City-level interest in Croatia grew 120.0% over the last two weeks while country-level interest grew 25.5%, a 94.5 percentage point gap, according to our travel intelligence network. That is a roughly 4.7x ratio between the two layers of the planning funnel in a single 14-day window. The country page is still climbing. But the city pages are climbing almost five times faster.
The Pattern
The two readings sit on top of the same destination but describe different research behaviors. Country-level pages, which typically catch travelers in the "should I go to Croatia" phase, posted 25.5% growth. City-level pages, which catch travelers already past that question and asking "what do I do in Split, Dubrovnik, Zagreb, Hvar," posted 120.0% growth.
The divergence is negative 94.5 percentage points from the country's perspective. In plain terms, the country tier is lagging its own cities by nearly a hundred points of two-week growth.
A spread of this size is not noise inside a normal browsing pattern. It indicates the audience hitting Croatia content right now is disproportionately weighted toward people who have already committed to the destination and are sequencing an itinerary, rather than people still evaluating whether Croatia belongs on their shortlist.
What The Data States
Right now, in our panel, Croatia's city-tier pages are capturing the dominant share of the country's interest growth. The country-tier page is still in positive territory at 25.5%, so top-of-funnel demand has not stalled. But the planning-stage attention is concentrating downstream, at the itinerary layer, not at the destination-selection layer.
For travel industry professionals working Croatia inventory, the present-tense read is straightforward: the marginal user arriving at Croatia content this fortnight is closer to booking than the marginal user a fortnight ago. Channel mix, content priorities, and bid strategies built around country-level keywords are addressing a smaller share of the active demand than city-level placements are. The data does not identify a trigger for the split, and no campaign, route announcement, or policy change is visible in this dataset. Whether the gap reflects a seasonal pull-forward, a shift in source markets, or a content-discovery effect on the city pages themselves is not resolvable from these two numbers alone. What is resolvable is that the audience composition has tilted, and tilted hard, toward the city tier.
One synthesis point worth holding: a 94.5 point spread sustained across a single window can be a one-print anomaly. A 94.5 point spread that survives the next print is a structural change in how Croatia is being researched, and that is a different planning problem than a hot week in Split.
Open Questions
The next reading is what tells us whether this is signal or noise. Specific data points to watch:
- Whether country-level growth in Croatia stays in the 20-30% band or accelerates toward the city-level rate. Convergence upward would suggest the country tier is catching up. Continued lag would confirm the funnel has split.
- Whether city-level growth holds near 120.0% or reverts. A reversion toward the country-level rate would mark this fortnight as a spike rather than a regime change.
- Whether the divergence stays negative from the country's perspective (cities outpacing) or flips. A flip would indicate top-of-funnel demand has resumed leadership.
- Whether the city-tier growth is concentrated in a small number of Croatian cities or spread across the urban portfolio. The current data reports the city aggregate, not the per-city distribution, so the breadth of the increase is an open question.
- Whether the 4.7x city-to-country growth ratio compresses, holds, or widens in the next two-week print. That ratio is the cleanest single number to track this pattern going forward.
Until the next reading lands, the defensible statement is the narrow one: Croatia's city pages are currently growing at almost five times the rate of its country page, and the planning-research stage for this destination is, for now, a city-led conversation.
Methodology
Data comes from Prospxct's proprietary travel intelligence panel, a network of 500+ destination-specific travel planning sites, each covering a single city, country, or region. All sites run on an unified analytics stack, allowing us to compare relative traffic patterns across destinations on a like-for-like basis.
For this study, we compare total traffic across two consecutive 14-day windows at both the country level and the city level, surfacing destinations where the two cohorts have diverged materially over the period.
We report percentages, ratios, and rankings, not absolute traffic volumes. All data reflects observed planning behaviour (users actively researching activities and logistics), not booking transactions or airport arrivals.
Last verified: .