Most travel sites try to do two jobs at once. They want to rank for "best eco-hotels in Galway" and they want to convert someone who already knows they need a room next Friday. Those are different jobs. They want different page shapes, different schema, different tones of voice, and — if you're being honest about it — different domains. That's the bet behind running impthotels.com and impttravel.com as two separate properties under the IMPT roof, instead of cramming everything into one stack.
Two domains, two jobs
The split is straightforward once you say it out loud.
impthotels.com is the directory. It's the place where a page exists for a hotel, a city, a region, and a country, and where the job of that page is to help someone book. It's transactional. The content is structured: rooms, amenities, location, sustainability notes, price range, availability. The user arrived because they searched for a hotel, or for a city plus the word "hotel", and they want a list of options that load fast and link straight through to a booking flow that offsets a tonne of CO₂ on-chain at our cost.
impttravel.com is the editorial site. It's where the writing lives. Itineraries, destination guides, food trails, sustainable travel essays, founder notes, the long stuff. The job of those pages is to be read, to be linked to, and to give someone a reason to care about a place before they care about a bed. The user arrived because they searched "weekend in west Cork" or "carbon-neutral travel Ireland", and they want a story, not a price grid.
You can do both jobs on one domain. Booking.com does. Tripadvisor does. But they're two-decade incumbents with brand authority that lets Google forgive a lot of crossed wires. We're not. So we split.
Why the split actually helps SEO
Search engines reward clarity of intent. A page that mixes long-form editorial with a booking widget and a price comparison table is a page that ranks for nothing in particular. A page that is plainly a directory listing, with the right schema and a clear conversion path, ranks as a directory listing. A page that is plainly an essay about cycling the Waterford Greenway, with no booking widget shoved into the middle of it, ranks as an essay about cycling the Waterford Greenway.
The other thing splitting does is let each domain build the kind of backlinks it can actually earn. Directories get linked to by aggregators, by city tourism pages, by people listing "where to find sustainable hotels". Editorial gets linked to by other editorial — by bloggers, by journalists, by people who are writing their own travel pieces and need to cite something. Those are different link economies. Trying to court both with one domain means you usually get neither.
And then there's the simple matter of crawl budget. A site with hundreds of thousands of hotel pages is a crawl problem. A site with a few hundred long-form articles is not. Letting the directory be a directory means we can be aggressive about sitemap structure, internal linking, and pagination without worrying that we're burying the editorial. Letting the editorial be editorial means each piece gets the attention it needs.
The hreflang strategy
IMPT books hotels in 195 countries. The directory has to speak to a user in Dublin, in Düsseldorf, in Dubai, in Denver. Hreflang is how you tell Google which version of a page belongs to which audience, and it's the part of international SEO that almost everyone gets wrong.
The rule we follow on impthotels.com is simple: a page exists once per language-region pair we genuinely serve, and every variant in the cluster points at every other variant, including itself, with the right hreflang tag. So the Clonmel directory page in English-for-Ireland references the same page in English-for-the-UK, in English-for-the-US, and so on. A user searching from Cork gets the IE variant. A user searching from Boston gets the US variant. The content can be substantially the same — the currency, the spelling, the date format, and the regional context vary.
What we don't do is generate machine-translated pages just to fill the matrix. If we don't have a real, reviewed version of a page in German, we don't pretend we do. Bad translation kills trust and Google can tell. The hreflang map only includes pairs we've actually built.
For impttravel.com the hreflang load is lighter, because editorial doesn't need to fork by region the way a directory does. An essay about a food trail through Tipperary reads the same in Sydney as it does in Stuttgart. We localise spelling and currency where it matters, and leave the rest alone.
Schema.org, properly
If you're building a sustainable hotel directory in 2026 and you're not shipping clean schema.org markup on every page, you're leaving rankings on the table. Google's hotel results, the rich cards, the price annotations, the review stars — all of that is downstream of structured data.
On impthotels.com, every hotel page ships:
Hotelas the primary type, withaddress,geo,starRating,amenityFeature, andcheckinTime/checkoutTimepopulated.aggregateRatingandreviewwhere we have real review data, never invented.Offernodes for the price information we're allowed to surface, with currency that matches the hreflang variant.BreadcrumbListso the country → region → city → hotel hierarchy is legible to crawlers.- Sustainability fields where the property has verifiable credentials. We don't tag a hotel as eco-certified without the certificate.
City and region pages use CollectionPage with an ItemList of hotels, plus Place for the location itself. It's the unglamorous plumbing that makes a schema.org hotel directory legible to search.
On impttravel.com, the schema is mostly Article, TravelAction, and TouristDestination, with author nodes that point to real people on real profile pages. The cross-link from an editorial piece to a directory page uses a plain anchor — no schema tricks, no sleight of hand.
The long-tail bet
Big travel keywords are won. "Hotels in Dublin" belongs to Booking, Expedia, Hotels.com and a handful of others, and we are not going to outspend them on brand search. So we don't try.
The bet is the long tail. "Eco-hotels in Kilkenny with electric car chargers." "Carbon-neutral hen party hotels in the west of Ireland." "Hotels near the Suir Blueway with on-site bike hire." Each of those queries has small monthly volume on its own. Stacked together, across a directory of 1.7 million hotels and a growing editorial archive, the long tail is where the traffic actually lives. And it's where intent is highest, because someone typing eight words into a search bar knows what they want.
For the Irish market specifically, that long-tail strategy lines up with the way people actually plan a trip here. Nobody books a generic "hotel in Ireland". They book a specific weekend in a specific town for a specific reason. The IMPT directory of eco-hotels in Clonmel is one example of what that looks like at a single-town level — a page built for the person who already knows where they're going.
The editorial side feeds the directory. A piece on impttravel.com about a sustainable weekend in a specific Irish town will, in its own time, rank for queries the directory page can't reach, and will pass a steady trickle of warm traffic across to the booking flow. The directory doesn't need to be a magazine. The magazine doesn't need to be a shop. They just need to know about each other.
What this costs us, and why we still do it
Two domains is more work. Two sets of analytics, two CDN configurations, two sitemap pipelines, two content calendars, two engineering teams thinking about caching. Cross-domain attribution is harder. A user who reads an essay on impttravel.com on a Tuesday and books on impthotels.com on a Friday is, in raw analytics terms, two different sessions on two different domains. We've built the plumbing to stitch those journeys together, and that plumbing was not free.
What we get for it is a site architecture that matches the way travel decisions actually get made. Inspiration first, transaction second, on different domains, with different schema, ranking for different queries. It's the boring version of green travel SEO — no tricks, no hacks, just two clean stacks doing one job each — and over the course of a couple of years it compounds in a way a single bloated domain doesn't.
What we're doing about it this quarter
If you run a travel or hospitality site and you've been trying to do editorial and transaction on one domain, the thing to do this week is open your search console and look at what your top editorial pages actually rank for, versus what your top product pages rank for. If those two lists overlap, you're cannibalising yourself, and the fix is structural, not cosmetic. On the IMPT side, we're spending this quarter tightening hreflang coverage on the directory for the markets where bookings are growing, expanding the editorial archive on impttravel.com with itineraries that point at real, verifiable sustainable properties, and continuing the slow, unglamorous work of making sure every page on both domains carries the schema.org markup it deserves. No shortcuts. Long tail, long game.