How an IR code library cut remote set-up from three weeks to four days

A 400-room hotel group with three properties can spend three weeks per property on remote-control commissioning and still carry forty-plus unresolved remote complaints into the next project. That is the shape of the problem in this case: not a failing remote, but a failing process for knowing which codes a remote should send. The group ran roughly sixty room types across three vintages of in-room TV hardware, with an in-house AV team of four and a contracted integrator handling the physical work. The numbers below are indicative of that engagement rather than a published benchmark, but the mechanism is the part worth reading.

Context: three properties, three hardware vintages, one small AV team

The estate had accumulated rather than been designed. The oldest property used TVs from the early flat-panel era, the middle one a later generation with different IR protocols, and the newest a mix of both plus set-top boxes. Across them, sixty room types meant sixty combinations of screen, box and remote — and each combination could need a different code set depending on how the room was wired.

Four in-house AV staff covered all three properties. The integrator handled installation and cabling but treated code capture as a favour rather than a deliverable. When a room was refurbished or a TV swapped, whoever did the swap had to find the right code by trial and error, or by asking someone who remembered doing it before. That works until the person who remembers leaves, or until a property runs two refurbishments in the same quarter.

The baseline: three weeks per property and a standing complaint queue

Before any new set-up began, the group logged more than forty unresolved remote complaints per month across the three properties. These were not obscure faults. Most were simple mismatches: a replacement TV that used a different power toggle, a set-top box whose volume commands had been learned from the wrong source remote, a housekeeping remote that had never been paired to the correct room. Each one was individually trivial and collectively expensive, because each consumed a site visit or a phone call from a team with no spare capacity.

The rollout itself took about three weeks per property, and most of that was manual code capture: pointing a remote at a receiver, pressing every button, recording the result, testing it against the actual device, and repeating when the captured code was marginal. Repetitive work, but not something you could script, because every property had its own spreadsheet and every spreadsheet had drifted from the hardware actually installed.

Two failure modes kept recurring. The first was treating the spreadsheets as authoritative when they were only a snapshot. Codes that worked in one property got copied into another where the TV firmware differed, and the copied code failed intermittently rather than outright — which is worse. The second was that code knowledge lived with individuals. When the integrator's lead technician was unavailable, the in-house team had no way to reproduce his capture method.

What was actually done: four actions, each with a cost

1. Build a searchable IR code library instead of a per-property spreadsheet

The team replaced the three spreadsheets with a single searchable IR code library, indexed by device model, protocol and button function. The reason was not tidiness. A spreadsheet assumes you already know which row you need; a searchable library lets a technician start from the device in front of them. The cost was real: roughly two weeks of one engineer's time to migrate the existing captures, plus the ongoing discipline of not letting the library fork back into local copies.

2. Prune the library to the models actually in the estate

Migration surfaced hundreds of codes for devices the group did not own — leftovers from previous integrators, trade-show samples, models decommissioned years earlier. The team removed anything not present in the current estate, which cut the library by roughly half. This was the least popular action and the most useful. A code library that contains everything is a code library nobody trusts, because the first three search results may be for hardware that no longer exists.

3. Script the hand-off into the commissioning tool

The library exported a machine-readable file that the commissioning tool consumed directly, so a technician selected a room type and the correct code set loaded without manual entry. Most of the time saving came from here, and it depended on the integrator exposing the commissioning tool's input format. The cost was a short integration effort and a commitment to keep the export schema stable; when the schema changed, the script broke, and the team learned to version it.

4. Assign one engineer as the library's owner

One named engineer became responsible for accepting new codes, retiring dead models and reviewing changes. The rule was simple: no code enters the library without a test against the physical device, and nobody edits the library without telling the owner. That cost perhaps a day a month of the engineer's time. It also stopped the slow re-fragmentation that had produced the original problem.

The result: two quarters, two metrics

Over two quarters, remote set-up time per property fell from about three weeks to four days. The backlog of remote-related work orders dropped by roughly two-thirds, from the forty-plus monthly figure to the low teens. Neither number should be read as a clean before-and-after. The first property to use the new library benefited from a refurbishment that had already replaced some of the oldest hardware, so part of its gain came from the estate getting simpler at the same time.

Metric	Before	After two quarters
Set-up time per property	About 3 weeks	About 4 days
Open remote work orders per month	40+	Low teens
Sources of code truth	3 spreadsheets plus individual memory	1 library with a named owner

The four-day figure is the one to treat with suspicion. It measures set-up time for a property whose hardware was already mapped in the library. The first property took longer than four days, because its codes had to be captured and tested before they could be loaded.

What transfers, and what does not

The library-first approach transfers almost anywhere. So does the single-owner rule. Both are organisational choices that cost time rather than money, and both fail in predictable ways when skipped: the library becomes a dumping ground, or it drifts back into per-site copies within a year.

The four-day result depends on three conditions I would not assume. First, hardware homogeneity: sixty room types is a lot, but they collapsed into a manageable number of device families once the old models were retired. A group with genuinely mixed hardware in every room would see less benefit. Second, an existing commissioning script or tool that accepts a machine-readable input. If every room is configured by hand through a menu, the library saves search time but not entry time. Third, an integrator willing to share codes and keep the export schema stable. An integrator who treats code capture as proprietary leverage will stall the whole approach, and no amount of internal discipline fixes that.

What does not transfer is the specific timeline. Two quarters was enough here because the estate was small and the team was motivated. A larger group, or one running simultaneous refurbishments, should expect the migration and the pruning to compete with live work — and should not promise a four-day set-up before the IR code library has been tested against every device family in the estate.

If you are weighing this up, the useful question is not whether you need a code library, since most multi-property operators do. It is whether you have a named owner and an integrator who will hand over codes. Without those two, the IR code library becomes another spreadsheet with better search. With them, the set-up time drops, because the technician stops guessing and starts loading. The library is the easy part; the ownership is what holds it together.