Plenty of businesses have a website, a monthly SEO invoice, and no ability to answer a simple question: is any of this producing customers?
Usually because the only numbers anyone looks at are visits and rankings, and neither tells you whether the site is doing its job.
The numbers that mislead
Total visits. Traffic without qualification means nothing. A thousand visitors looking for something you do not sell is worse than forty who wanted exactly what you offer, because the thousand cost you money to attract.
Keyword rankings on their own. Position three for a term nobody searches is a screenshot, not a result. Rankings matter only through the impressions and clicks they produce.
Bounce rate, as usually interpreted. A visitor reading your whole article, getting the answer and leaving is a success. The metric cannot tell that apart from someone leaving because the page was useless.
Time on page, for similar reasons. Longer is not automatically better — it can mean your pricing page is confusing.
None of these are useless. They are diagnostics. The problem is treating them as outcomes.
One more framing worth holding on to: if traffic is healthy but enquiries are not, the problem is usually the writing rather than the ranking — see why service websites get traffic but no enquiries.
What to measure instead
| Question | Metric | Where |
|---|---|---|
| Are people finding us? | Impressions, by query | Search Console |
| Are they choosing us? | Click-through rate | Search Console |
| Are the right people arriving? | Non-branded clicks | Search Console |
| Do they do anything? | Conversions (form, call, email) | GA4 |
| Which pages produce enquiries? | Conversions by landing page | GA4 |
| Is it worth continuing? | Enquiries that became customers | Your own records |
The last row is the only one that is actually revenue, and it is the one no analytics tool can give you. It requires asking new customers how they found you, and writing the answer down.
Branded versus non-branded — the split that matters most
If most of your search traffic comes from people typing your company name, your SEO is not working. Those people already knew you existed; search was just the route.
Split your Search Console queries into branded (containing your name) and non-branded (everything else). Non-branded clicks are the ones representing new demand. Growth there is the clearest evidence that SEO is doing something.
This single split reframes a lot of reports. A “traffic is up 40%” month that is entirely branded usually means your offline marketing worked.
Set up conversion tracking properly
Most sites either have no analytics or have GA4 installed and measuring nothing but pageviews. The second is more common and arguably worse, because it looks like measurement.
At minimum, track:
- Contact form submissions — fired on actual success, not on button click
- Phone number taps on mobile
- Email address clicks
- WhatsApp or Viber clicks, which in Nepal are often the primary contact route
That last one gets missed constantly. If most enquiries arrive via WhatsApp and you only track form submissions, your data says the site produces almost nothing while your phone says otherwise.
One caution on form tracking: fire the event when the submission genuinely succeeds. Tracking the click counts failed submissions as conversions and quietly inflates everything.
Attribution is imperfect. Use it anyway.
Analytics will never fully explain how customers found you. Someone reads an article, forgets it, sees you on Instagram, searches your name three weeks later and calls. That is recorded as a branded search, and it credits the wrong thing.
The practical response is not a more elaborate tracking setup. It is asking. Add one question to your enquiry form or your first call — “how did you come across us?” — and record the answer. Imperfect self-reported data across fifty customers tells you more about what works than any attribution model will.
A review rhythm that is actually sustainable
Monthly, 20 minutes: impressions, clicks, non-branded share, conversions, top landing pages. Note anything that moved sharply and why.
Quarterly, longer: which pages produce enquiries, which produce nothing, what to improve or retire. Compare against the same quarter last year rather than last month, since most businesses are seasonal.
Annually: is organic search worth what you spend on it, against every other channel?
The failure mode is checking dashboards weekly, seeing noise, and reacting to it. Search data is noisy at short intervals. Monthly is frequent enough to catch a real problem and infrequent enough to avoid chasing randomness.
What “working” looks like at each stage
Expectations matter, because judging month two by month twelve’s standard leads people to abandon work that was on track.
- Months 1–2: pages indexed, impressions appearing, technical problems resolved. Almost no traffic. This is normal.
- Months 3–4: impressions growing, first non-branded clicks, some long-tail rankings.
- Months 5–8: clicks growing, first enquiries attributable to organic search, commercial terms entering the top 20.
- Months 9–12: a compounding pattern where new content ranks faster than early content did.
If you are at month eight with no impression growth, something is wrong and it is worth investigating rather than waiting. See why isn’t my website showing on Google.
Frequently asked questions
What is the most important website metric?
Enquiries that turned into customers. Everything else is a step toward it. If you can only track one thing, track that — even in a spreadsheet.
Do I need Google Analytics and Search Console?
Yes, both, and they answer different questions. Search Console covers what happens in Google before the click — impressions, queries, rankings. GA4 covers what happens after. Neither replaces the other, and both are free.
How do I track WhatsApp enquiries?
Track clicks on your WhatsApp link as a conversion event. That captures intent from the website. For volume, use a business number and ask new contacts where they found you.
Why did my traffic drop suddenly?
Check Search Console for manual actions and security issues first, then what changed on your side — a migration, a redesign, a plugin update, a robots.txt edit. Sudden total drops are usually self-inflicted or a hack; gradual decline is more often competitive.
How long before SEO shows measurable results?
Impressions can move within weeks, particularly after technical fixes. Meaningful click and enquiry growth on commercial terms typically takes four to eight months. Anyone promising faster is targeting terms nobody searches.
Should I pay for rank-tracking software?
Not initially. Search Console gives you average position for every query you actually appear for, which is enough for most businesses. Paid tools become useful when you are tracking competitors or managing many sites.
The simplest place to start
If you have no measurement at all, do two things: verify Search Console, and set up GA4 with your contact form and WhatsApp link as conversions. That is an afternoon, it costs nothing, and it turns “the website doesn’t seem to do much” into a question you can answer.
If you would rather have it set up and reported on properly, our SEO packages include monthly reporting against these numbers rather than a screenshot of rankings, and how much SEO costs in Nepal covers what that work costs here.
