Hotel SEO Direct Bookings Case Study in Pakistan

The Client

A Murree-based boutique hillside resort with 52 rooms spread across a main building and a cluster of cottage-style suites, operating year-round but peaking hard during the summer domestic-tourism rush and the winter snowfall window. The property had built a respectable reputation over eight years — strong reviews on travel platforms, a loyal repeat guest base travelling up from Islamabad and Rawalpindi, and healthy occupancy from May through August. Revenue was not the problem. How that revenue was arriving was.

Roughly seven in ten of every booking came through an online travel agency — Booking.com, Agoda, and a couple of domestic aggregators — at commission rates between 18% and 22%. On an average nightly rate of PKR 18,000 for a standard room and PKR 32,000 for a suite, that commission worked out to between PKR 3.2M and PKR 4.0M leaving the business every month during peak season. The owner understood that OTAs had built the early demand, but the property had long outgrown that crutch. The direct website — the only channel where the resort kept full margin — was barely contributing.

When the resort approached WeProms Digital, organic search was responsible for a thin slice of direct bookings, and almost all of it came from people who already knew the resort’s name. The team had tried boosting the occasional social post and running a small paid campaign on the brand name, but no one had ever treated organic search as a serious direct-booking channel. They wanted an SEO audit and strategy that would restructure the site around how Pakistani travellers actually search for a place to stay in Murree — and then rank for it.

The Problem

Four issues were keeping the resort invisible to the travellers actively searching for accommodation it could serve:

1. The site was indexed for the brand name and almost nothing else. Google had indexed around 40 URLs, nearly all of them branded — homepage, contact, a single rooms page, dining. For the hundreds of unbranded queries Pakistani travellers type every week (“best hotel in murree for families”, “murree hotel with heater in winter”, “patriata hotel booking”, “hotel near mall road murree”) the resort did not exist in the results. Search Console confirmed it: the site showed impressions for “murree hotel” variants but earned a click-through rate under 1.2%.

2. One thin rooms page covered the entire property. Every room type — standard, deluxe, family suite, honeymoon cottage — shared a single page with a short paragraph pulled from an eight-year-old brochure and a gallery of low-resolution images. There was nothing for Google to rank for “family suite murree” or “honeymoon cottage murree with view”, and nothing for a traveller to evaluate before booking.

3. No local SEO presence to speak of. The Google Business Profile existed but was barely managed — outdated photos, an incomplete amenities list, slow or absent responses to reviews, and no Q&A. For destination searches with strong local intent (“hotels near mall road murree”), the Map Pack is the entire game, and the resort was not winning it.

4. Direct booking had no reason to exist. Even when a traveller found the site, the booking path routed them to an OTA widget, or to a phone number, or to a WhatsApp chat — there was no clean on-site booking engine that captured demand at the moment of intent. Organic traffic was not just small; it was structurally unable to convert into direct revenue.

The net effect was a property that ranked well in guests’ memories but was invisible at the moment of search. Every month, thousands of travellers were Googling exactly the kinds of stays this resort offered, finding competitors and aggregator pages instead, and booking through channels that cost the resort a fifth of the room rate.

Phase 1 — SEO Audit and Intent Diagnosis (Weeks 1-2)

Before building anything, we needed to understand what Pakistani travellers were actually searching for, where the site stood technically, and how the booking funnel leaked. The first two weeks were diagnostic.

Technical and crawl audit. The site was built on an aging custom CMS with a booking plugin that generated parameter-heavy URLs. Screaming Frog and Google Search Console turned up the usual suspects: 180+ crawl errors, a robots.txt that was blocking the JavaScript-powered booking engine entirely, no XML sitemap, missing canonical tags on duplicate room-filter URLs, and image alt text that read “IMG_2049” across the gallery. Mobile LCP sat at 4.6 seconds on the connection speed most domestic travellers on mobile data actually experience. None of it was catastrophic, but all of it compounded.

Intent research. This was the real diagnostic work. We pulled search demand for the destination using a mix of Search Console, keyword tools, and autocomplete scraping, then clustered roughly 260 queries into four intent stages:

Intent stage	Example queries	Volume signal
Destination exploration	”places to visit near murree”, “murree in december weather”	High, informational
Property comparison	”best hotel in murree for families”, “murree hotels with parking”	High, commercial
Specific booking	”mall road murree hotel booking”, “patriata hotel price”	Medium, transactional
Brand / direct	”resort name murree”, “resort name contact”	Low, branded

The pattern was the important part. The resort was present for the bottom row and absent for the top three — which together accounted for the overwhelming majority of actual booking-intent demand. The entire restructure would be built around closing that gap.

Booking-funnel and OTA analysis. We pulled six months of booking data and split it by channel: OTA, direct website, phone and WhatsApp, and corporate and walk-in. OTA carried 71% of bookings, direct online just 22%, and the remainder split between phone and walk-in. Commission leakage was quantified precisely so that every direct-booking gain could be valued in rupees saved, not just sessions gained.

Phase 1 results (by end of week 2): crawl errors triaged, sitemap and robots.txt fixed, mobile LCP improved to 2.8 seconds with image compression and lazy loading, and an intent map defining the entire Phase 2 build.

Phase 2 — Intent Restructure and Page Build (Weeks 3-7)

With a clean technical base and a clear intent map, we rebuilt the site’s architecture so that every significant booking-intent cluster had a dedicated page designed to rank and convert.

Room-type landing pages. The single thin rooms page was split into dedicated landing pages for each room and cottage type, each built around a specific intent: standard double, deluxe valley-view, family suite, honeymoon cottage. Every page carried a unique 250-400 word description covering capacity, view, heating, amenities, and the traveller profile it suited — written for the guest evaluating a stay, not for a search engine. Each targeted a cluster of queries (“family suite murree”, “murree hotel 5 person room”) rather than a single keyword.

Intent-led experience pages. Beyond room types, we built commercial-intent pages for the demand clusters the audit had surfaced: a “things to do near the resort” guide targeting destination-exploration queries, amenity-led pages (“heated rooms for winter stays”, “family-friendly resort murree”), and seasonal booking pages for summer and snowfall windows. These pages did two jobs at once — they captured top-of-funnel travellers early and they internal-linked into the room and booking pages, passing relevance and keeping visitors on the direct site rather than bouncing to an aggregator.

The direct booking engine. Structurally, the biggest change was giving organic traffic somewhere to convert. We routed every landing page into a clean on-site booking flow with real-time availability, removing the OTA widget from the primary path and replacing it with a direct-booking incentive (a complimentary breakfast or late checkout) that gave travellers a reason to book direct instead of defaulting to the aggregator they arrived from. You can see the same conversion-path logic in our on-page SEO services work — traffic without a conversion path is wasted spend.

Schema and structured data. Each room page received LodgingBusiness and Hotel schema with amenity, price-band, geo, and review data, making the resort eligible for rich results and the structured fields in the Map Pack. This is where technical SEO and local visibility converge for hospitality.

Phase 2 results (by end of week 7):

Build area	Before	After
Indexed URLs	~40 (mostly branded)	120+ intent-targeted pages
Room-type pages	1 thin page	6 unique landing pages
Direct booking path	OTA widget	On-site engine with incentive
Structured data	None	LodgingBusiness schema sitewide

Phase 3 — Local, Technical, and Content Optimisation (Weeks 5-12)

With the architecture in place, the third phase pushed the resort up the rankings by fixing the signals Google uses to rank hospitality businesses: local presence, technical health, and content authority.

Google Business Profile overhaul. For destination and proximity searches, the Map Pack decides most bookings, so we rebuilt the profile from scratch: accurate categories, a full amenity list, 40+ new high-resolution photos organised by room type and amenity, a complete services list, and a posting cadence of weekly updates tied to seasonality. We stood up a review-velocity programme — a post-stay email and WhatsApp message inviting recent guests to leave a review — that lifted the review count from roughly 180 to over 340 during the phase and raised the average rating by a tenth of a star. Review velocity is one of the strongest local ranking levers, and it compounds.

Core Web Vitals and mobile. Domestic tourism searches happen overwhelmingly on mobile, often on patchy data connections in transit. We pushed mobile LCP from 2.8 seconds down to 1.7 seconds by deferring the booking engine’s render-blocking scripts, serving responsive WebP imagery, and removing two third-party widgets that were loading on every page. Interaction latency came down to under 200 milliseconds. The speed gains improved rankings and reduced drop-off before the booking engine could load.

Internal linking and topic clusters. The experience and seasonal pages were wired into the room pages with descriptive anchor text, creating clear topical clusters around “murree family stay”, “winter murree”, and “honeymoon murree”. Internal link density rose from roughly 1.4 to 5.1 links per page, and link value flowed from the higher-authority guide content into the commercial room and booking pages.

Content velocity. We set a steady cadence of destination content — guides to nearby Patriata, Ayubia, and Bhurban, seasonal packing lists, weather explainers — that captured exploratory demand and funnelled it toward booking. Each piece linked into the relevant room or seasonal landing page. This is the same hospitality-vertical logic we apply across digital marketing for hotels, tuned to the destination.

Phase 3 results (by end of week 12): 54 non-branded keywords on page one (up from 9), search-to-site click-through up to 3.6% on target queries, and a Map Pack presence for “murree hotels” variants where the resort had previously not appeared.

Phase 4 — Measurement and Compounding (Weeks 10-20)

The final phase turned the restructure into a compounding channel the resort’s small team could sustain, with attribution that tied organic sessions directly to direct-booking revenue.

Direct-booking attribution. The earlier leakage analysis paid off here. We instrumented the booking engine so that every direct booking could be attributed back to its source — organic search, paid, referral, direct — and back to the landing page and intent cluster that started the journey. This replaced guesswork with a clear view of which pages earned which bookings, and it let us value organic growth in commission saved rather than vanity sessions.

Iteration and pruning. With attribution live, we doubled down on the intent clusters that converted and pruned or merged pages that attracted traffic but never booked. Two seasonal landing pages were expanded into year-round hubs; three thin amenity pages were consolidated. The site got sharper rather than just larger.

Seasonal campaigns built on the organic base. With organic direct-booking sessions compounding, paid media shifted from compensating for absent organic to amplifying proven pages — promoting the summer and snowfall landing pages to warm audiences during booking windows, rather than cold-targeting brand searches.

By the five-month mark, organic direct-booking sessions had grown 67% against the baseline, with non-branded organic traffic more than doubling. The resort was visible at the moment travellers were searching, and those travellers had a clean path to book direct.

Final Results at 5 Months

Metric	Before	At 5 months	Change
Organic direct-booking sessions	Baseline	+67%	Non-branded +118%
OTA share of bookings	71%	52%	-19 pts
Direct online revenue share	22%	38%	+16 pts
Indexed URLs	~40	180+	+350%
Search-to-site click-through	1.2%	3.6%	+200%
Page-1 non-branded keywords	9	54	+400%
Google Business Profile reviews	~180	340+	+88%

Every number traces back to a phase: crawl health and the intent map from Phase 1, the room and booking pages from Phase 2, the local and Core Web Vitals gains from Phase 3, and the durable direct-booking attribution and compounding from Phase 4.

What Made This Work

1. The site was restructured around intent, not rooms. The single biggest lever was the Phase 2 restructure. A hotel that only ranks for its own name is invisible at the moment of search. Building a page for every meaningful booking-intent cluster — room types, amenities, seasons, traveller profiles — is what made the resort appear for the queries travellers actually type.

2. Direct booking had to be earnable, not just possible. Traffic without a conversion path is waste. Giving organic visitors a clean on-site booking flow with a reason to book direct (a tangible perk) is what turned the 67% session gain into real commission saved, not just a larger analytics number.

3. Local SEO carried the transactional tail. For destination and proximity searches, the Map Pack decides most bookings. The Google Business Profile overhaul and review-velocity programme were not a side task — they were where a large share of the direct-booking lift actually came from.

4. Attribution made organic worth investing in. Tying each booking to its landing page and intent cluster let the owner see organic as a margin channel, measured in commission saved, rather than an abstract traffic line item. That is what unlocked sustained investment.

What Teams Can Apply

For Pakistani hotels and guesthouses that want to stop renting their guests from OTAs:

1. Map intent before you build pages. Cluster the queries travellers use to find your destination, then give each booking-intent cluster its own page. A single rooms page cannot rank for “family suite murree”, “honeymoon cottage murree”, and “heated room for winter” at once.

2. Give organic traffic somewhere to convert. A booking engine with a small direct-booking perk beats an OTA widget on every page. If a traveller who finds you organically still books through the aggregator, you have paid twice — once in lost margin, once in invisible SEO.

3. Treat your Google Business Profile as a primary channel. Accurate categories, fresh photos, and steady review velocity move the Map Pack more than almost anything else for destination searches.

4. Optimise for the mobile connection your guests actually have. Domestic tourism searches happen on mobile data, often in transit. A fast, lightweight page ranks better and converts better than a heavy one.

WeProms Digital has applied this intent-led hotel SEO framework across Pakistani properties in Murree, Nathia Gali, and northern destinations, and for city business hotels and guesthouse networks. The destination and keywords change with each property — the intent-first, conversion-ready, attribution-backed approach stays the same.