Kick Ads

Industry Insights

Education leads need time and follow-up.

Students and parents rarely decide from one ad. They see you first, search for you weeks later, and are still comparing after they have enquired. The advertising has to be there at each of those moments, and so does the follow-up.

What Makes This Industry Different

CPL does not tell the whole story.

An education lead may enquire months before making a decision. Students and parents compare destinations, institutions, courses, fees and events before progressing.

A cheap lead is therefore not automatically a useful lead. The account needs enough follow-up data to understand which campaigns are actually producing conversations and applications.

From the Accounts

What the education accounts we manage actually show

85%

Of one account's conversions were arriving on the contact page

1.0%

Were somebody actually trying to reach the school

3.5% → 82%

Conversion rate range inside one account, on one action

2x

Rise in non-brand click cost from June to late August

00Headline finding

In two education accounts, most of what was being counted as a lead was a page view.

85.0%Account 01: arriving on the contact page
1.0%Account 01: tapping WhatsApp
79.1%Account 02: five minutes on the site
14.0%Account 02: registering or submitting a form
Share of each account's counted conversions.

In two of the education accounts we took over, most of what the account counted was not an enquiry. In one, the action being optimised towards was arriving on the contact page, which was 85% of its counted conversions. The one action that records somebody actually trying to reach the school, a tap on WhatsApp, was 1.0%. In the other account, four in five counted conversions were a single event: spending five minutes on the site. Real registrations and form submissions came to 14%.

What we do with it

We check what the conversion action fires on before reading a single cost-per-lead figure, and we rebuild it where it fires on a page view. Until that is fixed, every budget decision is being made on the wrong number, and the campaigns that look best are usually just the ones sending the most traffic to that page.

Limitations

Two accounts, and one of them is an NGO running skills training rather than a fee-paying school. A third account in the same population runs clean conversion actions, so this is two of three rather than a rule. The 39 WhatsApp taps are themselves taps, not conversations, so the true enquiry count is lower still.

Two Hong Kong education accounts, two separate companies · 17 Jun to 26 Aug 2026 · shares only, no currency

01A check you can run yourself

One conversion action reported 3.5% in one campaign and 82% in another.

In the same tutoring account, the Performance Max campaign reported a conversion rate of 81.8%, two Search campaigns reported 45.4% and 39.7%, and three more reported 3.5%, 3.6% and 4.6%. All of them on the same conversion action.

What we do with it

A conversion rate that varies that far inside one account, on one action, is the signature of a goal that fires when somebody arrives at a URL. It takes a minute to check on your own account, and it is the first thing we look at.

One Hong Kong tutoring account, Google · conversions to 26 Aug 2026

02The enrolment window

A non-brand click cost twice as much in late August as it did in early June.

Where the quarter started
100

Week of 1 Jun

100

Week of 22 Jun

97

Week of 13 Jul

97

Week of 3 Aug

155

Last full week of Aug

204

In the tutoring account we manage, the cost of a non-brand click roughly doubled across the quarter, from the first week of June to the last full week of August. Impressions on the same campaigns peaked in mid-July and then fell by about 40%, so the account was paying twice as much for a click while there was less to buy. Spend on those campaigns rose 31% and bought 38% fewer clicks.

What we do with it

We plan the tutoring budget against this curve instead of holding a flat monthly figure, and we set expectations on cost per enquiry before the enrolment window rather than explaining the rise afterwards.

Limitations

One account, Hong Kong, Google, non-brand Search only. No conversion figure is involved in this one, so the way this data captures conversions does not affect it.

One Hong Kong tutoring account, HKD, non-brand Search · 1 Jun to 1 Sep 2026

How these figures were produced

Based on anonymised Google Ads and Meta Ads accounts managed by Kick Ads, Hong Kong education, mid-2026. Where two accounts run in different currencies, only shares are used. Conversion figures are shares between actions counted the same way, because conversion rows in this data are captured early and understate the level.

At a Glance

Kick Ads manages education paid media across Hong Kong, Malaysia and Taiwan, covering Google Search, Demand Gen, YouTube and Meta Ads. Our work can include course and programme campaign structure, event and seminar tracking, landing pages, and lead reporting against CRM feedback.

Diagnostics

What we check first

01

Search demand

Destination, institution, course, service and brand searches

02

Lead source

Search, Meta, Demand Gen, YouTube and event campaigns

03

Conversion flow

Landing page, WhatsApp, forms, seminar registration and consultation booking

04

CRM feedback

Whether the leads progress into conversations, appointments or applications

Channels

The campaign mix we usually run for education

Before they search

Demand Gen for events and seminars

Core

An event gives the campaign something specific to offer and a clear registration action, which is what makes this format work here

Meta Awareness

Support

Parents and students usually meet the brand on social before they ever search for it

While they are searching

Brand Search

Core

Catches the searches the awareness layer creates

Programme and destination Search

Core

Queries naming a course, a country or an exam are the closest thing to a ready enquiry in this industry

While they are comparing

Remarketing to registrants

Support

Someone who registered for a seminar has not yet applied, and the gap between the two is where the follow-up happens

What the results tell us

CRM feedback

Lead volume can rise long before applications do, so the platform's conversion count is not the number to steer on

Informs the next budget decision

The calendar

The education calendar we plan against

Education demand is driven by term dates and exam results, and the peaks are short. A week late on a peak is most of the peak missed.

  1. Jan to Feb

    Second term starts, and mock exams put pressure on students who are behind.

    We raise budget on subject-specific tutoring here, because the parent is reacting to a specific result rather than shopping generally.

  2. Mar to May

    The run-up to public exams. Intent is at its highest and so is competition.

    We expect cost per click to rise and plan for it, rather than reading the rise as something going wrong in the account.

  3. Jul

    Results day. The single sharpest demand spike in the industry, and it lasts days rather than weeks.

    Everything has to be live and approved before results day. An ad in review on the morning of results has missed it.

  4. Aug

    Enrolment for the new school year, and non-brand click costs run at their highest of the quarter.

    We set expectations on cost per enquiry before this window opens rather than explaining the rise afterwards.

  5. Sep to Nov

    Term settles. Demand drops to its floor for the year.

    We use the quiet stretch to rebuild tracking and landing pages, which is work nobody has time for during a peak.

This calendar comes from running these accounts, not from a measurement. Chinese New Year moves between January and February, so the dates around it shift each year.

One client

An education provider · Hong Kong, Malaysia and Taiwan

The problem

Enquiries were expensive and the always-on campaigns had nothing specific to offer. Someone who was still deciding had no reason to act now rather than in three months.

What we changed

We built the campaigns around the provider's own events instead of around the service in general. An event gives the campaign three things a general service page cannot: a date, so there is a reason to act now rather than in three months; something specific being offered, so the ad has content rather than a claim; and a registration action, which is a smaller thing to ask for than an enquiry. Registrations were tracked separately from general enquiries so the two were never averaged together.

What moved

−92%Cost per lead, through events
3Markets run from Hong Kong
8 yrPartnership

Common problems

  • Optimising only to cheap Lead Forms
  • No CRM feedback
  • All markets using the same campaign structure
  • No separation by destination or course intent
  • Event registrations mixed with standard enquiries
  • Awareness judged only on direct conversions
  • Weak remarketing between initial research and application

Results

Selected education results

View all results

Overseas Education · HK, MY & TW

8-year
Hong Kong partnership

Paid media support across three markets, Lead Generation and awareness

Where figures are shown, they are based on anonymised campaign data reviewed by Kick Ads.

FAQ

Common questions

Want to know which education campaigns bring useful enquiries?

We can review Search demand, campaigns, event tracking and CRM feedback to understand which areas deserve more budget.

Talk to us