There are now a dozen articles explaining what SEO costs in Nepal. I have read most of them and they broadly agree: somewhere between NPR 10,000 and NPR 100,000 a month, more for bigger sites. What none of them answer is the question the business owner is actually asking, which is not what should I pay but did I get anything for what I already paid.
That question comes up often enough that I want to write down how I answer it. Everything below is something you can check yourself in an afternoon, without access to anything your agency has not already given you, and without knowing any SEO.
First: the three numbers that always go up
Before the checks, the counter-checks. Three metrics appear in almost every monthly report and none of them, on their own, tell you anything.
Keywords ranked. "You now rank for 1,240 keywords." A site ranks for a keyword if it appears anywhere in the top hundred results. Publish a page and you will "rank for" dozens of phrases nobody searches, at positions nobody sees. The number only rises. It cannot fall unless the site is actively shrinking.
Impressions. Impressions rise when Google shows you to more people, including people who scroll straight past. Impressions can double while clicks stay flat, which usually means you started appearing for queries that are not yours. That is a diagnosis, not an achievement.
Backlinks acquired. A count with no quality attached. Twenty links from directories nobody reads is not better than one link from a publication your customers actually visit — it is worse, because the first kind occasionally attracts a penalty.
None of these are lies. They are just numbers chosen because they move in the right direction on their own. Ask about them and you will usually get a straight answer; the problem is only that they are presented as the result rather than as context.
The seven checks
1. Ask for the list of pages that changed
This is the single most revealing question, and it is not technical. Ask: which specific URLs did you change this month, and what did you change on them?
Real SEO work leaves a trail. Titles rewritten, pages merged, redirects added, sections rewritten, a new page published. If the answer is a list of URLs and edits, the work happened. If the answer is a paragraph about ongoing optimisation, ask again more specifically. I have reviewed engagements where the honest answer to that question, across six months, was nothing.
2. Look at clicks, not rankings — and look at them by page
Open Search Console yourself. If you do not have access, that is the first thing to fix, because it is your data on your property and there is no reason for it to live only in an agency account.
Go to Performance, set the range to the last six months, and switch to the Pages tab. You are looking for whether any specific page has more clicks than it did before the work started. Site-wide totals hide too much — a seasonal bump on one page can mask five pages going nowhere.
3. Check whether the money pages moved
Most businesses have three or four pages that actually earn: the service page, the pricing page, the main product category. Everything else is support.
Rankings and clicks on a blog post are pleasant. Rankings on the pages that lead to enquiries are the job. If six months of work has produced traffic growth entirely on informational articles while the service pages sit exactly where they were, the strategy is either wrong or has not reached them yet — and either way that is the conversation to have. I wrote about this gap between traffic and customers in more detail here.
4. Search for yourself the way a customer would
Open a private window and type what a customer would type. Not your business name — you will rank for that regardless, and ranking first for your own name is not SEO, it is arithmetic. Type the thing they need.
If you cannot find yourself in the first two pages for the query that describes your core service in your city, then whatever else has improved, the main problem has not been solved.
5. Check that the technical basics are actually fixed
Almost every audit promises technical fixes. Whether they were applied is checkable in about two minutes:
- View source on your homepage and search for
<h1. There should be exactly one. - Search the same source for
canonical. There should be one, and it should point at the page you are on. - Type a URL that does not exist on your domain. It should return a real 404, not a page that looks like an error while quietly returning success — a failure I have written about at length.
- Open your site on a phone on mobile data, not office wifi. Time it. If it takes more than about three seconds to show anything, that is a finding, and usually a fixable one.
When I audited 34 Nepali business websites, four in ten had no H1 at all. These are not obscure checks. They are the first things anyone competent looks at, which makes them a fair proxy for whether anyone looked.
6. Read one piece of the content out loud
If content is part of the engagement, read a paragraph aloud. You will know within two sentences whether a person wrote it for your customers or whether it was assembled to contain a keyword. Your customers will know too.
Watch for the same phrase appearing in every article, for paragraphs that restate the heading, and for text that would work equally well for a business in another industry. Content that could belong to anyone belongs to no one.
7. Ask what they would stop doing
This is the question I would ask if I could only ask one. Anybody can produce a list of things to do. Knowing what to stop — which pages to merge, which to delete, which keyword to abandon because the results page shows it is unwinnable — requires having actually looked at your site.
An answer here also tells you whether you are being sold volume. Publishing more is the easiest thing to bill for and often the wrong prescription; I have argued that case in the SEO content trap.
How long before any of this should show
Technical fixes can show in days to a few weeks. Content and authority work realistically takes three to six months before the trend is readable, longer in a competitive category.
That is a real constraint, not an excuse, and the difference is visible. A legitimate three-month answer comes with evidence of work: pages changed, problems found, things fixed. An excuse is a three-month answer with no trail behind it. If you cannot see what was done, the timeline is not the issue.
What a good report actually looks like
Short. Specific. It says what was done, what it was expected to affect, and whether it did. It names pages. It includes at least one thing that did not work, because across a quarter of real work something always does not.
If your report is mostly graphs that go up and to the right, ask for the list of URLs that changed. It is a fair question, it is easy to answer if the work happened, and how it is received tells you most of what you need to know.
If you would like a second opinion on work you are already paying for, I do that as a one-off audit — no retainer attached. More about how I work, or send me the domain and I will tell you what I find.