is a discipline demand, not a product feature. It puts unpaid work in front of the contract signature.

How are integrations prioritized?

Not every connection needs to exist on Day 1. Gamarnik splits integrations into three categories. Day 1 covers revenue and compliance: channel managers, CRS, GDS, booking engines, revenue management systems, payment gateways, mandatory fiscal interfaces, door locks, phones and call accounting. Phase 2 covers loyalty and CRM platforms, guest messaging, housekeeping apps, upselling tools and parking — though hotels with heavily used loyalty programs may need them from the start. The third category is retirement.

His test for the split: "If an integration fails on go-live morning, will the hotel lose revenue, break a legal requirement or negatively affect the guest experience? If yes, it should be a Day 1 priority." The framework also warns against automatic reconnection of every legacy system, noting the Shiji Platform consolidates spa, activities, golf and events management through Meridian Experiences on a shared guest profile — a pitch that conveniently favors Shiji's own stack.

How should testing work?

The stated goal is to find problems before guests do. Testing runs in layers. A UAT environment mirrors the expected production configuration. Sandbox-to-sandbox testing applies wherever third-party vendors offer it; where no sandbox exists, production-to-production testing may be necessary under controlled conditions. Real guest data, including personally identifiable information, should never enter test environments. One to two weeks before go-live, on-premise integrations such as door locks or call accounting can be switched from the legacy PMS to the new one for roughly an hour, then switched back. Large cloud distribution partners that forbid production switching must instead confirm message receipt through their system logs.

What is the timeline?

The plan assigns owners and deadlines across six stages: planning and scoping at 8 to 12 weeks out, prioritization at 6 to 8 weeks, UAT build and mapping validation at 4 to 6 weeks, training and workflow testing at 2 to 4 weeks, final production checks at 1 to 2 weeks, and launch monitoring through the first 72 hours after go-live.

Gamarnik's closing argument is that migration should not recreate yesterday's ecosystem in a new system. For operators who follow the sequence, the reward is fewer failures on the first morning. For those who skip the inventory, the framework offers no rescue plan.

pms-migrationhotel-integrationsworkflow-testinguser-acceptance-testing-uatconnectivity-inventory

More from Marcus Bennett

Marcus Bennett

Show full bio

Market editor covering technology and product news in Restaurant and hotel operations at The Pass Brief.

798 articles

Pairings

« Previous article