# Hotel mapping solutions for metasearch and affiliate sites

> URL: https://hotelmapping.org/en/solutions/hotel-mapping-solutions-for-metasearch-and-affiliate-sites/ · Language: English

Metasearch and affiliate sites show prices from different sellers side by side for the same hotel. A false match directly breaks this comparison. hotelmapping.org matches offers at hotel, room and board level; instead of merging uncertain records automatically, it sends them to human review.

## Problems you face

- **Two different hotels on one card:** If two hotels with similar names a few hundred meters apart are merged by mistake, users see another hotel's price. This error turns into disappointment after the click and a loss of trust.
- **Different rooms in the same price row:** One seller sends a standard room, another a sea-view suite. Without room-level mapping, the two prices are compared as if they were the same product.
- **Board differences are invisible:** If a half board (HB) offer is placed next to an all inclusive (AI) offer, the lowest price becomes misleading. Because sellers write board in different codes and languages, the difference is overlooked.

## Recommended package: Hotel + room + board mapping

Hotel mapping links every seller's hotel to a single HM ID. Room and board type mapping limit the comparison to the same room and the same board type. The result: only truly comparable prices sit side by side.

- [Hotel mapping solutions](https://hotelmapping.org/en/products/hotel-mapping-solutions/)
- [Room mapping solutions](https://hotelmapping.org/en/products/room-mapping-solutions/)
- [Board mapping solutions](https://hotelmapping.org/en/products/board-mapping-solutions/)

## Sample flow

1. **Ingesting seller data:** Each seller's hotel list and static data are received via API, CSV or FTP.
2. **Hotel-level mapping:** Records are linked to HM IDs by evaluating name, address, coordinates, phone and website together. Scores between 0.80 and 0.97 are not merged automatically; they go to human review.
3. **Room and board key:** Each offer is keyed by HM ID, room tag and standard board code; for example HM-1048213-R03 and HB.
4. **Comparison and feedback:** Prices are compared only among offers with the same key. Incorrect matches you report are corrected and added to the model as training data.

## Why is a mapping error expensive in metasearch?

The value of metasearch and affiliate sites lies in correctly showing users the price of the same product at different sellers. Every offer from [OTAs](https://hotelmapping.org/en/glossary/ota/) and other sellers must be linked to a hotel card on the site. If this link is wrong, the comparison is wrong too. When a user clicks a price and sees a different hotel or a different room on the seller's page, the loss of trust is not limited to that offer. For the basic concepts, see the [what is hotel mapping](https://hotelmapping.org/en/guide/what-is-hotel-mapping/) guide.

## Mapping at three levels

| Level | Question | Product |
| --- | --- | --- |
| Hotel | Are these two records the same property? | [Hotel mapping](https://hotelmapping.org/en/products/hotel-mapping-solutions/) |
| Room | Are these two offers the same room type? | [Room mapping](https://hotelmapping.org/en/products/room-mapping-solutions/) |
| Board | Are these two offers the same board type? | [Board mapping](https://hotelmapping.org/en/products/board-mapping-solutions/) |

A fictional example: the names "Kemer Pine Garden" and "Kemer Pine Garden Beach" are similar and their coordinates are close. Even if name similarity is high, if the phone and website differ, the [match score](https://hotelmapping.org/en/glossary/match-score/) may stay below the automatic match threshold. In that case the record is not merged automatically and goes to a specialist's review.

## How is the confidence score used?

Every match comes with a score between 0 and 1. If the score is 0.97 or higher, the match is automatic; records between 0.80 and 0.97 go through a specialist's review; below 0.80 opens a new HM ID. You can also use the score on your side; for example, you can include only high-scoring offers in the comparison table.

When you notice an incorrect match, you report it with the HM ID and supplier code. The record is corrected and the decision is added to the model as training data; errors of the same kind are measured against the test set in later releases.

You can find why room names are so inconsistent in the [what is room mapping](https://hotelmapping.org/en/guide/what-is-room-mapping/) article, and what to ask when evaluating a provider in the [12 questions when choosing a hotel mapping API](https://hotelmapping.org/en/guide/12-questions-hotel-mapping-api/) list.

## Frequently asked questions

### Which is more harmful, a false match or a missed match?

For metasearch, a false match is usually more harmful, because it shows the user an incorrect price. That is why uncertain scores are not merged automatically; a specialist reviews them.

### What is your accuracy rate?

Accuracy is measured with every release against a manually labeled test set. We do not publish figures before the measurement is complete; measured values will appear on the coverage page.

### Is hotel mapping alone enough without room mapping?

It gives you a single card at hotel level, but price rows may compare different rooms. For sites that do price comparison, we recommend room and board mapping as well.

## Related articles

- [What is hotel mapping? How it works: a complete guide](https://hotelmapping.org/en/guide/what-is-hotel-mapping/)
- [What is room mapping? Why are room names so messy?](https://hotelmapping.org/en/guide/what-is-room-mapping/)
- [12 questions to ask when choosing a hotel mapping API](https://hotelmapping.org/en/guide/12-questions-hotel-mapping-api/)
