This is the dataset behind the audit write-up, published separately so it can be checked, cited and re-run. If you only want the argument, read the article. If you want to verify it or use the numbers, everything is here.
What was measured
On 2026-08-23, the homepage of 34 live Nepali websites across 7 sectors was fetched once over HTTPS with a mobile user agent, following up to five redirects.
Only signals that are objectively present in the HTTP response or the HTML were recorded — status codes, headers, meta tags, heading counts, image attributes, script counts, transfer sizes. Nothing about content quality, keyword choice, backlinks or rankings, because none of that can be established from a single page fetch.
The five-point check referenced throughout is: exactly one <h1>, a canonical tag, a meta description between 50 and 160 characters, any valid structured data, and alt text on every image.
Results
| Check | Result |
|---|---|
| Passed all five basic checks | 3 of 34 (9%) |
| Exactly one <h1> | 12 of 34 (35%) |
| No <h1> at all | 14 of 34 (41%) |
| Canonical tag present | 22 of 34 (65%) |
| Meta description present | 27 of 34 (79%) |
| Description within 50–160 characters | 11 of 34 (32%) |
| Any structured data | 18 of 34 (53%) |
| Broken structured data | 0 of 34 (0%) |
| Local-business schema | 6 of 34 (18%) |
| BreadcrumbList schema | 3 of 34 (9%) |
| Serves over HTTPS | 34 of 34 (100%) |
| Compresses HTML | 34 of 34 (100%) |
| Mobile viewport tag | 33 of 34 (97%) |
| Declares Nepali as page language | 1 of 34 (3%) |
| Runs WordPress | 12 of 34 (35%) |
Images
Across the sample there were 2167 images. Of those, 864 had no alt text and 1584 had no width and height attributes.
Performance
Median time to first byte was 849 ms, ranging from 31 ms to 10367 ms. 13 sites exceeded one second and 5 exceeded 2.5 seconds. Median HTML transfer was 27411 bytes. The worst page carried 21 render-blocking scripts in its <head>.
Declared page language
| <html lang> | Sites |
|---|---|
en | 19 |
en-US | 10 |
en-us | 1 |
en-gb | 1 |
zxx | 1 |
ne | 1 |
zxx is the ISO 639 code for "no linguistic content", almost certainly set by a plugin trying to block automatic translation.
By sector
| Sector | n | One h1 | Canonical | Schema |
|---|---|---|---|---|
| trekking / tourism | 7 | 4 | 5 | 4 |
| health / clinics | 3 | 0 | 1 | 1 |
| hospitality | 2 | 1 | 2 | 2 |
| news / media | 7 | 3 | 4 | 3 |
| finance / commerce | 6 | 2 | 3 | 2 |
| education / institutions | 3 | 0 | 2 | 1 |
| seo & digital agencies | 6 | 2 | 5 | 5 |
Limitations
These matter more than the numbers, and anyone citing this should cite them too.
- It is a convenience sample, not a random one. Sites were taken from public search results by sector. That biases towards sites that already rank, which if anything makes these results flattering.
- 34 sites is small. Treat it as a strong signal about a real pattern, not as a national statistic. It does not support a claim about "all Nepali websites".
- Homepages only. Interior pages were not audited and may differ.
- One moment in time. Server response times in particular vary by hour.
- Two sites dropped out after timing out or failing to respond, and are excluded from every percentage.
The data
- nepal-website-audit-2026.json — every aggregate on this page, machine-readable.
- The audit script — the PHP that produced it. Point it at your own list of sites and it will produce the same measurements.
Per-site results are not published. The finding is the pattern, not which particular business forgot its H1, and a ranked list of named local businesses by failure would be a different and considerably less useful project.
How to cite this
Bhatt, Diwakar. Technical SEO Audit of 34 Nepali Websites. 2026-08-23.
https://diwakarbhatt.com.np/research/nepal-website-seo-audit-2026.html
If you are writing about Nepali web quality, digital readiness or local SEO and want a figure checked, or a specific cut of the data that is not here, email me — I would rather answer than have a number travel without its limitations attached.
Re-running it
The most useful thing anyone could do with this is repeat it: on a bigger sample, on a different sector, or on the same sites in a year to see whether anything moved. The script is published for exactly that. If you run it, I would like to see what you get, including if it contradicts this.
Related reading: the audit write-up, why most Nepali business sites stay invisible, and SEO in Nepal in 2026.