Live

Hotel data pipeline

An automated pipeline that delivers each hotel's daily import file, so nobody exports and reformats data by hand.

Details anonymised to protect client confidentiality.

The dashboard: every property's daily run at a glance. Names and figures replaced with placeholders.
The daily flow

Every day, each property's file goes from its PMS to the revenue tool.

  1. Step 1

    Pull

    Signs in to each property's PMS and downloads the forward-looking stay data.

  2. Step 2

    Convert

    Turns it into the exact file the revenue tool imports, and checks it before anything is sent.

  3. Step 3

    Deliver

    Emails each property's file to the revenue tool, with a copy to anyone on CC.

Each property runs on its own schedule: the data is pulled first, then sent at its delivery time.

Context

A hotel revenue-management firm uses an AI-powered revenue tool to set and recommend rates across the properties it manages. Each property's data lives in that hotel's property management system.

The boring work

Every day, someone pulled forward-looking stay data out of each hotel's property management system, converted it into the exact file format the revenue tool imports, and sent it on. Repeated across many properties, every day, and easy to get wrong.

What we built

An automated pipeline that does the whole chain on a schedule.

Before / after

Five manual steps became one glance.

PDone by a personDone by software
Before
01A person logs in to each hotel's property management system.P
02They export the forward stay data.P
03They reformat it into the revenue tool's import format.P
04They send or upload the file.P
05They repeat for every property, every day.P
After
01The pipeline pulls each hotel's forward stay data from its property management system.
02It converts the data into the import file the revenue tool needs.
03It emails each property its file daily.
04A person only looks if something is flagged.P
Checked before it's sent

Every file is checked before the revenue tool sees it.

  • Every stay date is there, in order, with no gaps
  • Total rows are removed
  • Rooms sold never exceed rooms available

If a check fails, nothing is sent and the team is told why.

A single run: pull, convert, deliver, with the data checks. Placeholders throughout.
When something goes wrong

It retries first, then tells you what to fix.

  1. The PMS can't be reached, or it rejects the sign-in.
  2. It tries again, and again a little later.
  3. Still failing? Owners and Admins get an email, and the dashboard shows the cause in plain words.

What changed

  • The daily job no longer depends on someone remembering or having time.
  • Files arrive consistently in the right format.
  • Staff time goes to analysis instead of reformatting.
  • Fewer manual errors, by design.