Part A:Cleaning and Analysing Data

1. Research Question

Does price competitiveness affect hotel booking likelihood, and does this effect differ across hotel star ratings?

This is a descriptive analysis. We examine the association between price competitiveness and booking likelihood across different star ratings. The goal is to summarise patterns in the data rather than predict future outcomes or claim causation.

Why descriptive? Our research question asks whether price competitiveness is associated with higher booking rates, not whether it causes them - We do not predict future booking probabilities - We do not use experimental manipulation to infer causality - The analysis focuses on summarising historical patterns using aggregation, cross-tabulation, and visualisation.

1.1 Hypotheses

Hypothesis	Statement
H1	Hotels with higher price competitiveness (cheaper than same-star average) have higher booking likelihood
H2	The effect of price competitiveness on booking likelihood differs across hotel star ratings

1.2 Definition of Price Competitiveness

In this analysis, price competitiveness is defined as:

Price Ratio=Current Price/Average Price of Same-Star Hotels

• Price Ratio < 1: Cheaper than same-star average (competitive)

• Price Ratio = 1: At same-star average price

• Price Ratio > 1: More expensive than same-star average (uncompetitive)

This definition is more meaningful than comparing to a hotel’s own historical price because customers typically compare hotels within the same category.

rm(list = ls())
library(tidyverse)

── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr     1.2.1     ✔ readr     2.2.0
✔ forcats   1.0.1     ✔ stringr   1.6.0
✔ ggplot2   4.0.3     ✔ tibble    3.3.1
✔ lubridate 1.9.5     ✔ tidyr     1.3.2
✔ purrr     1.2.2     
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

library(scales)

Attaching package: 'scales'

The following object is masked from 'package:purrr':

    discard

The following object is masked from 'package:readr':

    col_factor

options(digits = 4, scipen = 999)

load the initial dataset:

# Load data
df_raw <- read_csv("data/expedia.csv")

Rows: 158269 Columns: 54
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr  (31): visitor_hist_starrating, visitor_hist_adr_usd, prop_review_score,...
dbl  (22): srch_id, site_id, visitor_location_country_id, prop_country_id, p...
date  (1): date_time

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

#Show the structure of the dataset
glimpse(df_raw)

Rows: 158,269
Columns: 54
$ srch_id                     <dbl> 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61…
$ date_time                   <date> 2013-06-22, 2013-06-22, 2013-06-22, 2013-…
$ site_id                     <dbl> 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, …
$ visitor_location_country_id <dbl> 219, 219, 219, 219, 219, 219, 219, 219, 21…
$ visitor_hist_starrating     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ visitor_hist_adr_usd        <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ prop_country_id             <dbl> 219, 219, 219, 219, 219, 219, 219, 219, 21…
$ prop_id                     <dbl> 16517, 31767, 39799, 68629, 73060, 78003, …
$ prop_starrating             <dbl> 3, 2, 2, 2, 3, 3, 3, 3, 3, 3, 2, 3, 2, 0, …
$ prop_review_score           <chr> "4.5", "3.5", "3.0", "2.0", "3.0", "4.5", …
$ prop_brand_bool             <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ prop_location_score1        <dbl> 2.48, 1.79, 0.00, 2.48, 2.48, 2.64, 1.95, …
$ prop_location_score2        <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ prop_log_historical_price   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …
$ position                    <dbl> 7, 4, 3, 13, 12, 9, 10, 8, 6, 1, 14, 2, 22…
$ price_usd                   <dbl> 127, 55, 47, 65, 83, 135, 139, 129, 71, 68…
$ promotion_flag              <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …
$ srch_destination_id         <dbl> 18040, 18040, 18040, 18040, 18040, 18040, …
$ srch_length_of_stay         <dbl> 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 1, 1, …
$ srch_booking_window         <dbl> 359, 359, 359, 359, 359, 359, 359, 359, 35…
$ srch_adults_count           <dbl> 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, …
$ srch_children_count         <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, …
$ srch_room_count             <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ srch_saturday_night_bool    <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, …
$ srch_query_affinity_score   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ orig_destination_distance   <chr> "1955.51", "1955.76", "1959.38", "1955.92"…
$ random_bool                 <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ comp1_rate                  <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp1_inv                   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp1_rate_percent_diff     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp2_rate                  <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp2_inv                   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp2_rate_percent_diff     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp3_rate                  <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp3_inv                   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp3_rate_percent_diff     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp4_rate                  <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp4_inv                   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp4_rate_percent_diff     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp5_rate                  <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp5_inv                   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp5_rate_percent_diff     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp6_rate                  <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp6_inv                   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp6_rate_percent_diff     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp7_rate                  <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp7_inv                   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp7_rate_percent_diff     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp8_rate                  <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp8_inv                   <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ comp8_rate_percent_diff     <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ click_bool                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, …
$ gross_bookings_usd          <chr> "NULL", "NULL", "NULL", "NULL", "NULL", "N…
$ booking_bool                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …

#Show the number of rows and columns
dim(df_raw)

[1] 158269     54

# Check missing and invalid values in key variables
data_quality_check <- df_raw %>% 
  summarise(
    total_rows = n(),
    missing_price = sum(is.na(price_usd)),
    invalid_price = sum(price_usd <= 0, na.rm = TRUE),
    missing_star = sum(is.na(prop_starrating)),
    invalid_star = sum(prop_starrating <= 0, na.rm = TRUE),
    missing_position = sum(is.na(position)),
    missing_booking = sum(is.na(booking_bool))
  )

print(data_quality_check)

# A tibble: 1 × 7
  total_rows missing_price invalid_price missing_star invalid_star
       <int>         <int>         <int>        <int>        <int>
1     158269             0             0            0         5836
# ℹ 2 more variables: missing_position <int>, missing_booking <int>

2. Data Cleaning

df_clean <- df_raw %>%
  mutate(
    # Dependent variable
    booking = booking_bool,
    
    # Handle missing values
    visitor_hist_starrating = ifelse(is.na(visitor_hist_starrating), 0, visitor_hist_starrating),
    visitor_hist_adr_usd = ifelse(is.na(visitor_hist_adr_usd), 0, visitor_hist_adr_usd),
    prop_review_score = ifelse(is.na(prop_review_score), 0, prop_review_score),
    prop_location_score1 = ifelse(is.na(prop_location_score1), 0, prop_location_score1)
  ) %>%
  filter(prop_starrating > 0) %>%
  filter(!is.na(position)) %>%
  filter(!is.na(price_usd))

# Cleaning summary
cleaning_summary <- data.frame(
  step = c("Original data", "Cleaned data"),
  n_rows = c(nrow(df_raw), nrow(df_clean))
)
print(cleaning_summary)

           step n_rows
1 Original data 158269
2  Cleaned data 152433

Edit Attributes (F4)

3. Descriptive Analysis

# Data Analysis - Create price competitiveness variable
df_clean <- df_clean %>%
  group_by(prop_starrating) %>%
  mutate(
    avg_price_by_star = mean(price_usd, na.rm = TRUE),
    price_ratio = price_usd / avg_price_by_star
  ) %>%
  ungroup() %>%
  filter(price_ratio > 0.3 & price_ratio < 3.0)

# Create star categories and price groups
df_clean <- df_clean %>%
  mutate(
    star_category = case_when(
      prop_starrating <= 2 ~ "Budget (1-2★)",
      prop_starrating == 3 ~ "Mid (3★)",
      prop_starrating >= 4 ~ "Luxury (4-5★)",
      TRUE ~ "Unknown"
    ),
    price_comp_group = case_when(
      price_ratio < 0.7 ~ "Very Competitive",
      price_ratio >= 0.7 & price_ratio < 0.9 ~ "Competitive",
      price_ratio >= 0.9 & price_ratio < 1.1 ~ "At Market Price",
      price_ratio >= 1.1 & price_ratio < 1.3 ~ "Uncompetitive",
      price_ratio >= 1.3 ~ "Very Uncompetitive",
      TRUE ~ "Unknown"
    )
  )

3.1 Booking Rate by Star Ratingating

df_clean %>%
  group_by(star_category) %>%
  summarise(
    Count = n(),
    Booking_Rate = paste0(round(mean(booking) * 100, 1), "%"),
    Avg_Price_Ratio = round(mean(price_ratio, na.rm = TRUE), 2)
  ) %>%
  knitr::kable(caption = "Table 1: Booking Rate by Hotel Star Rating")

Table 1: Booking Rate by Hotel Star Rating

star_category	Count	Booking_Rate	Avg_Price_Ratio
Budget (1-2★)	41800	2.5%	0.91
Luxury (4-5★)	40116	3.2%	0.98
Mid (3★)	68969	3%	0.97

Key observation: Luxury hotels have the highest booking rate (3.2%), while Budget hotels have the lowest (2.5%). All star categories have average price ratios below 1, indicating that Expedia generally displays competitive prices.

3.2 Booking Rate by Price Competitiveness

df_clean %>%
  group_by(price_comp_group) %>%
  summarise(
    Count = n(),
    Booking_Rate = paste0(round(mean(booking) * 100, 1), "%"),
    Avg_Star_Rating = round(mean(prop_starrating, na.rm = TRUE), 1)
  ) %>%
  arrange(desc(Booking_Rate)) %>%
  knitr::kable(caption = "Table 2: Booking Rate by Price Competitiveness")

Table 2: Booking Rate by Price Competitiveness

price_comp_group	Count	Booking_Rate	Avg_Star_Rating
Very Competitive	44024	3.9%	3.0
Competitive	37775	3.1%	3.0
At Market Price	26913	2.8%	3.0
Uncompetitive	16372	2.1%	3.1
Very Uncompetitive	25801	1.5%	3.1

Key observation: Very Competitive hotels have a booking rate of 3.9%, which is 2.6 times higher than Very Uncompetitive hotels (1.5%). This strongly supports H1.

3.3 Cross-Tabulation: Star Rating × Price Competitiveness

cross_table <- df_clean %>%
  group_by(star_category, price_comp_group) %>%
  summarise(
    Booking_Rate = round(mean(booking) * 100, 1),
    .groups = "drop"
  ) %>%
  pivot_wider(
    id_cols = star_category,
    names_from = price_comp_group,
    values_from = Booking_Rate,
    values_fill = NA
  )

knitr::kable(cross_table, caption = "Table 3: Booking Rate (%) by Star Rating and Price Competitiveness")

Table 3: Booking Rate (%) by Star Rating and Price Competitiveness

star_category	At Market Price	Competitive	Uncompetitive	Very Competitive	Very Uncompetitive
Budget (1-2★)	2.7	2.5	2.1	2.8	1.7
Luxury (4-5★)	2.9	3.4	2.3	4.9	1.3
Mid (3★)	2.8	3.2	1.9	4.3	1.6

Key observation: Within each star category, booking rate increases as price competitiveness improves. However, the magnitude varies by star rating.

3.4 Price Sensitivity Analysis

sensitivity_table <- df_clean %>%
  group_by(star_category) %>%
  summarise(
    Booking_Rate_Cheaper = round(mean(booking[price_ratio < 0.9]) * 100, 1),
    Booking_Rate_Expensive = round(mean(booking[price_ratio > 1.1]) * 100, 1),
    Difference = Booking_Rate_Cheaper - Booking_Rate_Expensive,
    .groups = "drop"
  )

knitr::kable(sensitivity_table, caption = "Table 4: Price Sensitivity by Star Rating")

Table 4: Price Sensitivity by Star Rating

star_category	Booking_Rate_Cheaper	Booking_Rate_Expensive	Difference
Budget (1-2★)	2.6	1.8	0.8
Luxury (4-5★)	4.3	1.7	2.6
Mid (3★)	3.7	1.8	1.9

Key observation: Price sensitivity ranking: Luxury (+2.6 pp) > Mid (+1.9 pp) > Budget (+0.8 pp). This contradicts the intuition that budget travelers are most price-sensitive.

4. Visualization

4.1 Price Competitiveness vs Booking Rate by Star Rating

#Price Competitiveness vs Booking Rate by Star Rating
plot1_data <- df_clean %>%
  filter(star_category != "Unknown") %>%
  mutate(price_bin = cut(
    price_ratio,
    breaks = seq(0.4, 2.2, 0.2),
    labels = c("0.4–0.6", "0.6–0.8", "0.8–1.0",
               "1.0–1.2", "1.2–1.4", "1.4–1.6",
               "1.6–1.8", "1.8–2.0", "2.0–2.2"),
    include.lowest = TRUE
  )) %>%
  filter(!is.na(price_bin)) %>%
  group_by(star_category, price_bin) %>%
  summarise(
    Booking_Rate = mean(booking) * 100,
    n = n(),
    .groups = "drop"
  ) %>%
  filter(n > 30)

ggplot(plot1_data,
       aes(x = price_bin, y = Booking_Rate,
           color = star_category, group = star_category)) +
  geom_line(linewidth = 1.2) +
  geom_point(size = 2.5) +
  geom_hline(
    yintercept = mean(df_clean$booking) * 100,
    linetype = "dashed",
    color = "gray50"
  ) +
  labs(
    title = "Price Competitiveness vs Booking Rate by Hotel Star Rating",
    subtitle = "Price ratio < 1 indicates cheaper than same-star/destination average",
    x = "Price Ratio (Current Price / Benchmark Average)",
    y = "Booking Rate (%)",
    color = "Star Category"
  ) +
  theme_minimal(base_size = 12) +
  theme(
    axis.text.x = element_text(angle = 45, hjust = 1),
    legend.position = "bottom",
    plot.title = element_text(face = "bold", size = 14, hjust = 0.5)
  )

Figure 1 shows that for all star categories, booking rates decrease as prices become less competitive. The effect is strongest for Luxury hotels.

4.2 Boxplot: Booked vs Not Booked

ggplot(df_clean, aes(x = ifelse(booking == 1, "Booked", "Not Booked"), 
                     y = price_ratio, fill = star_category)) +
  geom_boxplot(alpha = 0.7, outlier.size = 0.5) +
  geom_hline(yintercept = 1, linetype = "dashed", color = "darkred", size = 1) +
  labs(
    title = "Price Competitiveness Distribution: Booked vs Not Booked",
    subtitle = "Red dashed line = same-star average price (ratio = 1)",
    x = "Booking Outcome",
    y = "Price Ratio (Lower = More Competitive)",
    fill = "Star Category"
  ) +
  theme_minimal() +
  theme(legend.position = "bottom",
        plot.title = element_text(face = "bold", size = 14, hjust = 0.5))

Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
ℹ Please use `linewidth` instead.

Figure 2 shows that booked hotels have significantly lower price ratios (more competitive) than non-booked hotels.

4.3 Price Sensitivity by Star Rating

sensitivity_table %>%
  filter(star_category != "Unknown") %>%
  ggplot(aes(x = star_category, y = Difference, fill = star_category)) +
  geom_bar(stat = "identity", alpha = 0.8) +
  geom_text(aes(label = paste0(Difference, " pp")), vjust = -0.5, size = 4) +
  labs(
    title = "Price Sensitivity by Hotel Star Rating",
    subtitle = "Difference = Booking rate (cheaper) - Booking rate (expensive)",
    x = "Star Category",
    y = "Price Sensitivity (percentage points)"
  ) +
  theme_minimal() +
  theme(legend.position = "none",
        plot.title = element_text(face = "bold", size = 14, hjust = 0.5))

Figure 3 clearly shows that Luxury hotels have the highest price sensitivity, followed by Mid hotels, with Budget hotels being the least price-sensitive.

5. Conclusion

5.1 Summary of Findings

Finding	Evidence	Supports
More competitive pricing increases booking likelihood	Table 2, Figure 1, Figure 2	H1: Yes
Price sensitivity differs across star ratings	Table 4, Figure 3	H2: Yes
Luxury hotels are most price-sensitive	Table 4, Figure 3	Surprising finding
Budget hotels are least price-sensitive	Table 4, Figure 3	Surprising finding

5.2 Answer to Research Questions

Q1: Does price competitiveness affect hotel booking likelihood?

Yes. Hotels with more competitive pricing (lower price ratio) have significantly higher booking probabilities. Very Competitive hotels have a 3.9% booking rate, compared to only 1.5% for very uncompetitive hotels — a 2.6x difference.

Q2: Does this effect differ across hotel star ratings?

Yes. The price sensitivity ranking is:

Rank	Star Category	Price Sensitivity
1	Luxury (4-5★)	+2.6 pp
2	Mid (3★)	+1.9 pp
3	Budget (1-2★)	+0.8 pp

5.3 Possible Explanation for Counter-Intuitive Finding

Why are Luxury hotels most price-sensitive?

Factor	Explanation
Absolute savings	A 10% discount on a 500 room saves50, while the same discount on a 100 room saves only10
More alternatives	Luxury travelers have more comparable options to choose from
Cross-platform comparison	Higher-end travelers are more likely to compare prices across platforms

5.4 Business Recommendations for Expedia

1. Luxury hotels: Price promotions are most effective. Prioritize displaying luxury hotels with competitive prices.

2. Mid hotels: Price competitiveness is key for conversion. Maintain competitive pricing for this segment.

3. Budget hotels: Price beyond a certain point has diminishing returns. Focus on other factors like reviews and location.

6. Limitations

1. Price competitiveness definition: Uses same-star average as benchmark; ideal measure would include competitor prices

2. Correlation vs causation: Observational data cannot prove causality

3. Sample period: Data may not reflect current post-pandemic travel behavior

Part B: Analytics Team Memo

To: Analytics Team
From: Group 4
Subject: Expedia Price Competitiveness Analysis – Preliminary Findings
Date: May 28,2026

1. Research Question

• Does price competitiveness affect hotel booking likelihood, and does this effect differ across hotel star ratings?

• This question matters because understanding price sensitivity by segment can help Expedia optimize ranking algorithms and promotion strategies.

2. Data & Approach

• Dataset: Expedia hotel search impressions (June 2013, 158k rows, 6,286 unique searches).

• Key variables: booking_bool (outcome), price_usd, prop_starrating, position.

• Price competitiveness definition: price_ratio = current_price / same-star_avg_price. Values < 1 indicate competitive pricing.

• Method: Descriptive analysis using cross-tabulation and visualization. No causal claims.

3. Preliminary Findings

• Finding 1 (H1 supported): Price competitiveness strongly predicts booking rates. Very Competitive hotels have 3.9% booking rate vs 1.5% for very uncompetitive hotels (2.6x difference).

• Finding 2 (H2 supported): Price sensitivity varies significantly by star rating. Luxury hotels show the largest difference (+2.6 percentage points between cheap and expensive prices), followed by Mid (+1.9 pp) and Budget (+0.8 pp).

• Surprising finding: Luxury hotels are most price-sensitive, contradicting the intuition that budget travelers care most about price. Possible explanation: absolute dollar savings are larger for luxury hotels.

4. Assumptions & Limitations

• Assumption: Same-star average price is an appropriate benchmark for price competitiveness.

• Limitation 1: Observational data → correlation, not causation.

• Limitation 2: 2013 data may not reflect current travel behavior.

• Limitation 3: Visitor history variables have >90% missing rates and were excluded.

5. Open Questions for Feedback

• Should we refine the price competitiveness benchmark to include destination as well as star rating?

• Would the team recommend adding a robustness check using the random_bool experiment variable?

• Should we explore interaction effects with review scores or location scores?

6. Next Steps

• Validate findings with more recent data if available.

• Test whether results hold after controlling for search position and promotion flags.

• Prepare executive brief for management with actionable recommendations.

Part C: Corporate Brief

To: Management
From: Group 4
Subject: Price Competitiveness and Booking Performance
Date: May 28,2026

Executive Summary

Does offering competitive prices help Expedia increase hotel bookings? Our analysis of June 2013 search data shows a clear answer: yes, and the effect varies significantly by hotel star rating.

Contrary to common belief, luxury hotels (4-5★) are the most price-sensitive segment, not budget hotels.

---

Key Insights

Star Category	Booking Rate (Cheap)	Booking Rate (Expensive)	Difference
Luxury (4-5★)	4.3%	1.7%	+2.6 pp
Mid (3★)	3.7%	1.8%	+1.9 pp
Budget (1-2★)	2.6%	1.8%	+0.8 pp

Overall: Hotels priced below their star category average are 2.6 times more likely to be booked than those priced above average.

Why luxury hotels are most price-sensitive: A 10% discount on a $500 room saves $50, while the same discount on a $100 room saves only $10. Luxury travelers also have more comparable options and are more likely to cross-compare prices.

---

Business Implications

• For luxury hotels: Price is a powerful competitive weapon. Competitive pricing drives significant conversion gains. Expedia should prioritise displaying luxury hotels with attractive prices.

• For mid-range hotels: Maintain price competitiveness. This segment responds well to attractive pricing. Price promotions are likely effective for this group.

• For budget hotels: Price competition has diminishing returns. Focus on highlighting reviews, location scores, and other value-adds rather than competing solely on price.

---

Recommended Actions

1. Prioritise price competitiveness signals in ranking algorithms for luxury and mid-range hotel searches.

2. Test “below average for this star” badges on search result pages to highlight competitive prices.

3. For budget hotel searches, consider giving more weight to non-price factors (review scores, location scores) in ranking decisions.

---

Caveat

These findings are based on 2013 data. We recommend validating with recent data before implementing major changes to ranking or promotion strategies.