Case brief: Education consultancy Search ads
Best for: Week 4 structure & landing pages · Week 2 CRO of paid LPs
Source of truth: /case-studies/nepal-education-consultancy-ppc/ · Narrative: /blog/education-consultancy-ppc-case-study-nepal/
Situation
Kathmandu study-abroad consultancy. One messy campaign, mixed intent, homepage landings, incomplete call tracking. Spend held constant at NPR 300,000/month for six months.
Numbers (as published)
| Metric | Before | After |
|---|---|---|
| Leads/mo | 85–90 | 180–200 (~2×) |
| CPL | NPR 3,500 | NPR 2,100 (−40%) |
| Landing CVR | 2.1% | 6.8% |
| Enrollment CAC | NPR 28,000 | NPR 12,000 (−57%) |
| Reported ROAS | 1.8× | 5.2× |
| Irrelevant spend | ~35% | <8% |
What they did
- Rebuild by destination country + intent
- Phrase/exact + 300+ negatives
- Dedicated country landing pages
- Call tracking → then tCPA / RLSA when data allowed
MBA discussion
- Same budget, better outcomes — which lever moved CVR most (structure, LP, or bidding)?
- How would you set LTV for an enrollment business before setting Target CPA?
- Compare to the home-services Ads case: shared pattern is negatives + themed structure + matching LP.
Concepts in Play
- Search intent: what the searcher wants: research (“study in Australia cost”) or action (“Australia student visa consultancy Kathmandu”).
- Negative keywords: terms that stop ads showing for irrelevant searches (e.g. “jobs”, “free”, “scholarship form”).
- Cost per lead (CPL): ad spend divided by leads; enrolment CAC divides spend by students who actually enrol.
- Smart bidding (tCPA/RLSA): automated bidding towards a target cost per acquisition, and adjusting bids for past visitors. Both need reliable conversion data first.
Reading the Numbers
Work these from the published figures only; do not add new KPIs.
- Check the consistency of the published figures: at NPR 300,000 a month, what CPL would 85–90 leads imply, and what would 180–200 leads imply? Compare with the published CPLs and suggest why they might not reconcile exactly (for example, which leads and which spend are counted).
- Landing-page conversion rose from 2.1% to 6.8%. If clicks had stayed the same, roughly how many times more leads would that alone produce?
- Irrelevant spend fell from ~35% to under 8%. Of NPR 300,000, how much was being wasted before, and how much after?
Limits of the Evidence
- Several changes were made together (structure, negatives, landing pages, tracking, bidding), so the case cannot say which one caused how much of the gain.
- Call tracking was incomplete at the start, so part of the “before” leads may have been undercounted. That makes the improvement look larger than it was.
- Reported ROAS depends on the value assigned to an enrolment; ask what value was used before quoting 5.2×.
Apply It to Your Sandbox
- For your sandbox, write two ad groups by intent (one research, one ready-to-act) with five keywords each.
- List ten negative keywords that would waste your budget.
- Specify the landing page for each ad group. It must not be the home page.
- State the conversion you will track and how (form, call, WhatsApp click).
Cite in pitches
“Education PPC case (Kathmandu): CPL NPR 3,500 → 2,100 (−40%) and ~2× leads on the same NPR 300k/month (source).”
