Menu Close

Data Disclaimer

1. Financial data provided by Hotelysis is an output of artificial intelligence.

There are various public statistics and private data related to the Korean lodging industry, in and out of the country, and Hotelysis collects most of them. Please check the Data Sources page for the complete list of data sources collected by Hotelysis. However, Hotelysis does not just pass them over to the users. It is mainly because we think they can cause confusions, contrary to our goal to provide complete visibility over the Korean lodging market. In detail, there are two reaons:

First, there exists widespread inconsistency among public statistics. Public statistics related to the lodging industry in Korea are compiled in accordance with regulatory classification of lodging business segments and property types, and statistics for each category is compiled independently by the relevant authority. Please check Lodging Property Types by Regulation page for the classifcation details. On the other hand, due to the difference in the focus of each authority's work scope according to the relevant laws, there are also significant differences in the way each statistics is compiled and the scope of the information contained. In other words, you will get different answers for the same question, depending upon who you ask.

  • Different standards in compiling: results vary depending on whether the statistics are about lodging business or loding property because the regulatory lodging business segments and property types do not match. A lodging property can be used for businesses other than the lodging business, a lodging business does not have to be run only at a lodgin property, and the type of business run at a property can change at any time. This implies that the results of statistics vary depending on whether the management target is a business or a property, and whether the property is managed by parcel or by building. For example, if a single company operates a resorts complex consisting of multiple buildings in multiple parcels, it may be viewed as one lodging property or multiple lodging properties depending on which authority is looking at it.
  • Different scopes in compiling: although there are discrepancies between statistical data due to different compiling standards, statistics on the supply of lodging properties in Korea are relatively abundant. On the other hand, there are limited statistics that provide visibility over operating performance such as demand, prices, revenues, and costs. First of all, in the case of revenues, the Korean Hotel Association publishes the Operating Statistics of Hotel Industry every year, and it provides somewhat detailed information about demand, prices, and revenues. However, the sample of this statistics covers only about 50% of all hotels, which is 4% of all lodging properties. The Economic Census published by Statistics Korea every five years contains high-level data about revenues and costs, but the results for the same sample as the Operating Statistics of Hotel Industry are different.

Second, contamination of survey-based statistics is inevitable. In most cases, supply statistics are not survey-based, but there still is some contamination due to mix-up of address systems. Operating performance statistics are based entirely on voluntary surveys from individual businesses, and there is no real way to find and correct errors if respondents provide incorrect data. Of course, techniques for verifying and correcting statistical significance are widely used, but there are clear limitations in association with the volatile and irregular cash flow data of the lodging industry.

  • Contamination by mistake: in most cases, the contamination is caused by mistakes. For example, some of the data are submitted missing, some irrelevant data are submitted together, the entries are mixed up, or there are mistakes in the unit notation of the numbers. In this case, it is relatively simple to improve the accuracy: eleminate the additional cost burden to improve accuracy, and provide incentives to improve the accuracy. However, since the accuracy issue is not only for the lodging industry statistics, there could be an issue whether it is realistic for the government to take the steering wheel in doing so.
  • Contamination by intention: although the proportion is relatively small, there are cases where false data are submitted to contaminate statistical data. In this case, the ripple effect is large compared to the small proportion. It is because such small defects can destroy the credibility of the entire statistical data. Although there are supposed to be complex reasons behind it, we think the main purpose is to control the competitive environment through asymmetry of market information. There is virtually no way to solve this problem because there is no way to identify the false data, and it cannot be punished even if verified as they are voluntary survey-based statistics.

The advantage of cleaning up the data through artificial intelligence is that validated data are provided in a consistent format. The artificial intelligence algorithms developed by Hotelysis match various public statistics and private data, whether within the country and overseas, related to the Korean lodging industry, according to preset standards. Then, the significance of each is verified to correct the data points determined as errors, and to estimate or forecast the missing data points according to the preset standards. The algorithms are applied to the individual property level financial data as well as the aggregate market level data, and the competitive market level data, aggregated again after verifying the individual property level data, are provided to the users.

However, there is still a limitation that the accuracy of the data cleaned up by artificial intelligence cannot be fully guaranteed. Hotelysis tests the error rate by comparing the output of artificial intelligence with the original market level data. In this process, the output data are verified by MAPE and SMAPE, with the identical statistical samples of the original market level data. In addition, the error rates of individual property level data are tested with the actual data of the property, if applicable, but the availability of individual property level data is limited compared to market level data. Hotelysis provides only data that passes the verification tests, yet cannot guarantee the accuracy of the data provided.

2. The data provided at HBI Dashboard is based upon the financial data at the individual property level.

The reason that Hotelysis breaks the market level data down to the individual property level, instead of providing the original market level data, is to reflect the unstable supply status and market dynamics of the Korean lodging market more accurately and promptly in the data.

First of all, the unstable supply status means that the pace of new openings is high, while the pace of closings is also high due to the short life cycle. In other words, there is a high degree of uncertainty about when a newly opened property will close down. It also implies that the competitive environment is dynamically changing in Korea, and the delivery of accurate market data on time is critical. The unstable supply status in Korean lodging market can be captured by two indicators:

  • Average life cycle of a lodging business: in general, the economic life of a lodging property is 40 years, which is the depreciable life, and the closer the average life cycle of a lodging business is to the economic life, the more stable the market is. However, the average life cycle of a lodging business in Korea is only 23 years, which is 58% of the economic life, while it is 36 years in the United States, which is 90%.
  • Annual average growth of rooms: from 2005 to 2019, the number of rooms of lodging properties in Korea increased by 1.0% per annum with a standard deviation of 1.1%. During the same period, it increased byof 1.3% per annum with a standard deviation of 1.0% in the United States. In terms of the overall rate of increase or decrease, Korea's supply status seems more stable, but behind it, there are relatively high opening and closing rates.

On the other hand, the volatility of demand in Korean lodging market is also estimated to be high, but it is difficult to be measured accurately. The reason that demand is estimated to be highly volatile is that the proportion of external demand is high, and the seasonal leisure demand accounts for most of domestic demand. It is difficult to be measured accurately because the compiling standards for demand-related statistics differ from supply-related statistics, and demand-related statistics are not quite in shape in terms of significance and timeliness.

  • Gap between regulatory types and market types: the Korea National Tourism Survey and the International Visitor Survey, published by Korea Culture & Tourism Institute every year, provide data to determine domestic and foreign demand by lodging property type. However, since they use a market types rather than regulatory types, it includes pensions or guesthouses for which supply statistics do not exist. Unless individual lodging properties use the regulatory types in their names, it is not feasible to use the regulatory types in the survey for travelers. This means that demand statistics cannot be linked to supply statistics.
  • Limited financial data for lodging businesses: although limited to 4% of all lodging properties, the only detailed revenue-related data available for the hotel industry is released approximately one to one and a half years after the end of the year. The Economic Census data is also be released a year and a half later, but unfortunately, the subject years so far have been very unusual. 2010 was just after the global financial crisis, 2015 was a time when the lodging industry was hit hard by MERS and 2020 by COVID-19. In other words, we only have revenues and costs data deviating from normal conditions.

The advantage of financial data broken down to the individual property level is that it allows to handle discrepancies and changes of samples in various public statistics effectively. For this purpose, Hotelysis has developed its own classification system by further segmenting and clustering the regulatory types in accordance with market conditions, to maintain the utilization of available statistical data and the accurate capture of the market conditions. Please check the Lodging Property Types by Hotelysis page for the details. In addition, the algorithms developed by Hotelysis is based on the correlation between the individual items of the base data and key economic indicators, enabling independent estimating or forecasting. This means that financial data with significance is provided in advance and confirmed later when relevant statistics are published.

However, the financial data for individual properties are not provided as is, and users will receive data aggregated into competitive market level. The biggest reason is the legal restriction in association with privacy. However, this limitation is complemented by allowing users to set up their own competitive set of properties to generate datasets. For example, the users can select and organize a competitive set of properties directly competiting with the subject property, extract the aggregated data for them, and comparatively analyze the financial performance of the subject property. In this case, in principle, the financial performance data for the subject property shall be provided by the user. A legal review for providing the financial data for individual property is currently underway, and we will decide whether to revise or maintain our policies as soon as we get the opinions.

3. Financial data provided by Hotelysis is in compliance with accounting standards.

There are two sets of accounting standards used by Hotelysis. One is the Uniform System of Accounts for the Lodging Industry (USALI), which is customized for lodging industry, and the other is tje International Financial Reporting Standards (IFRS), which is used generally across all industries. We use them based upon following criteria:

  • USALI: We apply USALI for operating performance data, which is presented in the form of an income statement. Although USALI is not a regulatory format for filing purposes, we think it is more effective in analyzing productivity and efficiency of operations for lodging properties.
  • IFRS: We apply IFRS for financial position data, which is presented in the form of a balance sheet. However, if the actual data obtained and utilized by Hotelysis is created in accordance with GAAP, it is not converted separately, so the user needs to be careful in interpreting the book value data.

The financial data of the lodging property, even for the same operating performance, can vary significantly depending on the ownership and operational structure. For example, if a lodging property is operated under a lease structure, the depreciation expense may not be included in the lessee’s financial statements, while it will be included in case of an owner-operator structure, incurring a huge difference in profit and loss while the operating performance is the same. In principle, the data provided by Hotelysis excludes the variables of ownership and operation structure, and provides data attributed to the assets. Therefore, all data are based on the operation by the ownership, and if the original data obtained and utilized by Hotelysis is under a lease structure, it is converted back to the owner-operator model.

If there is any further questions regarding data provided by Hotelysis, contact us at data@hotelysis.com. Please note that we cannot provide answers in relation to the artificial intelligence algorithms developed by Hotelysis.