Himali GreenHimali Green logoHimali Green

Measured, not claimed

The fastest ecommerce site in Nepal

87 on Google's Lighthouse performance score, against 69 for the next fastest — measured 23 August 2026.

Our Lighthouse score
87Our Lighthouse scorenext best 69
Sites compared
5Sites comparedhead to head
Runs per page
5Runs per pagemedian reported
Less JavaScript
2.8×Less JavaScriptthan the typical rival

Last measured 23 August 2026. Every number on this page is written by the measurement itself — nothing here is typed in by hand, and the claim above changes on its own when a run goes against us.

The comparison

Homepage, on a simulated mid-range Android phone over Slow 4G — the device and connection most Nepali shoppers actually use. We come first on Google’s composite performance score.

  • Himali Green(us)87
  • Daraz Nepal69
  • Muncha.com52
  • Bhatbhateni Online45
  • HamroBazaar42

    Classifieds marketplace rather than a storefront — listed because it is one of the most-visited buy/sell sites in Nepal.

Google's own weighted composite of the five metrics below, 0–100. The standard single number for “how fast is this page”. Higher is better. Bars show the median across the measured runs; the lighter band marks the full range observed.

Largest Contentful Paint

The Core Web Vital Google ranks on, shown separately because it is the metric a reader is most likely to want to check independently — including when it is not the one that flatters us.

  • Himali Green(us)4.0 s
  • Daraz Nepal4.7 s
  • Muncha.com9.3 s
  • Bhatbhateni Online11 s
  • HamroBazaar15 s

    Classifieds marketplace rather than a storefront — listed because it is one of the most-visited buy/sell sites in Nepal.

How long until the biggest piece of content is on screen. A Core Web Vital, and the one Google ranks on. Lower is better. Bars show the median across the measured runs; the lighter band marks the full range observed.

Every number from this run

All 9 metrics, for every site and page type we could measure. Hover any figure for the range across the 5 runs behind it.

Page
Device
MetricHimali GreenDaraz NepalBhatbhateni OnlineMuncha.comHamroBazaar
Lighthouse performance score↑87#169#245#452#342#5
Largest Contentful Paint↓4.0 s#14.7 s#211 s#49.3 s#315 s#5
First Contentful Paint↓1.6 s#11.9 s#25.6 s#54.7 s#44.6 s#3
Speed Index↓2.0 s#15.4 s#27.5 s#45.8 s#313 s#5
Total Blocking Time↓53 ms#1337 ms#472 ms#356 ms#2568 ms#5
Cumulative Layout Shift↓0.000#10.100#30.251#50.160#40.000#1
Page weight↓597 KB#11.7 MB#31.3 MB#23.1 MB#441.7 MB#5
JavaScript shipped↓272 KB#1854 KB#4656 KB#3528 KB#21.7 MB#5
Requests↓40#1142#446#290#3254#5
ReportOpenOpenOpenOpenOpen

Not measured on this page type: SastoDeal (the page did not load for our test browser). These sites are left out of the ranking entirely rather than counted as slow — a page we could not load has not been shown to be anything.

How we measured

Read this before believing the chart. A benchmark is only worth the method behind it, so ours is published in full — including the machine it ran on.

Measured with
Google Lighthouse 13.4.1, headless Chrome, driven by make bench
Runs per page
1 warm-up run discarded, then 5 measured. median of the measured runs; range shown as [min, max]
Cache
cold — a fresh Chrome profile per run, no disk or HTTP cache
Order of measurement
round-robin across sites, so any drift in network conditions during the run is shared equally rather than landing on whichever site was measured last
Simulated mobile device
Mobile (Moto G Power, Slow 4G, 4x CPU slowdown)150ms round-trip · 1638kbps throughput · 4× CPU slowdown
Simulated desktop device
Desktop (1350x940, Dense 4G, no CPU slowdown)40ms round-trip · 10240kbps throughput · 1× CPU slowdown
Machine
Developer machine · Apple M2 · 8 cores · 16 GB RAM · darwin 23.2.0 (arm64)
Measured from
not recorded
Node / Chrome
v26.5.0 · Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) HeadlessChrome/151.0.0.0 Safari/537.36
Code at time of run
0b0bd79ebd

Same pages for everyone

The homepage, a category page and a product page on each site, fixed in advance and listed in full further down this page. No page was chosen after seeing its score, and the list is published so you can check that for yourself.

Interleaved, not batched

We measure one page on every site, then go round again. If the connection wobbles mid-run, every site takes the same hit — measuring one site to completion before starting the next is the easiest way to publish a false result by accident.

Medians, with the spread shown

A single run is noise. We take 5 measurements per page after discarding a warm-up, report the median, and print the full range beside it so you can see how noisy each figure is.

Failures excluded, never counted as slow

A site that blocks our test browser is dropped from the ranking and named at the bottom of the table. Banking an unreachable site as the slowest entry would be the cheapest way to fake a win.

Check it yourself

You do not have to trust our tooling, and you should not have to. Every page in this comparison is a public URL — ours included — measured with a tool that ships inside the browser you are reading this in.

  1. 1. Open any URL below in Chrome, then DevTools → Lighthouse.
  2. 2. Set Mode to Navigation, Device to Mobile, and tick only Performance — the same settings listed under “How we measured” above.
  3. 3. Run it a few times and take the median — one run is noise, which is why we take 5.

The exact pages we measured

Nothing here was picked after seeing its score, and this list is what makes that checkable.

We measure our own site the same way we measure everyone else’s: over the public internet, at www.himaligreen.com, with no local shortcut and no special case. Our own Lighthouse reports are linked from the table above, so you can compare them against whatever you get.

Every run we have published

We re-measure regularly and keep the results, including the runs where we do worse. A number that only ever gets published when it is flattering is a marketing figure, not a measurement.

RunHimali GreenDaraz NepalBhatbhateni OnlineSastoDealMuncha.comHamroBazaar
2026-08-23876945—5242
2026-08-09896344—5145
2026-08-07646046—5349

Lighthouse performance score, homepage, mobile.

What this does not prove

The limits of the claim, stated by us rather than left for someone else to find.

  • It is a snapshot. Every published figure describes the day it was measured, and that date is printed beside every claim. Every site here is under active development, and any of them could be faster than us tomorrow — which is why we re-measure and republish rather than quoting one good afternoon forever.
  • It measures page speed, not the shop. Daraz has a catalogue we cannot match and a logistics network we are years from. Loading quickly is one thing a site can be good at, and it is the only thing this page is about.
  • Bigger sites carry more weight for a reason. Much of what makes a large marketplace slower is functionality we simply do not have yet. The comparison is fair as a description of what a shopper experiences; it is not a judgement of anyone’s engineering.
  • Lab numbers are simulated. Lighthouse models a mid-range phone on Slow 4G rather than renting one. That is what makes the run reproducible, and it is also why the field-data section exists.

Something look wrong? Tell us — if a run is mistaken we would rather fix it than defend it.