Why published benchmarks mislead here
A cost-per-lead figure is the product of an auction, an audience and an offer. Change the country and all three change: fewer advertisers bidding, cheaper clicks, smaller audiences, different average order values and a very different mix of devices and payment behaviour. A US benchmark of USD 40 per lead for legal services tells a Kathmandu law firm nothing useful, and a Nepali benchmark averaged across dentistry and real estate tells nobody anything at all.
The second problem is measurement. Most Nepali accounts under-report conversions because phone calls and WhatsApp messages are not tracked, so any “average CPL in Nepal” figure collected across accounts is averaging numbers that were measured differently. That is why this page does not print a table of industry averages: I would be inventing precision I do not have.
Building your own benchmark, in four numbers
- Cost per click, from your own account over the last 90 days, split by campaign rather than averaged.
- Conversion rate, with every genuine enquiry counted: form, call, WhatsApp click. If calls are untracked, fix that before trusting anything else.
- Cost per lead, which follows from the first two.
- Maximum affordable cost per lead, from your close rate and the profit on a customer. This is the only benchmark that matters, because it tells you whether to scale or stop.
The CPL calculator does the arithmetic, including the maximum you can afford. Once you have those four numbers for one month, the following month has a benchmark: your own.
The one result I can publish
A fertility clinic in Kathmandu ran Meta Ads for IVF enquiries and produced more than 150 qualified leads at roughly NPR 180 per lead, using doctor-led video rather than stock imagery, Nepali-language creative and a lead form built for phones with WhatsApp follow-up.
Read that as one campaign in one category with one creative approach, not as a benchmark. Healthcare in Nepal responds unusually well to a named doctor on camera; the same cost per lead would be optimistic for real estate and pessimistic for a low-value consumer product. The full write-up, including the creative and the targeting, is in the Angel Fertility Clinic case study.
Patterns that hold, without numbers attached
These are directional observations from client accounts rather than measured averages, and they are worth knowing while you gather your own data.
- Clicks in Nepal cost a fraction of the same keyword in Australia or the US, and lead volumes are correspondingly smaller. Budget is rarely the constraint; audience size is.
- Meta usually produces cheaper leads than Google in Nepal, and Google’s leads usually close at a higher rate, because search intent is stronger. Judge the two on cost per customer, not cost per lead.
- Untracked phone and WhatsApp enquiries are the single largest source of apparent underperformance. Accounts routinely look far more expensive than they really are.
- Lead quality varies more than lead cost. A cheap lead form with no qualifying question produces volume that the sales team cannot use.
- Seasonality is sharp around Dashain and Tihar, in both directions depending on the category, so month-on-month comparisons across those weeks mislead.
Next steps
Get the four numbers for your own account, then use them. If tracking is the gap, the conversion tracking setup and the Pixel and Conversions API setup are where to start; if you want the account reviewed first, the free Google Ads audit and free Meta Ads audit both report cost per enquiry as measured rather than as reported.
Updated