But what about demand? Where visa refusals do (and don't) reshape postgradate search interest

One of the (presumably) unintended consequences of AI is that it's now much easier to visualise UK student visa application and outcome data in new, exciting and occasionally terrifying ways. We might call it the UKV(A)I apocalypse. And it's happening, now, in a LinkedIn post near you.

This means that the story of the UK's relative fall in international student visa applications (down around 20% in 2026 to date) and the spike in refusals (up more than 50% in the first half of the year) is becoming very familiar. But I need to retell part of it in order to get where I'm going in this post.

Fewer applications

The story has also moved on slightly since we last looked at it. We now have application data up until September and outcome data up until June.

Here's what the former looks like:

The lines here give the cumulative total of student visa applications (main applicants, not dependents) in each of the past three years. Remember, this is prospective international students applying to UKVI with CAS from their university. We'll get to the outcome of those applications in a moment.

Because this is cumulative (the overall gap grows further each month 2026 is behind 2025 and 2024) the distance widens. But it hides a slight silver-lining. Whereas monthly applications were down around 30% compared to 2025 for most months between January and June, the gap narrowed to around 20% in the all-important June-September period (when most applications actually happen).

It's important to remember that this is probably a supply and demand story. Students may be becoming less interested in UK study. But, given the change to compliance metrics we're going to talk about, universities may also be less inclined to progress some applications to CAS. Which is to say that the rate at which international applications are being rejected by universities themselves isn't in this data. Even though it probably impacts interest in turn.

The tl;dr is that applications aren't looking as bad as we thought, but they're still down around as much as they were when the 2024 dependent ban came in (a point made with typically astute candor over on Wonkhe).

And fewer of those fewer applications are getting through.

More refusals and withdrawals

This time we're looking at the breakdown of visa outcomes in each quarter from Q1 24 to Q2 26 (the most recent data as of writing).

• The FAU-lilac section of each column is the percentage of applications accepted by UKVI, standing at 88% in Q2 26, down from 97% in Q2 25.

• The definitely-not-FAU-red section is the percentage of applications refused by UKVI, standing at 7.2% in Q2 26, up from 2.8% in Q2 25. 

• And the also-definitely-not-FAU-yellow section is the percentage that have been withdrawn, standing at 5% in Q2 26, up from a tiny 0.6% in Q2 25.
  
  

The refusals are where most of the attention has been paid this year as, sector-wide, they're now above the 5% threshold for Basic Compliance Assessments. The key term there is 'sector-wide'. Any one institution with a refusal rate that high will risk having its CAS allowance (the ability to sponsor visas and therefore accept most international applicants) cut; in reality many will be safely below and some could be well above. We'll see why when we look at (very different) refusal rates for specific student nationalities.

The withdrawals are receiving more attention now, with the assumption that they're potentially a product of the refusals. Withdrawals happen when a student or a university pulls an application in progress before a decision has been made by UKVI. All data for this is anecdotal, but it's possible (and rumoured) that universities and/or agents might be encouraging students to withdraw applications they expect to be refused. Universities may also cancel CAS themselves. In both cases the withdrawal 'protects' the university from a refusal in its statistics and, theoretically, protects the student from having a visa refusal on their own record (though this may not always work perfectly and, in any case, having an application-in-progress cancelled isn't typically a good thing for students).

The other thing to remark in this visualisation is the pattern. Refusals (and withdrawals) are higher in Q1 and lower in Q2. This is true of every year; it's just that the proportion of refusals+withdrawals is that much higher this year. We would now expect Q3 to have the fewest refusals+withdrawals of any point in the cycle, before rising again, relatively speaking, in Q4. This is especially important because a high refusal rate in Q3 impacts significantly more applications.

We won't know for certain until November, but anecdotal stories and proxy data so far suggests Q3 26 might see an unprecedented spike.

So what about demand?

This is the part of the story you won't have seen on LinkedIn. Unless you arrived here after future-me shared it there.

Because (one) obvious question we can ask about rising UK student visa refusals is "what is this doing to demand?". And, unlike the broader question of "why is this happening?" there's a relatively clear answer. We can measure it with our Keystone Share of Search data, like this:

We're now plotting the change in search interest in the UK against the proportion of refused and withdrawn student visa applications over the same period of time. Search data is via FindAMasters, so it's Masters-level interest (that's how that works) whereas visa applications are across all levels of study, but will be substantially Masters-level in practice.

To make comparison easier, the search trend is indexed to Q1 24. So we're not looking at the absolute 'volume' of search interest in the UK: we're looking at how it's grown over time since the start of 2024, measured at half-yearly intervals. The search data in each case is for the half-year period that encompasses the two quarterly visa columns either side of it. It reflects the broader period in which those outcomes are happening and in which they can potentially be influencing searching audiences.   

With that in mind, we see more or less what we'd expect. Search interest starts to fall as refusals and withdrawals start to rise. This makes sense if international audiences are hearing stories of higher UK visa refusals (they are) and shifting their behaviour accordingly.

It also makes sense when we look at some individual audiences. Here's Pakistan as an example:

The country with the highest and most sudden spike in visa refusals also sees the most dramatic fall in UK search interest across the same period. Remember, the fall in search interest that seems to precede the Q4 25 refusal+withdrawal spike reflects search across Q3 and Q4.

It's not always quite as neat. Here's the pattern for Sri Lanka (another audience with relatively high refusals+withdrawals):

The drop in Q4 25 is there again, but it levels off (for now).

Some audiences also 'buck the trend'. Nigerian search interest does fall in Q4 25, but it rises again in 2026:

There are likely to be other factors at play here, including the fact that the USA added Nigeria to a travel ban at the start of 2026 (when its UK interest picks back up).

Most importantly, some audiences trend in the opposite direction. The USA itself is a good example of this:

The refusal+withdrawal rates here are never high enough to worry the BCA threshold and they're actually down vs last year. Increased US interest in the UK (and Europe) is another logical story in the current climate.

So, what does this mean?

I think there are the requisite three takeaways from this attempt to cross-examine search and visa trends.

The first is the most obvious: International search interest and application outcomes are joined by a feedback loop. As audiences increasingly experience rejection, refusal and requests to withdraw, they share those stories with other prospective students and search interest in the UK drops. This matters to universities because it exposes the need to reassure and re-engage the audiences you do care about, particularly in markets you are active in. It probably also exposes the need to communicate clearly with markets you'd like to be active in again in the future. Trust is hard to rebuild.

The second point nuances the first: Of course I like our Share of Search data and of course I think it's valuable. But search is still window shopping. Audiences don't necessarily stop browsing just because the 'product' has become harder to buy. Sometimes that news can stimulate curiosity. This matters because it exposes the opportunity to reassure and re-engage above. Audiences are still searching even when they're less interested. And some aren't less interested.

The third is that international education is complicated: Refusal rates vary by audience and so does search interest. There are some remarkably resilient or compliant markets. There are also some resilient and compliant markets. We are still here to help you find and reach them.

And if you are staying up-to-data on LinkedIn, I massively recommend following folks like Dave Amor, Nicholas Dillon and Spencer Withrington.