Why we developed our own Inventory Mapping Technology

Integrating hundreds of thousands of hotels into a marketplace is a very complex task.

Data needs to be gathered from multiple sources, however each of them can use varying formats and query logic.

Not only because the data needs to be gathered from multiple partners (who all use different data formats and have adopted varying query logic) — but more importantly, the same hotels can be differently named by each partner. Let’s take a look at one example:

The hotel can be named Melia Hotel Wien by one provider, Melia Vienna Hotel by another provider and Melia Hotel by a third one. Simply combining all Melia hotels into the same basket would be a mistake, since there may be two or more Melia hotels within the same city.

Coupling this with different data formats and incomplete information fields, can make it prone to a number of inaccuracies. In addition, what if there are five different hotel descriptions from five different providers? Which one is most accurate?

Address information can vary too, both in language and format. A district in Vienna can be called by its name or by its number (Donaustadt vs. 22nd district).

You get the point. There are endless scenarios that make it very difficult for a computer to match duplicate inventories. Unless it understands the data.

Machine Learning and Neural Nets

Understanding data is something which humans do very well, but at an incredibly low speed. It not only takes very long, but also adds significant labor costs to the process.

To combat this, we have taken the approach of training a neural net to understand the data and map it accordingly.

Image result for neural network

While this may not eliminate the human factor entirely, we estimate that it will reduce the workload by around 95%. Leaving a manageable portion for human intervention; limited to difficult cases where the AI can’t arrive at the outcome with sufficient accuracy.


A quicker and more efficient integration tool is of high value. But the most important aspect is the ability to automate the maintenance process.

With 400,000 hotels, there are hundreds of changes happening each day:

  • New hotels are added to the database, while existing ones close temporarily or even permanently
  • Hotel names are changing
  • Descriptions, images and other info are frequently subject to changes

Maintenance can add significant ongoing costs without a proper mapping tool.

Why did we build our own technology?

We could have bought the service from third party providers. However, developing the technology in-house has several advantages:

  1. Cost. Mapping providers are very expensive and consume a lot of resources. Our 0% commission business model works best when coupled with vertical integration.
  2. Flexibility. The mapping tool is unique to LockTrip. It is built to take full advantage of wholesale providers who offer unique pricing in specific Geo-locations. We are able to make quick adjustments whenever needed.
  3. Technology. As a software company, we can offer our powerful technology and tools to travel start-ups who want to build on our LT blockchain. In addition to the newly developed mapping tool, our crypto to fiat exchange-module also falls into this category.
  4. Independence. Maintaining our own tool keeps us independent from third parties — something which we have always aimed for wherever possible.
  5. Future Applications. A powerful mapping tool will become very useful when we launch our decentralized database.

Update on our new Inventory

The delay in the integration of our new inventory has been mostly caused by the development of our mapping tool, which we regard as an investment with immediate returns.

As of now, we have completed the synchronization and mapping of all inventory related elements. Our database currently consists of approximately 400,000 hotels.

We have also concluded the mapping on all operations related to price fetching. This means that we are now able to perform searches among the full 400,000 hotel inventory, receive prices, display rooms and hotels effectively.

The only work left to be done relates to the actions that take place after a booking is initiated. We will update our community once this final task has been completed.

Join us on Telegram

LockTrip is the first marketplace with 0% commission where you can save on average 20% on your hotel and rental bookings compared to anywhere else. Read how to buy LOC tokens here!