The complete SEO guide for Nepali businesses.

Nepal’s commercial web is small and shallow. In most Western markets, ranking for “accountant” means fighting past a decade of established sites, directories, and content farms. In Nepal, the first page for a great many commercial queries is still a mix of Facebook pages, a Wix site from 2019, and two listings nobody maintains. That is not an insult to the market, it is an opportunity: the bar to reach page one is genuinely low, and it will not stay low.

It also means the shape of the opportunity is local rather than national. Very few people in Nepal search for “interior designer”. They search for “interior designer in Lalitpur”, or they search for “interior designer” on a phone in Lalitpur and Google adds the location for them. Those two paths land on the same result set, and both are won mostly by signals that have nothing to do with your website’s copy.

For any business with a physical address or a service area, the map pack sits above the normal results. It is also the one ranking surface you control almost entirely through data entry rather than through competing on authority. We have watched businesses go from invisible to receiving direct calls inside three weeks on this alone.

What to fill in, in order

1. 1Claim the listing at business.google.com and complete the postcard or phone verification. Do not skip this because an unverified listing exists: an unclaimed listing can be edited by anyone.
2. 2Set the primary category precisely. “Web designer” and “Advertising agency” surface for very different queries, and the primary category carries far more weight than the secondary ones.
3. 3Write the full address the way Nepal Post and the locals both write it, including the tole or chowk. Consistency with how the address appears elsewhere online matters more than any particular format.
4. 4Add a local phone number that a person answers. A Viber-only number or a number that rings out will cost you the enquiry after the ranking has already done its job.
5. 5Upload at least ten real photographs: the storefront from the street, the interior, the team, the work. Stock imagery is obvious and it suppresses engagement.
6. 6Fill in opening hours and keep them accurate through Dashain and Tihar. Google surfaces holiday-hour prompts, and answering them is a small freshness signal.
7. 7Add your services or products as individual entries rather than one paragraph. Each becomes matchable text.

Reviews, honestly

Review count and recency both feed local ranking, and review text is matched against queries. A review that says “they rebuilt our restaurant’s website and set up online ordering” helps you rank for restaurant website work in a way that your own copy cannot. Ask every satisfied client, in person or over a message, with a direct link to the review form. Never buy reviews: bought reviews arrive in clusters from unrelated accounts, Google is good at spotting the pattern, and the recovery from a filtered listing costs more than the reviews were worth.

Reply to all of them, including the bad ones, without arguing. A calm, specific reply to a one-star review reads well to the next hundred people who see it, which is the actual audience.

This is the part generic SEO advice gets wrong about Nepal. Your customers search across three overlapping registers, and they use different ones for different intents.

Most Nepali businesses optimise only for the first row, then wonder why their traffic is thin. The second and third rows have far less competition, and Romanized Nepali in particular is nearly unclaimed for most trades. You do not need a separate Nepali site to capture it. A single naturally written paragraph using the phrasing your customers actually use, plus an FAQ answered in that phrasing, is usually enough.

Finding the phrases without paying for a tool

• Type the start of a query into Google on a phone in Nepal and read the autocomplete suggestions. Those are real query volumes, not estimates.
• Scroll to “People also ask” and “Related searches”. Each of those is a section heading you could write.
• Read the questions people actually ask you on Facebook, Viber, and the phone, then use their exact wording. Nobody asks for a “bespoke digital solution”.
• Open Search Console’s Queries tab, sort by impressions, and look for anything with impressions and no clicks. Those are queries where Google already thinks you are relevant and your title is not winning the click.
• Google Keyword Planner is free with any Ads account and its Nepal data, while coarse, is better than guessing.

Technical SEO does not make a bad site rank. It stops a good site from being held back, which is a different and more modest claim than most agencies make. The list is short and finite, and once it is done it stays done.

• One canonical hostname. Pick either example.com or www.example.com and 301 the other to it. Serving both at status 200 splits your signals across two sites, and it is the single most common misconfiguration we find.
• HTTPS everywhere, with no mixed-content warnings.
• A unique title and meta description on every page. Not a template with the business name repeated: the title is the advert, and it decides your click-through rate at whatever position you have earned.
• One H1 per page, describing that page rather than the company.
• A sitemap.xml submitted in Search Console, and a robots.txt that points to it.
• Real HTML that exists before JavaScript runs. If your site renders its content client-side only, assume the crawler sees an empty shell.
• Structured data for what you are: LocalBusiness or ProfessionalService with your address and geo coordinates, plus FAQPage where you genuinely have FAQs.
• No 404s on URLs that used to rank. When you redesign, map every old URL to its closest replacement with a 301. Deleting a page that had traffic throws away everything it had earned.

Speed, measured on the phone your customer owns

A site that scores well on a fibre connection in an office in Kathmandu can be unusable on 4G in Nepalgunj. Test on a mid-range Android device on mobile data, not on your laptop. The usual culprits, in the order they usually matter: uncompressed hero images, three separate font families, an embedded map or video loading on first paint, and a page-builder plugin stack shipping several hundred kilobytes of JavaScript to render text.

Serve images as WebP or AVIF at the size they are actually displayed. Load anything below the fold lazily. Self-host or subset your fonts. Those four changes usually take a site from a failing Core Web Vitals assessment to a passing one without touching the design.

The most common structural mistake we see is a single Services page listing eight services in eight sentences. That page competes with itself. It cannot rank for “branding agency in Nepal” and “custom software development Nepal” at once, because it does not say enough about either to be the best answer to either.

Split it. One page per service, each one answering the questions a buyer actually has: what exactly you deliver, what it costs or how pricing works, how long it takes, what you need from them, and proof that you have done it before. That is not padding, it is the information the enquiry email would have asked for anyway.

What a page that earns its ranking contains

• A specific H1 that names the service and the place, not a slogan.
• An opening paragraph that would make sense to someone who arrived from a search result and knows nothing about you.
• The deliverables, itemised. Vagueness reads as inexperience to buyers and gives Google nothing to match.
• Honest pricing signals. Even a range, or “projects typically start at”, outperforms silence, and it filters out enquiries you did not want.
• Three to six real questions with real answers, marked up as FAQPage.
• Links to work you have done in that service, and links to your related services.

Link building advice written for the US market is mostly useless here, because the tactics assume a dense ecosystem of niche blogs and journalists taking pitches. Nepal has a thinner web, but it has its own routes, and they are less contested.

• Local directories and chambers: your municipality’s business listing, the Federation of Nepalese Chambers of Commerce and Industry, your district chamber, and industry associations. These are citations more than links, and consistency of your name, address, and phone across them is what feeds local ranking.
• Nepali news and trade press. Tech and business outlets here run genuinely useful contributed pieces. A single data-backed article about your sector, offered rather than pitched aggressively, is worth more than fifty directory submissions.
• Universities and colleges. Guest lectures, internship programmes, and sponsored student projects produce .edu.np links that are hard for competitors to replicate.
• Your own clients. If you built it, ask for a credit link in the footer. Most will agree, and it is the most natural relevance signal available to an agency.
• Events and sponsorships. Hackathons, startup weekends, and local meetups almost always list sponsors with a link.
• Suppliers and partners. Many have a “where to buy” or “our partners” page and simply have not thought to add you.

Avoid the two things sold most aggressively in this market: bulk directory submissions and paid guest posts on link farms. Both are visible to Google, both are worthless at best, and the second is an active liability.

Search Console is the only free source of truth about how you appear in Google, and four numbers on the Performance report tell you almost everything. What matters is which of them is broken, because each one points at a different fix.

Two habits worth forming. First, read the Queries tab more often than the totals: the totals move slowly and tell you nothing actionable, while the queries tell you what Google currently thinks you are. Second, check the Pages report under Indexing for anything excluded. A page that is not indexed cannot rank, and the reason given is usually specific and fixable.

Set your expectations against the timescale rather than the effort. Technical fixes can show up within a fortnight. A new page typically takes one to three months to settle at its real position. Local ranking changes from a Business Profile can move in weeks. Nothing about search rewards impatience, and the businesses that win here are simply the ones still doing it in month nine.

• Redesigning the site without redirecting the old URLs. This is the most expensive mistake on the list and the easiest to avoid. Every page that had traffic needs a 301 to its replacement before launch, not after.
• Serving the site on both www and non-www with no redirect, so every signal is split in half.
• Keyword stuffing the footer with “web design Nepal Kathmandu Pokhara Biratnagar Chitwan”. It has not worked since roughly 2011 and it makes the site look cheap.
• One Services page trying to rank for eight services.
• Hiding the phone number and address behind a contact form. For local search these are ranking signals, not just conveniences.
• Blocking the site in robots.txt after a staging deployment and never noticing.
• Buying reviews, buying links, or hiring anyone who guarantees a number-one ranking. Nobody can guarantee that, and the only way to appear to is with tactics that eventually cost you the site.