Blog post: Juliano Calixto • 12/08/2026
How students choose a language school in 2026
How students choose a language school in 2026
The shortlist now gets built before anyone visits your website, and half the decisions happen in places you have no analytics for. Here's the journey as it actually runs, and what each step demands from your site.
The shortlist now gets built before anyone visits your website, and half the decisions happen in places you have no analytics for. Here's the journey as it actually runs, and what each step demands from your site.
Ask a school director how students find them and you'll usually hear a version of the 2015 answer: Google, agents, word of mouth, maybe Instagram. None of that is wrong. It's just no longer the whole path, and the parts that changed are the parts that decide whether you make the shortlist. Here's the journey as it runs now, step by step.
Step 1: The question gets asked, and often answered, before your site is in play
The first move used to be a search that produced ten links. Now it frequently produces an answer.
Google AI Overviews appeared on roughly 48% of queries as of March 2026, up from about 31% a year earlier, according to Advanced Web Ranking data compiled by Digital Applied. When one appears, organic click through rates drop somewhere between 34% and 61% depending on the study. The link is still there. Fewer people need it.
Meanwhile the students themselves moved. ICEF Monitor reports that use of ChatGPT for college search doubled between 2024 and 2025, that education is among the most prominent categories for website traffic referred by ChatGPT, and that prospective students increasingly use AI tools to compare and evaluate study options before they contact anyone.
What this demands from your site: content a machine can quote. Direct questions with direct answers, specific facts instead of adjectives, pages that resolve a doubt in a paragraph rather than burying it in a brochure. If your site can't be summarized accurately, it won't be summarized at all.
Step 2: The shortlist gets built by something that read your website without visiting it
This is the step that breaks the old mental model. A student asking an AI assistant to compare three language schools in Dublin is running a comparison based on whatever those assistants can extract from each school's site and from third party sources talking about them.
You're being evaluated by a reader who never fills in your form, never triggers your analytics, and never sees your homepage hero video. That reader is deciding whether you belong on the list.
ICEF's own advice to schools is blunt about the first move: find out whether AI tools mention your school at all, and what they say when they do. Most directors have never checked. It takes five minutes and the answer is usually uncomfortable.
What this demands from your site: clarity over cleverness. Course names that describe the course. Levels, durations, prices and start dates written as facts, not as "contact us for details". Everything an assistant would need to recommend you accurately.
This is the step that breaks the old mental model. A student asking an AI assistant to compare three language schools in Dublin is running a comparison based on whatever those assistants can extract from each school's site and from third party sources talking about them.
You're being evaluated by a reader who never fills in your form, never triggers your analytics, and never sees your homepage hero video. That reader is deciding whether you belong on the list. ICEF's own advice to schools is blunt about the first move: find out whether AI tools mention your school at all, and what they say when they do. Most directors have never checked. It takes five minutes and the answer is usually uncomfortable.
What this demands from your site: clarity over cleverness. Course names that describe the course. Levels, durations, prices and start dates written as facts, not as "contact us for details". Everything an assistant would need to recommend you accurately.
Step 3: The site visit is a verification, not a discovery
By the time someone reaches your website, they usually arrive with a hypothesis: this school might be right for me. The visit exists to confirm or kill it.
That flips what the page is for. It isn't there to impress, it's there to answer, fast, on a phone, in the language the visitor actually reads. Most school sites still open with the institution talking about itself and hide the answers three clicks down. The visitor doesn't dig. There's another tab open.
What this demands from your site: the answers above the fold on the pages that matter, which are almost never the homepage. Course pages are where enrollment decisions get made.
By the time someone reaches your website, they usually arrive with a hypothesis: this school might be right for me.
The visit exists to confirm or kill it.That flips what the page is for. It isn't there to impress, it's there to answer, fast, on a phone, in the language the visitor actually reads. Most school sites still open with the institution talking about itself and hide the answers three clicks down. The visitor doesn't dig. There's another tab open.
What this demands from your site: the answers above the fold on the pages that matter, which are almost never the homepage. Course pages are where enrollment decisions get made.
Step 4: The inquiry is a test, and most schools fail it
Then someone fills in your form. What happens next is measured, and the numbers are not flattering to the sector: a 2025 mystery shopper study reported by ICEF Monitor found only one in four inquiries received follow-up communications.
Students notice. They contacted more than one school. The first useful reply shapes the whole comparison, and silence is an answer too.
What this demands from your site: a plan for the minutes after submission. Who owns the reply, how fast it goes out, and what it says.
Step 5: The decision takes months, and your reporting probably doesn't
The last step is the one that quietly ruins marketing decisions. Families don't decide on Tuesday because they clicked on Monday. In LanguageBird's account, we measured the real cycle: an average of 4.7 months from first contact to enrollment.
Any campaign judged on a 30 day window in that business is being judged before the story ends, and the campaigns that get paused for underperformance are frequently the ones that were about to pay.
What this demands from your site: measurement that follows the student, not the platform. If your reporting can't connect an enrollment back to the click that started it four months earlier, you're managing on guesswork.
Ask a school director how students find them and you'll usually hear a version of the 2015 answer: Google, agents, word of mouth, maybe Instagram. None of that is wrong. It's just no longer the whole path, and the parts that changed are the parts that decide whether you make the shortlist. Here's the journey as it runs now, step by step.
Step 1: The question gets asked, and often answered, before your site is in play
The first move used to be a search that produced ten links. Now it frequently produces an answer.
Google AI Overviews appeared on roughly 48% of queries as of March 2026, up from about 31% a year earlier, according to Advanced Web Ranking data compiled by Digital Applied. When one appears, organic click through rates drop somewhere between 34% and 61% depending on the study. The link is still there. Fewer people need it.
Meanwhile the students themselves moved. ICEF Monitor reports that use of ChatGPT for college search doubled between 2024 and 2025, that education is among the most prominent categories for website traffic referred by ChatGPT, and that prospective students increasingly use AI tools to compare and evaluate study options before they contact anyone.
What this demands from your site: content a machine can quote. Direct questions with direct answers, specific facts instead of adjectives, pages that resolve a doubt in a paragraph rather than burying it in a brochure. If your site can't be summarized accurately, it won't be summarized at all.
Step 2: The shortlist gets built by something that read your website without visiting it
This is the step that breaks the old mental model. A student asking an AI assistant to compare three language schools in Dublin is running a comparison based on whatever those assistants can extract from each school's site and from third party sources talking about them.
You're being evaluated by a reader who never fills in your form, never triggers your analytics, and never sees your homepage hero video. That reader is deciding whether you belong on the list. ICEF's own advice to schools is blunt about the first move: find out whether AI tools mention your school at all, and what they say when they do. Most directors have never checked. It takes five minutes and the answer is usually uncomfortable.
What this demands from your site: clarity over cleverness. Course names that describe the course. Levels, durations, prices and start dates written as facts, not as "contact us for details". Everything an assistant would need to recommend you accurately.
Step 3: The site visit is a verification, not a discovery
By the time someone reaches your website, they usually arrive with a hypothesis: this school might be right for me. The visit exists to confirm or kill it.
That flips what the page is for. It isn't there to impress, it's there to answer, fast, on a phone, in the language the visitor actually reads. Most school sites still open with the institution talking about itself and hide the answers three clicks down. The visitor doesn't dig. There's another tab open.
What this demands from your site: the answers above the fold on the pages that matter, which are almost never the homepage. Course pages are where enrollment decisions get made.
Step 4: The inquiry is a test, and most schools fail it
Then someone fills in your form. What happens next is measured, and the numbers are not flattering to the sector: a 2025 mystery shopper study reported by ICEF Monitor found only one in four inquiries received follow-up communications.
Students notice. They contacted more than one school. The first useful reply shapes the whole comparison, and silence is an answer too.
What this demands from your site: a plan for the minutes after submission. Who owns the reply, how fast it goes out, and what it says.
Step 5: The decision takes months, and your reporting probably doesn't
The last step is the one that quietly ruins marketing decisions. Families don't decide on Tuesday because they clicked on Monday. In LanguageBird's account, we measured the real cycle: an average of 4.7 months from first contact to enrollment.
Any campaign judged on a 30 day window in that business is being judged before the story ends, and the campaigns that get paused for underperformance are frequently the ones that were about to pay.
What this demands from your site: measurement that follows the student, not the platform. If your reporting can't connect an enrollment back to the click that started it four months earlier, you're managing on guesswork.
The uncomfortable
summary
The uncomfortable summary
The uncomfortable
summary
Your website used to be a brochure that people found.
Now it's the source material for machines that recommend you, the verification step for people who already half decided, and the front door of a process that takes months and gets abandoned in the first hour of silence.
Most school websites are still built for the 2015 journey. That's the gap, and it's why the sites that get this right are pulling ahead of schools with better classrooms, better teachers and worse pages.
On August 26 the Platty team is running a free live session on exactly this: how to build a website that informs, converts, and gets cited by AI. Real numbers from real school rebuilds, the checklist you can run on your own site the same week, and 5 complete website audits raffled among everyone in the room.
Keep Reading
Keep Reading
Explore more insights, ideas, and perspectives from our latest articles.
