Applying advanced Marketing Mix Modeling (MMM) methodologies, it was estimated that 6.2% of the total number of visitors from the United States to the Dominican Republic during 2024 were directly driven by digital marketing actions developed by the Ministry of Tourism. This implies a total of 278,913 incremental tourists at a 47x ROI. The model used suggests that the maximum effect of advertising campaigns occurs approximately 50 days after exposure to the ad, particularly in the channels with the highest investment—Google Search and Meta—and that it extends up to 100 days later as a residual effect.

Context: Marketing as a catalyst for tourism development

Developing the tourism industry is a strategic pillar for promoting sustainable economic growth, especially in developing economies and emerging destinations. Academic literature has extensively documented the positive role that investment in tourism promotion plays in generating international demand.

For example, Brida and Schubert (2008) concluded that an increase in promotional spending has a direct and significant effect on the increase in tourists, which in turn stimulates production, employment, and infrastructure investment.

Similarly, Divisekera and Kulendran (2006) demonstrated, based on an analysis of the Australian case, that sustained advertising is a key determinant of international arrivals from markets such as Japan and New Zealand.

Finally, Longwoods International (2021) introduced the concept of the "halo effect", noting that tourism marketing campaigns not only improve a country's image as a travel destination but also positively influence its perception as a place to invest, study, or reside.

In Latin America, the upward trend in international tourism has been clear: 117 million tourists visited the region in 2023, according to data from the World Tourism Organization (UNWTO). However, this dynamic occurs in an environment of increasing competition between destinations. Countries such as Mexico (45 million), Brazil (6.7 million), and Argentina (5.3 million) led the regional ranking in 2023. Mexico also experienced year-over-year growth of 7.4%, while Brazil grew by 12.6%. Argentina, with 5.3 million tourists in the first nine months of the year, achieved its second-highest record in more than a decade.

In parallel, destinations in Central America and the Caribbean—such as the Dominican Republic—are implementing more sophisticated marketing strategies and improving their infrastructure to capture greater market share. The Dominican Republic consolidated its position as the most popular destination in the Caribbean, surpassing 10 million visitors in 2023, while Colombia reached 6.7 million visitors in 2024, representing an 8.5% increase over the previous year. This regional competition is pushing countries to diversify their offerings with ecotourism, cultural and gastronomic tourism, and adventure tourism, responding to an increasingly demanding, informed, and experience-oriented demand.

Challenge: Gap in impact measurement

Despite the strategic importance of tourism marketing, there are few systematic public efforts to measure its real impact using econometric tools and aggregated behavioral data. Some relevant background information helps to measure the challenge:

Oxford Economics, in its evaluation of the Brand USA program, estimated that advertising campaigns contributed to a 1.4% increase in international visitors in 2016 and a 0.6% increase in 2022 (post-pandemic), and reported a return on investment (ROI) of 34x and 11x the value invested.

Google, through a three-month randomized controlled experiment in New Zealand, estimated an 11% lift in visitors between the ad-exposed group versus the control group, and a ROAS (Return on Ad Spend) of 12x.

These cases illustrate both the potential of marketing and the complexity of measuring its real effect. Therefore, having rigorous metrics that allow for the identification of truly incremental visitors is essential for governments when deciding how to efficiently allocate their international promotion budgets.

Measurement strategy: design, methodology and data

Faced with this need, the Dominican Republic's Ministry of Tourism, in collaboration with Bunker, designed and implemented a comprehensive measurement plan based on industry best practices and academic research. This approach combined:

Marketing Mix Models (MMMs), specifically the open-source Robyn framework, developed by Meta, and Google's Meridian.

Geographic experiment design and time series analysis aimed at validating the stability of results under different configurations.

Post-click and post-view digital attribution analysis complemented the estimation with individualized information on campaign exposure.

This measurement system was designed to triangulate results and thus strengthen the causal validity of the conclusions obtained. The focus was on investment in digital media targeting the US market during the period from January to December 2024.

Model design: econometric estimation of advertising impact

The impact of digital media investment on the number of visitors from the United States to the Dominican Republic in 2024 was estimated using a robust econometric approach based on the Marketing Mix Modeling (MMM) methodology.

To do this, the Robyn model was used, an open-source solution developed by Meta that integrates machine learning and classical statistics techniques to quantify the incremental effect of marketing on business variables. Robyn applies a multi-channel frequentist approach, supported by regularized regression models—specifically Ridge and Lasso—which allows for handling typical situations of high multicollinearity and overfitting, common in environments with multiple advertising channels and large data volumes. This approach allows for obtaining parsimonious, interpretable, and stable estimates, even in contexts of high structural complexity.

To ensure the robustness of the model, the dataset was systematically divided into three sets: training, validation, and testing. The results were consistent across all time intervals, achieving an adjusted R² of 0.83 in training, 0.89 in validation, and 0.832 in testing. These values reflect the model's high predictive capacity and good generalization, providing confidence in the statistical robustness of the estimates.

Variables included in the model

In addition to the digital media advertising investment variables (Meta and Google), a battery of key control variables were tested and incorporated to isolate the causal effect of marketing on tourist volume. These variables were selected based on empirical evidence and previous qualitative studies on traveler behavior and include:

Average relative hotel prices in the Dominican Republic versus competing tourist destinations.

Relative prices of flights to the Dominican Republic compared to similar alternatives in the region.

Evolution of digital searches related to terms of interest to potential tourists to the destination.

Calendar of holidays and special events in the United States, including key dates that influence travel flow.

Climate seasonality and exogenous shocks, such as weather events or health conditions, that may affect travel intentions.

The explicit inclusion of these variables aims to control for sources of variability not attributable to marketing, ensuring that the estimated impact of digital advertising campaigns represents only their incremental effect on travel decisions, and not the influence of structural or circumstantial factors. The variables were constructed using various data sources from the Ministry of Tourism, Google Trends, Google Flight Prices, and various weather information sources such as OpenWeatherMap.

Main results

Estimated incremental impact

The Robyn model estimated that digital campaigns targeting the United States market achieved a 47x ROI, generating 6.2% of total visits from that country in 2024. This equates to an additional 278,913 tourists, directly and statistically significantly attributable to digital advertising investment.

This percentage is especially relevant when compared to other documented cases internationally. For example, the Brand USA program generated a 1.4% impact on international visits to the United States in 2019, according to official figures (thebrandusa.com). Another study, conducted on Spanish hotel chains, found that advertising on Meta platforms accounted for 16.8% of web bookings in 2021 (Marketing Insider Review). The estimated result of 6.2% represents a conservative estimate, as it focuses exclusively on short-term effects and does not consider long-term benefits such as brand building or the return of repeat visitors.

Efficiency and economic return

The analysis yielded an average return on advertising spend (ROAS) for the period analyzed of:

56x ROAS on Google Search

53.7x ROAS on Meta (Facebook and Instagram)

48x ROAS on Google Demand Gen

24x ROAS on YouTube Video campaigns

Considering that the average spending per tourist (non-resident foreigner) in the Dominican Republic was US$168 per night for an average of 8 nights, totaling US$1,343 per person, digital marketing showed an average direct return of 47 times for every dollar invested during the period.

This indicator exceeds the metrics of other international programs and confirms the high profitability of the promotional actions carried out during the period. It should be noted that these estimates do not include indirect effects or the multiplier impact on other sectors of the economy (lodging, gastronomy, transportation, commerce).

Temporality of the effect

The model's temporal analysis identified that the peak advertising impact occurs 50 days after exposure to the ad, with a significant effect extending up to 100 days later. This insight is key for campaign planning: it allows for aligning peak advertising investment periods with travelers' effective decision-making windows.

Model robustness: cross-validations and alternative models

In line with best practices recommended by marketing science teams such as those at Google and Meta, various robustness tests were applied to minimize bias and validate the stability of the results:

1. Consistency across different model configurations: Multiple models (8 model families) were trained with variations in training, validation, and testing periods, all maintaining adjusted R² greater than 0.8 on the test datasets.
2. Alternative Bayesian models (Meridian): Twelve configurations were implemented using hierarchical Bayesian models with spatial and temporal components, disaggregating investment and visits by region of origin. Although these models showed better training fits (0.85–0.94), testing performance was more modest (0.5–0.65), confirming the robustness of the baseline approach.
3. Post-click and post-view digital attribution: Behavior was evaluated during a 28-day window prior to tourists' arrival in October 2024. Using anonymized data, the percentage of visitors exposed to digital campaigns was estimated, as an indirect validation of their actual reach. While this information does not establish causality, it supports the model's hypothesis regarding the order of magnitude of the effect.

Next steps: scaling measurement

This study represents a significant step toward more accurate, evidence-based measurement of the impact of digital marketing on international tourism. The result—278,913 additional tourists from the United States in 2024—validates the effectiveness of the campaigns implemented by the Dominican Republic and demonstrates that it is possible to build replicable, scalable, and useful models for public decision-making.

Measuring the impact of tourism marketing not only allows for budget optimization but also ensures that each investment effectively contributes to economic development, job creation, and the country's competitiveness as an international destination.

Therefore, based on the results obtained and the methodological soundness of the approach adopted, the following lines of work are proposed to deepen and scale the model:

• Design and implementation of a controlled geographic experiment that exploits the variability in investment levels across different U.S. regions, with the goal of generating complementary causal evidence by creating synthetic control groups.
• The model will be periodically updated with data from the first quarter of 2025 to monitor the evolving impact, including the potential addition of new channels like TikTok, geared toward younger audiences.
• Application of the model in new source markets, especially those with lower advertising investment and visitor volume, will require adapting the model structure and exploring new predictor variables relevant to those contexts.

References

• Banco Central de la República Dominicana. (s. f.). Sector turismo. Recuperado el 4 de julio de 2025, de https://www.bancentral.gov.do/a/d/2537-sector-turismo
• Brand USA. (2023, 1 de mayo). The return on investment of Brand USA marketing (FY2022): Oxford Economics report. Recuperado el 4 de julio de 2025, de https://www.thebrandusa.com/system/files/FY2022_ROI%20Report_TE-Oxford%205-1-2023.pdf
• Brida, J. G., & Schubert, S. F. (2008). The economic effects of advertising on tourism demand. Disponible en https://ideas.repec.org
• Divisekera, S., & Kulendran, N. (2006). Effects of advertising on tourism demand: An econometric study. Sustain PATA, SAGE Journals. Disponible en https://journals.sagepub.com
• Duan, Q. (2017). Measuring long-term advertising effects in the tourism and hospitality industry (Tesis doctoral). University of South Carolina. Disponible en https://scholarcommons.sc.edu
• International Trade Administration. (s. f.). National Travel and Tourism Office. Recuperado el 4 de julio de 2025, de https://www.trade.gov/national-travel-and-tourism-office
• Jones, N. B., Miles, P., & Beaulieu, T. (2021). The value of social media advertising strategies on tourist behavior: A game-changer for small rural businesses. Journal of Small Business Strategy. https://www.jsbs.org
• Kim, H., Jang, S., & Fu, S. (2020). A model of tourism advertising effects. Tourism Management, ScienceDirect. https://www.sciencedirect.com
• Kulendran, N., & Divisekera, S. (2006). Economic effects of advertising on tourism demand: A case study. Sustain PATA, SAGE Journals. https://journals.sagepub.com
• Longwoods International. (2021). Advertising impact on economic development beyond tourism: “Halo Effect”. Disponible en https://longwoods-intl.com
• Ministerio de Turismo (República Dominicana). (s. f.). Tableros. Recuperado el 4 de julio de 2025, de https://situr.mitur.gob.do/tableros/
• Think with Google. (2019, septiembre). Experiment: How Tourism New Zealand measured the incremental impact of display ads on tourist visits. Recuperado el 4 de julio de 2025, de https://www.thinkwithgoogle.com/intl/en-apac/marketing-strategies/search/experiment-how-tourism-new-zealand-measured-incremental-impact-display-ads-tourist-visits/
• Steinmetz, J. L., Park, S., & Fesenmaier, D. R. (2010). Measuring tourism advertising: The Destination Advertising Response (DAR) model. University of Massachusetts Amherst, ScholarWorks. https://scholarworks.umass.edu