Google Ads for Hotels and Guesthouses in Nepal: Competing With the OTAs for Direct Bookings

The short answer

Most advice about hotel search advertising is written as though the auction is a contest between hotels. It usually is not. Search your own property name and look at what sits above your listing: in a great many cases it is an online travel agency, sometimes several, plus a metasearch site, all bidding on a term that describes your building. They are competing with each other for the right to sell your rooms, and the budget they are doing it with is funded by a commission they earn on those rooms.

That is the asymmetry the whole category runs on, and it is genuinely uneven. An OTA is bidding across thousands of properties with a known expected value per click at portfolio scale, backed by conversion machinery it has spent years refining and a brand travelers already trust with their card details. You are one property with one set of rooms and a site your cousin built. Treating that as a fair fight you can win by outbidding is the fastest way to spend a season's marketing budget on nothing.

The way out is not outbidding. It is understanding that you and the OTA have different cost structures for the same booking, and that the difference is what justifies a bid. An OTA booking costs you a commission on that stay, and it costs you the guest relationship — the email address, the repeat visit, the direct rebooking next season. A direct booking costs you the acquisition once. Whether the gap between those two things is bigger than the cost of buying the click is the single calculation that decides whether a hotel should be in the auction at all. Most small properties never do it.

Who you are actually competing against

Before deciding anything, look at the auction as it actually is rather than as you assume it is. Run the searches yourself, on a phone, several times over several days, and record what appears. Three categories of competitor show up on hotel queries, and they behave differently enough that lumping them together produces bad decisions.

• Online travel agencies bidding on your property name specifically. This is the defensive problem. They are selling your inventory, so the click is worth something concrete to them regardless of what you do.
• Online travel agencies and metasearch sites bidding on category and location terms — the phrases people use before they know which property they want. Here they are not attacking you; they are simply better funded than you on a term you would also like. This is where a small property loses money fastest.
• Other hotels, which are usually the least aggressive presence in the auction and the one small properties worry about most. A neighboring guesthouse bidding on the same location term is a normal competitor with roughly your constraints. The OTA is not.

That ordering matters because the correct response differs in each case. Against an OTA on your own name, the question is defensive and testable, and the rest of this post takes it apart. Against an OTA on generic location terms, the honest answer is often that you should not be competing head-on at all, and should instead be buying the narrower, more specific queries where a large aggregator has nothing better to say than you do. Against other hotels, ordinary competitive discipline applies.

One more thing to record while you are looking: whether the OTA result on your name is a paid ad or an organic listing. An OTA ranking organically above you costs you nothing in the auction and cannot be outbid — that is a search visibility problem, not an advertising one, and buying clicks will not fix it. People conflate the two constantly and then wonder why the ad spend did not change the picture.

The direct-versus-OTA math, as a method

I am not going to give you a commission percentage, because commission rates vary by platform, by contract, by property tier and by whatever program the hotel has been enrolled into, and a number I invented would be worse than no number at all. What follows is the method. Fill it in with your own actual rate, taken from your own actual statements.

Start with a single booking and follow the money in both channels. For the OTA booking: the room revenue, minus the commission actually deducted, minus any payment processing the platform applies, gives you what lands in your account. For the direct booking: the same room revenue, minus your own payment processing, minus whatever it cost to acquire that specific booking through advertising. The gap between the two is the per-booking value of a channel shift.

Then extend it beyond the single stay, because that is where the real difference is and where the arithmetic usually gets stopped too early. An OTA booking gives you a guest whose email address, in most arrangements, you do not own. If that guest returns, they return through the same channel and you pay again. A direct booking gives you the guest record, the consent to contact them, and the ability to have the second stay cost you nothing at all. Whether that matters depends entirely on your repeat rate, which is a number sitting in your own reservation system right now.

1. Take your own commission rate from your last statement. Not a rate someone quoted in an article — the deduction actually applied to your bookings.
2. Compute the per-booking cash difference between an OTA booking and a direct booking of the same room, same length, same season.
3. Look up your own repeat rate and how repeat guests currently rebook. If most repeats come back through the OTA, the lifetime difference is much larger than the single-stay difference.
4. Set the resulting figure as your ceiling for what a direct booking may cost to acquire. Not a target — a ceiling, above which the shift stops being worth making.
5. Divide that ceiling by your realistic site conversion rate to get the most a click can be worth to you. If you do not know your site conversion rate, that is the first thing to measure, and until you do, every bid you place is a guess.

That last step is where most hotel accounts quietly fail. A property with a booking engine that converts poorly can have an excellent commission argument on paper and still lose money on every campaign, because the cost of the clicks needed to produce one booking exceeds the value of shifting it. The math does not just tell you whether to advertise. It tells you what has to be true about your website before advertising can work, which is a much more useful output.

The brand-defense question, answered honestly

This is the question every hotel asks first, and the confident answers on both sides are both wrong. One camp says never pay for traffic you already own. The other says always defend your brand or the OTA takes it. Neither is a rule, because the answer genuinely depends on three variables that differ property to property.

Variable one: is anyone actually bidding on you?

Check before assuming. Many small properties defend a brand term nobody is attacking, which is the purest form of wasted spend — you buy a click you would have received free, at a price set by an auction with only one participant. Search your property name from a phone, from a browser with no personalization, and ideally from a connection in one of your actual source markets, because ad presence varies by the searcher's location. Do it more than once; OTA bidding on individual property names is not always constant.

If nothing is above your organic listing on your own name, the defensive argument disappears entirely and brand bidding becomes a question of whether your own listing is doing its job — which is a Business Profile and search visibility question, not an ads question. Google Business Profile is available in Nepal, confirmed on Google's own supported countries list, and for a hotel it is the highest-return unpaid asset there is.

Variable two: what does brand traffic do when it arrives?

A brand click is a person who already decided on you and is looking for the way to book. If your site converts that person reliably, defending the term has obvious value, because every one you lose to an OTA is a full-price booking that becomes a commissioned one. If your site loses them anyway — no availability calendar, no rate visible, a booking flow that breaks on a phone — then defending the term is paying to send a warm guest into a broken funnel, and the OTA is arguably doing you a favor by converting them at all.

This is uncomfortable and it is true. There are properties for which the honest sequence is: stop bidding on your own name, spend the equivalent on fixing the booking path, and revisit the auction question when the site can hold a booking. The OTA is not beating you on bid price. It is beating you on the checkout.

Variable three: would you have had the booking anyway?

This is the incrementality problem and it is the one that makes brand campaigns look so good in reporting. Someone who searches your property name by name is already yours. If they click the ad, the campaign records a conversion. If the ad had not existed, most of them would have scrolled slightly and clicked the organic result, and you would have recorded the same booking with no media cost. The conversion is attributed; it is not necessarily caused.

So the brand campaign will always look like the best-performing thing in the account, because it is buying an audience that had already decided. Cost per conversion will be the lowest, return on ad spend the highest, and every automated recommendation you receive will suggest putting more money into it. None of that tells you whether it produced a single booking you would not otherwise have had.

How to actually test brand bidding, rather than asserting a rule

The only way to answer the incrementality question is to turn it off and watch what happens to the business, not to the account. This is a real test with real risk, so run it deliberately rather than by accident.

1. Define the measurement before you touch anything. The number you are watching is total direct bookings from the property's own reservation system, and direct share of total bookings. Not clicks, not conversions in the ads interface, not sessions.
2. Choose a period long enough to be readable and comparable. Hotel demand is seasonal and lumpy, so a week against a week in a different season proves nothing. Compare like against like, and expect this to take longer than you want.
3. Pause the brand campaign only. Leave every other campaign running and unchanged, so the only variable that moved is the one you are testing.
4. Watch organic and direct traffic as well as bookings. If brand paid clicks were largely cannibalizing your own organic clicks, you will see organic clicks rise to partially replace them — that recovery is the clearest single signal that the spend was not incremental.
5. Watch the OTA channel too. If direct bookings fall and OTA bookings rise by a similar amount over the same window, that is the interception the defensive argument predicts, and you now have evidence for it instead of a theory.
6. Decide from the business numbers. If total direct bookings held up without the campaign, the campaign was buying traffic you already owned. If they fell and OTA volume rose, defending the term is doing real work and you should fund it.

Two honest caveats about this test. It is noisy for a small property, because a hotel with modest volume may not produce enough bookings in any comparable window to distinguish a real effect from ordinary variation — in which case the correct conclusion is that you cannot measure it, not that the effect is zero. And a test run during a period when the OTA happened not to be bidding on you tells you nothing about periods when it is. Record what the auction looked like during the test, not just the outcome.

The general discipline underneath this — deciding in advance what evidence would change your mind, and reading business outcomes rather than platform metrics — is the same one behind the bid governance framework, and it applies with unusual force here because the brand campaign is the single most flattering line item in a hotel account.

What Nepal specifically does not give you

Most hotel marketing advice online is written for markets with a fuller Google feature set than Nepal has. Several of the tools that article assumes are not available here, and each absence removes something a hotel would specifically want. These are not opinions about what works badly — they are absences from Google's own published country lists, checked in July 2026. The full country guide covers the platform and regulatory picture in detail, including payment routes and foreign-exchange ceilings; what follows is only the part that changes a hotel account.

Calls cannot be measured, and hotels run on calls

Google publishes the complete list of countries eligible for call reporting and Google forwarding numbers, and Nepal is not on it. Get the nuance right: a call asset carrying your own phone number will still serve on your ads. What does not exist is the forwarding number that substitutes for yours, the resulting call counts, and call-duration conversions. So calls generated by a Nepali hotel's Google Ads cannot be counted natively at all.

For a category where a meaningful share of enquiries — group bookings, long stays, anything with a question about transfers or dietary needs — arrives as a phone call, this is a serious gap. Automated bidding allocates budget toward conversions it can observe. If your calls are invisible and your form fills are not, the system will move money toward whatever produces forms and away from whatever produces calls, and it will do so while reporting an improving cost per conversion. That is a structural trap, not a configuration error, and closing it takes third-party call tracking or at minimum a distinct phone number used only on paid landing pages, reconciled by hand.

Lead form assets are unavailable

Google's documentation states that lead forms are only eligible to serve in some countries and that audiences in countries which do not allow them will not see the form. The eligible list runs to eighty-five countries and includes Bangladesh, India, Pakistan and Sri Lanka. Nepal is not on it. So in-ad enquiry capture — the format a property might reasonably want for group and event enquiries — is off the table, and the traffic has to reach your own page and your own form. That is better practice regardless, since the data lands in your systems, but it removes a fallback that hotels in neighboring markets have.

Feed-based products, and what I could not verify

Google Merchant Center is verifiably unavailable for Nepal — the country does not appear on Google's supported countries and currencies list, and neither do Pakistan, Bangladesh or Sri Lanka, while India does. That removes retail-goal Performance Max and every feed-driven product that depends on it.

On Google's hotel-specific booking feed products I am going to be straightforwardly honest: I could not verify their availability for Nepal from a Google-published country list, in either direction. The feature lists I was able to fetch and read cover Merchant Center, call reporting, lead forms, language targeting and geo targets. Hotel booking feeds are not among them. So this post does not tell you that they work here, and does not tell you that they do not. Check Google's own current documentation and the requirements attached to those products before building a plan that depends on them — and treat any article that states the answer confidently without showing you where it came from as unreliable, because that is the same pattern that produces the invented cost figures discussed in the post on what Google Ads actually costs in Nepal.

Nepali is not a targetable language

Google's published table of languages available for ad targeting lists fifty-one languages, and Nepali is not among them, while Hindi, Bengali, Urdu, Tamil and several other South Asian languages are. Google also states that ads created in a language not supported for ads language targeting will be disapproved. For an inbound tourism property this constraint bites less than it might, because your guests are searching in English from abroad and English creative is the natural fit. For a property chasing domestic weekend business it is a genuine limitation, and the practical route practitioners take is English creative with geographic targeting carrying the load — which is common practice, not documented Google guidance.

What Nepal does give you: geography, in detail

The one thing Nepal has in abundance is geographic targeting granularity. Google's official geo-targets file, in the release dated 6 July 2026, contains 114 active Nepal targets: the country itself, five provinces, five legacy development regions, plus districts, cities, divisions and roughly eighty municipalities. For a hotel that is useful in both directions — targeting domestic travelers by origin city, and excluding areas you know do not convert. It is also, for reasons the next section explains, less central to an inbound property's setup than most people expect.

Targeting the guest, not the country you are in

Here is the mistake that costs inbound properties the most, and it is a settings mistake rather than a strategic one. A hotel in Nepal sets its location targeting to Nepal, because that is where the hotel is. But the guest is not in Nepal when they book. They are at home, in another country, three weeks or three months before arrival, comparing options. Targeting Nepal reaches people who are already here — a real audience, but a different and much smaller one than the audience you are trying to fill rooms with.

Source-market targeting

The correct starting point is your own arrivals data. Every property knows which countries its guests come from, because it records nationality at check-in. That list, ordered by volume and by revenue rather than by headcount, is your geographic targeting plan. Build campaigns against the source markets that actually produce your bookings, not against a generic list of wealthy countries, and treat markets with high volume but low revenue per stay differently from the reverse.

There is also a settings detail that matters more for travel than for any other category: Google's location options distinguish between people present in a location and people showing interest in it. For a dental clinic, presence-only is almost always right. For an inbound hotel it is almost always wrong, because the entire point is reaching someone who is interested in your location while sitting somewhere else. Get this setting explicitly right rather than accepting whatever the default was, and keep the domestic campaign — where presence often is the right signal — structurally separate so the two never share it.

Language and device follow from the source market

Once targeting is set by source market, language targeting mostly answers itself: English creative for international campaigns, since Nepali is not available as a targeting language anyway and your inbound audience is not searching in it. Do not extrapolate Nepal's device split onto a foreign audience either. Nepal's own device mix — mobile at 62.13 percent against desktop at 37.46 percent, per StatCounter for June 2026 — describes traffic inside Nepal and tells you nothing about how a traveler in another country researches a trip. Read device performance per campaign from your own data, because trip research and trip booking frequently happen on different devices.

The consideration window is long, and it changes what you buy

Travel has one of the longest gaps between first search and purchase of any consumer category. Someone plans a trek months ahead, reads for weeks, shortlists, and books later — often from a different device, often after several visits, often after being reminded. That length has three practical consequences for the account.

• Last-click reporting will systematically under-credit everything that happens early in the journey and over-credit the final brand search. Look at the assisting paths, not only the closing click, before cutting an early-funnel campaign.
• Remarketing is not optional in this category the way it is in some others. The audience genuinely does leave and come back; the whole shape of travel demand assumes it.
• Your conversion window and attribution settings need to be long enough to actually contain the decision. A window shorter than your typical booking lead time will report that campaigns produced nothing, and you will act on that and be wrong.

Seasonality compounds all of it. Nepal's inbound demand concentrates into particular parts of the year, driven by trekking and mountain tourism weather, and the research that produces those arrivals happens well before them. I am not going to put a figure on the shape of that curve, because I have no source for one that would apply to your property — but you do. Your own arrival records, by month, across several years, are the seasonality dataset that matters, and they should be read together with the lead time between booking and stay to decide when campaigns need to be live. Advertising during the season you want to fill is usually too late.

Campaign structure for a small property

Structure exists to make budget controllable and results readable, and both matter more when the budget is small. Four campaign groups cover what a single property actually needs. Anything beyond them should earn its place by answering a question the existing four cannot.

Brand

Your property name and its common misspellings, plus name-plus-modifier queries such as name with the word booking or with your location. Keep it separate from everything else for one reason above all: it lets you pause it independently for the incrementality test described earlier, and it prevents its flattering numbers from contaminating the reporting on everything else. If brand sits inside a general campaign, its low cost per conversion will make the whole campaign look healthy while the non-brand half quietly loses money.

Non-brand location terms

The queries where someone knows where they are going but not where they are staying — the phrase pattern of hotel in Pokhara, or a neighborhood plus accommodation, or a landmark plus a place to stay. This is the campaign that grows the business, and it is also where you are directly outgunned by aggregators. Compete by being more specific than they can be. A large OTA's landing page for a city is generic by necessity; yours can answer the actual question behind a narrow query, and narrow queries are cheaper.

Match type discipline and negatives matter more here than in almost any other vertical, because travel queries wander. Job searches, property rentals, dissertation research, people looking for a restaurant rather than a room, and an endless tail of place names you do not serve will all find their way in on loose matching. The negative keyword list every service business needs is a reasonable starting frame, and for a hotel I would add employment terms, long-term rental terms and the names of nearby properties you do not want to pay to be compared against unless that comparison is deliberate.

Category and intent terms

Queries describing the kind of stay rather than a place — boutique, family rooms, pet friendly, near the trailhead, budget guesthouse. These are only worth buying where the property genuinely and distinctively matches, and that qualifier is the whole discipline. A guesthouse bidding on luxury terms because they look premium is buying expensive disappointment. The right test for any category term is whether a visitor arriving on it will find the promise confirmed within one screen of your landing page.

Remarketing

Reach people who visited, checked dates, or started a booking and did not finish. Given the length of the travel consideration window, this is where a small hotel budget most reliably earns its keep, because it is the one place you are talking to people who have already shown specific interest in your property rather than in the category. Segment by depth of engagement — someone who reached the booking form and abandoned is a different person from someone who read the location page — and cap frequency, since a property that follows a traveler around for two months reads as desperate rather than attentive.

The booking engine problem

This is the section that gets skipped and the one that decides outcomes. A hotel with no direct booking engine, or with one that is hostile on a phone, cannot convert paid traffic no matter how well the campaign is built. Everything above is downstream of this. There is no bidding strategy, creative test or audience segment that compensates for a guest being unable to complete a reservation.

The comparison a traveler is making is not between your site and another hotel's site. It is between your site and an OTA checkout they have used before, which remembers their details, shows availability instantly, confirms in seconds and takes a card without friction. That is the standard your booking path is being judged against, whether or not that seems fair. A booking flow that requires an email enquiry, a wait for a reply, and a bank transfer is not competing with that. It is losing to it in the first ten seconds.

• Availability and rate must be visible without contacting anyone. A traveler who has to ask the price has already opened another tab.
• The whole flow has to complete on a mid-range phone, on a mobile connection, with one thumb. Test it on the cheapest Android handset in the office, not on the manager's laptop.
• Confirmation must be immediate and unambiguous. Confirm on request is a different product from confirmed, and travelers treat it that way.
• The rate on your site must be at least as good as the rate on the OTA. If your own channel is more expensive than the commissioned one, no amount of advertising will fix the reason people are not booking direct.
• The booking engine has to be able to fire a conversion event with the actual transaction value. If it cannot, your measurement is capped at counting bookings, and the next section explains why that is a problem.
• Phone must be a tap-to-call link on mobile, and someone must actually answer it during the hours your source markets are awake, which are not your office hours.

If several of those are not true today, the recommendation is not a better campaign. It is to spend the money on the booking path first and come back to the auction afterward. That sequencing feels like a delay and is in fact the shortest route, because the same fix raises the conversion rate on every channel you already have, including the traffic arriving free. The general version of this diagnostic work is in where paid traffic actually leaks on landing pages, and every point in it applies with more force to a property whose competitor for the same click is a company that has spent a decade refining exactly this checkout.

Measurement: why a flat conversion value misleads the bidding

Hotel bookings are not interchangeable, and this is where hotel measurement differs most sharply from lead generation. A single night in the cheapest room and six nights in your best suite are both one conversion. If you assign a flat value to every booking, you are telling the bidding system those two outcomes are worth the same, and the system will rationally buy more of whichever is cheaper to win — which is the small one. The account will report more conversions at a lower cost per conversion while revenue goes sideways or down.

The fix is conceptual before it is technical: pass the actual value of each booking through to the conversion, so the system is bidding toward revenue rather than toward booking count. Most booking engines can supply the real transaction value at confirmation, and where they can, this is the highest-value tracking change a hotel account can make. It changes the objective from filling rooms to filling them profitably.

What to be careful about when you do it

• Decide whether the value you pass is gross room revenue or something closer to contribution after variable costs, and apply the same definition consistently. Mixing the two across campaigns makes comparisons meaningless.
• Handle cancellations deliberately. A booking that cancels is not revenue, and if cancellations are common in your property, feeding gross confirmations into the bidding teaches it to buy the bookings most likely to cancel. Importing the confirmed-after-cancellation-window outcome is more work and considerably more honest.
• Do not let a single unusually large group booking distort a small account's learning. Consider whether an outlier belongs in the signal at all.
• Deduplicate. A confirmation page that can be refreshed, or a thank-you page reachable from an email link, will double-count, and value-based bidding amplifies that error rather than diluting it.

Underneath all of it is the ordinary tracking foundation, which has to be correct before value-based anything can be trusted — one properly configured conversion action per real outcome, primary and secondary conversions separated deliberately, and no accidental double counting. The four-step conversion tracking audit is the fastest way to find out whether that foundation is sound, and it is worth running before any bidding change, not after one goes wrong.

The metric that lies

Here is the specific way a hotel account degrades while its dashboard improves. Total bookings rise. Occupancy holds or improves. The ads account reports more conversions at a lower cost. Everyone concludes the marketing is working. Meanwhile the share of bookings arriving direct is falling, and margin per occupied room is falling with it, because a growing proportion of that rising volume is arriving through a commissioned channel.

Total bookings is the metric that lies here, and it lies in an unusually convincing way, because it moves in the direction everyone wants. A hotel can be busier and poorer at the same time, and no metric inside the ads account will tell you that, because the ads account cannot see the channel mix of the bookings it did not cause.

The second number that catches it is direct share of total bookings, tracked over time alongside revenue per available room or whatever margin measure your property uses. If bookings are up and direct share is down, the growth is being bought from a channel that charges you on every future stay too. That combination should trigger a review of where the demand is actually coming from, not a budget increase.

Read the table as a trade rather than a verdict. The OTA column is not a list of abuses; it is the price of distribution you did not have to build. The point is that the price is ongoing and the direct column's price is not, and that difference is exactly what a bid is buying.

Where hotel accounts break in practice

• The brand campaign becomes the account. It reports the best numbers, so budget migrates toward it, and within a few months most of the spend is buying people who already knew the property's name. Growth stops and nobody can see why, because every metric on the screen improved.
• Location terms are bought at aggregator prices with a page that cannot answer the query. The click costs what an OTA is willing to pay and lands on a homepage with a photo carousel and no rate.
• Calls are assumed to be counted because a call asset is showing. The asset serves, the measurement does not exist in Nepal, and the account optimizes away from the channel producing the enquiries.
• Attribution windows shorter than the booking lead time. Early-funnel campaigns look like they produce nothing, get cut, and volume falls a month later in a way nobody connects to the decision.
• A flat conversion value teaching the bidding to prefer short cheap stays over long valuable ones, while the conversion count rises reassuringly.
• Advertising the season while it is happening rather than while it is being planned, and concluding that paid search does not work for hotels.
• Nobody answering enquiries in the source market's waking hours. A traveler comparing three properties books the one that replied.

The pattern connecting most of those is the same: a metric inside the ads account improved, and a number inside the business did not. When those two disagree, the business number is the one that is right.

When a hotel should not run Google Ads at all

Plainly, because the answer here is no more often than the category's marketing suggests. Work through these before spending anything, and stop at the first clear no.

1. You cannot take a booking on your own site. No engine, no instant confirmation, no mobile checkout. Then paid traffic has nothing to complete and the money is spent introducing people to a channel that cannot serve them. Fix this first, without exception.
2. Your direct rate is worse than your OTA rate. You are paying to send people to the more expensive version of your own product. No campaign survives that.
3. You do not know your commission rate, your repeat rate or your site conversion rate. Without them there is no way to judge whether any click price is acceptable, and in a market with no published benchmarks to fall back on, those internal numbers are the only reference you will have.
4. Nobody answers enquiries promptly, in the hours your guests are awake. Response time destroys more hospitality marketing budget than bidding decisions ever will.
5. Your funding route caps spend below what a test would need. Nepal's foreign-exchange ceilings bind on ad spend abroad, and if your route allows only a token monthly amount, treat it as paid research with a defined question rather than as a channel, and be honest that it will not produce a readable result quickly.
6. You have not yet built the things that cost no media at all — a complete Google Business Profile, which is available in Nepal, a site that loads on a phone, current photography, and reviews. For a great many small properties these return more than a small ads budget, and they compound instead of stopping when the spend stops.

There is also a category answer worth saying out loud. A property whose demand comes overwhelmingly from walk-ins, from trekking agencies, from tour operators or from a long-standing referral relationship is not being held back by an absence of search advertising, and buying clicks will not change the mechanism that fills its rooms. Advertising captures demand that is being expressed as a search. If your guests never express it that way, the channel is wrong regardless of how well it is run.

A starting sequence

The order matters more than the individual steps, because most of these block the ones after them. The first campaign does not launch until step seven, and that is deliberate.

1. Pull your own numbers: actual commission rate from your last statements, repeat rate, direct share of bookings, and average booking value by room type and length of stay. Nothing downstream can be judged without them.
2. Audit the auction. Search your property name and your main location terms from a phone, from a source market, several times over several days, and record who appears and whether it is paid or organic.
3. Fix the booking path before anything else. Instant availability, mobile-completable checkout, immediate confirmation, and a direct rate that is not worse than the OTA rate.
4. Build the measurement layer. Conversion tracking on completed bookings with real transaction values passed through, deduplicated, with a conversion window long enough to contain your actual booking lead time.
5. Solve the call gap deliberately, because Google's native call reporting is unavailable in Nepal. Third-party call tracking, or at minimum a dedicated number on paid landing pages plus logging the source of every enquiry at first contact.
6. Map source markets from your own arrivals data, and set location targeting to reach people interested in your location rather than only people standing in it.
7. Launch narrow: brand as its own campaign, and one tightly scoped non-brand location campaign with a negative list from day one. Nothing else yet.
8. Add remarketing once there is enough traffic to build a meaningful audience, segmented by how far into the booking flow the visitor reached.
9. Run the brand incrementality test once you have a clean baseline and a comparable period, and let the property's own reservation records decide the answer rather than the ads dashboard.
10. Review monthly on direct share of bookings and margin per occupied room, not on conversions and cost per conversion. If those two sets of numbers ever disagree, believe the first.

Where to start

If you take one thing from this: the auction you are in is not a hotel-versus-hotel contest, and the campaign is not the first thing to build. Work out what an OTA booking actually costs you over the life of a guest relationship, using your own commission rate rather than a number from an article. That figure is the ceiling on what a direct booking is worth buying, and it is also, more usefully, the argument that justifies fixing the booking engine — because the same fix pays off on every channel at once, including the free ones.

Then test rather than assert. Whether to defend your own name is an empirical question with a property-specific answer, and the test is cheap compared to years of quietly paying for traffic you already had. The country-level constraints that shape all of this — payment routes, feature availability, what can and cannot be measured from Nepal — are set out in the complete guide to Google Ads in Nepal, and the sequencing method behind this post is documented as the Signal-to-Revenue Framework. I run that practice through Arcetis, and the reason this post spends more time on commissions and booking engines than on bidding is that for a hotel those are the two things that decide whether an account can work at all.

Related reading