Connecting a USD 22M Promotional Investment to USD 1.8B in Economic Impact: A DMO Attribution Model

Back to Perspectives

There is a distinction in destination marketing that most of the sector avoids making. Destination management organizations regularly report on how the tourism industry performed: arrivals, room nights, hotel occupancy, average daily spend. These are legitimate and important measures of industry health. They are not measures of what the DMO produced.

Industry performance data tells you how the destination did. DMO attribution tells you what the organization's promotional investment contributed to that outcome. These are different questions. The first is answered with statistics from immigration services, hotel chains, and national tourism authorities. The second requires building a system that connects a dollar of promotional spend to a dollar of economic activity in the destination economy. Almost no DMOs build that system.

At PROMTUR Panama, we built it. A USD 22M annual promotional investment was connected to USD 1.8B in economic impact. This article explains the methodology, the data infrastructure required, where the model had gaps, and why those gaps are not a reason to avoid building the capability.

Why the Distinction Matters

When a DMO reports that arrivals grew 13% year-on-year, that number captures the combined effect of every factor influencing demand: airline capacity additions, regional economic conditions, exchange rate movements, competitor destination events, word-of-mouth, and the DMO's promotional activity. The DMO did not produce all of that growth. In most markets it produced a fraction of it.

The problem is that the sector has developed a habit of presenting industry performance as DMO performance. Boards are shown arrival charts. Finance ministries are shown room night data. The implicit claim is that the organization's budget drove the outcomes being reported. That claim is rarely supported by evidence because the attribution work required to support it has not been done.

This matters for three reasons.

• Budget justification. A DMO that cannot separate its contribution from background industry performance is vulnerable in every budget cycle. Finance ministers and development banks will eventually ask: what would have happened without you? Without attribution data, there is no answer. There are documented cases across the sector where DMO funding was cut and tourism performance declined. The problem is that those cases only prove the value of promotion after the damage is done. A destination should not have to lose its marketing budget, watch arrivals fall, and rebuild from a weakened position in order to demonstrate that its DMO was producing results. Attribution capability makes the case before the cut, not after it.
• Strategic decision-making. If you do not know which promotional investments are producing economic outcomes, you cannot allocate budget rationally. You are managing by intuition dressed as strategy.
• Sector credibility. Tourism competes for public investment against sectors that can demonstrate economic return. Infrastructure, manufacturing, and technology investment all come with impact assessments. Tourism rarely does. Attribution capability changes that conversation.

The PROMTUR Attribution Model

The PROMTUR promotional program operated across four streams: direct marketing (paid media, social media, PR and earned media), strategic alliances (airlines, online travel agencies, and bed banks), travel trade development, and MICE promotion. Each stream was measured using a different methodology, reflecting the different data available at each stage of the promotional funnel.

The model was not built to claim full credit for Panama's tourism performance. It was built to isolate the incremental economic contribution of PROMTUR's promotional investment, separate from the broader factors driving arrivals. The total USD 22M promotional investment is reported here as a combined figure. Stream-level budgets are not disclosed.

Direct Marketing: Attribution via the Arrival Chain

Direct marketing covered paid media campaigns across 10 international source markets. Channels included YouTube, connected television, programmatic display, social media, and paid search. Attribution required connecting a marketing exposure to an arrival to a set of economic behaviors in the destination. That chain has four links.

Applying this chain to attributed US-market arrivals produced a direct marketing return ratio of approximately 21.8x on the measurable segment alone. The full return across all source markets was higher but could not be directly attributed with the same evidentiary standard.

Strategic Alliances: Purchase-Stage Attribution

The alliance program operated across airlines, online travel agencies, and bed banks. Partners included Copa Airlines, Expedia, and a range of additional airline and OTA partners. Bed banks, which aggregate hotel inventory and distribute it through wholesale channels, were included because they represent a significant conversion pathway for international leisure travelers that sits outside the direct booking funnel.

Attribution was tracked at the point of purchase: a completed booking of a Panama itinerary was the measurable outcome. Purchase-stage attribution is more straightforward than exposure-to-arrival attribution because the booking event is discrete, timestamped, and confirmed by the partner. It does not require arrival data matching. The limitation is that bookings are not the same as arrivals, and cancellations are not always captured in the impact calculation.

Partner-specific figures and deal structures are confidential and not disclosed here. The alliance program generated the largest single component of the total economic impact figure.

MICE Promotion: Event Impact Calculator Methodology

MICE attribution used the Destinations International Event Impact Calculator (EIC), the recognized standard methodology for meetings industry impact measurement. Event organizers reported participant counts, event duration, and spending profiles. These inputs were applied through the EIC to produce per-event economic impact estimates.

MICE is the most defensible stream to measure because the event is a contained, documented economic activity. Participant counts are verifiable. Duration is known. Spending profiles follow established research benchmarks. In the first half of 2024, MICE participation grew 22% year-on-year, from 19,000 to 23,000 participants.

Travel Trade: The Unmeasured Stream

Travel trade promotion was not included in the economic impact attribution model. This is a deliberate disclosure, not an accidental omission.

The structural problem with travel trade measurement is well known in the sector. Wholesalers, tour operators, and travel agencies are generally unwilling to report booking volumes, conversion rates, or revenue data because doing so reveals commercially sensitive information about margins, client relationships, and competitive positioning. Attempts to develop a pay-per-performance reporting model with travel trade partners were paused due to the level of commercial disclosure required.

The result is that a meaningful portion of the promotional investment produced real economic outcomes that remain untracked and unclaimed in the attribution model. The USD 1.8B figure does not include travel trade impact. The true return on the full promotional program was larger than reported.

The decision to exclude rather than estimate was deliberate. An unverifiable estimate in the attribution model would have undermined the credibility of every other figure in it.

The Data Infrastructure Required

Building this attribution capability required assembling seven data sources into a single operational system. No single source produced the model. The value came from connecting them.

• Flight booking and forward-looking demand data: seat capacity and booking trends across 10 source markets, approximately 70% air coverage, sourced via ForwardKeys
• Hotel performance data: reporting hotels covering the majority of Panama City and interior market capacity, sourced via STR
• Vacation rental data: real-time occupancy and revenue tracking across major short-term rental platforms, sourced via Lighthouse
• International arrival data: monthly visitor counts by nationality from Panama's National Migration Service and INEC
• Average daily spend data: per-tourist expenditure benchmarks from the Autoridad de Turismo de Panama (ATP)
• Arrival attribution data: marketing exposure to arrival matching for US carrier travelers
• Alliance and MICE partner reporting: conversion and participant data from partner systems and event organizers

These sources were integrated into a unified business intelligence environment. A machine learning layer was added to optimize budget allocation and measure incremental campaign impact: time-series forecasting using Prophet, gradient boosting using XGBoost and LightGBM, Marketing Mix Modeling using LightweightMMM, and causal inference using CausalImpact. The ML stack was not required for basic attribution. It was built to answer a harder question: what would have happened to arrivals if we had not run the campaign at all? Causal inference allows you to estimate that counterfactual, which is the only way to claim true incrementality.

Building this infrastructure took three years and a dedicated Business Intelligence team. It required sustained executive commitment, government data-sharing agreements, and a board willing to fund capabilities whose full value would not be visible for 12 to 18 months after implementation. That institutional commitment is the real barrier for most DMOs. The technology is available. The organizational will to invest in it is not.

Start Measuring Where You Can

The PROMTUR attribution model had gaps. Direct marketing attribution covered only one of ten source markets at the arrival level. Travel trade was excluded entirely. The full economic contribution of the promotional program was larger than the USD 1.8B reported. We knew all of this.

We reported it anyway, because a conservative and defensible number is more valuable than a large unverifiable one. The USD 1.8B figure was real. The direct marketing return ratio was real. The gaps were documented and disclosed. The reported outcome was the floor, not the ceiling.

The argument against building this capability is almost always the same: the data is imperfect, the methodology is contested, and the gaps will be used against you. That argument protects the organization from accountability. It does not serve the destination, the industry, or the public investment that funds the DMO.

Many destinations have access to some form of arrival data through immigration services, national statistics agencies, or tourism authorities, though the quality, granularity, and accessibility of that data varies considerably by country. Most established tourism markets have hotel performance benchmarking available. Most tourism authorities conduct periodic visitor expenditure surveys, though frequency and methodology differ widely. The building blocks for a basic attribution model exist in more destinations than currently use them. What is typically missing is not the data itself but the decision to invest in connecting it, the government relationships required to access it, and the institutional commitment to sustain the effort.

Start with what you can measure. Report what you can prove. Disclose what you cannot. Build more capability each year. The gap between what tourism investment produces and what the sector can demonstrate it produces is the single largest obstacle to securing the budgets, the government mandates, and the development bank partnerships that destinations need. Closing that gap is not a technical challenge. It is a governance decision.

Woodrow Oldford is Managing Principal of Oldford Global Consulting. He served as Chief Marketing Officer, Chief Strategy Officer, and Acting CEO of PROMTUR Panama from 2020 to 2023.