Signed-in loads are around 30% faster, with roles shown instantly.
Repeat visits to a board now show roles instantly from a locally cached snapshot, while the latest data loads quietly behind it. Signed-in page loads are around 30% faster, and skeleton placeholders replaced the old loading text. Members on slower connections feel this most.
What makes it faster
The change is not a single optimisation, it is a change in what happens first. Instead of waiting for fresh data before showing anything, the board shows what it already has and updates in place.
- Repeat visits render roles immediately from a locally cached snapshot.
- The latest data loads in the background and replaces the snapshot without a jump.
- Signed-in page loads are around 30% faster.
- Polished skeleton placeholders replaced the old loading text, so the shape of the page is visible while it fills.
That is the whole of it. No settings, no migration, and it applied to boards that were already running.
What it means for your community
Members stop deciding the board is slow, which is a judgement people make in the first second or two and rarely revise. A careers hub competes with every other tab someone has open, and the ones that hesitate on load are the ones quietly dropped from the routine.
An alumni network sees this in return visits. Alumni check in occasionally rather than daily, and an occasional visitor has no patience banked. If the board is showing roles before they have finished deciding whether to wait, they browse. If it is showing a spinner, a good proportion close the tab and the institution never learns why engagement is flat.
A skills bootcamp has a version of this shaped by devices. Learners are often on phones and on connections that are not a fast office network, and they are checking between other commitments. Showing cached roles instantly is worth more to them than to anyone reviewing the board from a desk, which is exactly the group whose engagement the placement rate depends on.
What it means for your team
Nothing to do, which is the point. There is no cache to configure, no setting to weigh up and no risk of members seeing stale roles, because the fresh data arrives moments later and replaces what was shown.
It also removes a confound from your reporting. When a board is slow, low engagement has two possible explanations and you cannot separate them. Once loading is not the bottleneck, the numbers in member analytics are telling you something about relevance rather than about patience, which makes them worth acting on.
Skeleton placeholders matter more than they sound too. A page that shows its own shape while filling reads as working, where a blank area with a line of text reads as broken, and members who think something is broken tell you so.
It changes what you can reasonably do in a campaign as well. Sending a cohort a filtered link only works if the page arrives quickly enough that the click survives, and that is doubly true when the link goes out by email and gets opened on a phone somewhere with poor signal. Speed is what decides whether the filtered view you carefully built ever gets seen.
Do members need to clear anything?
No. The cache is handled for them and refreshes on its own, so there is nothing to clear and no way for a member to end up looking at yesterday's roles. What they see on arrival is the last known state, replaced within moments by the current one. If a role closed since their last visit, the update lands before they could realistically have acted on it.
Getting started
There is nothing to enable. The improvement applies to boards already running and to every member signed in. The marketplace admin guide covers the board as members see it, and the knowledge base handles questions about member experience. The work is recorded in the page load changelog entry.
If member engagement is what you are trying to move, the job board for alumni networks and job board for skills bootcamps pages go into it further. Spacewalk runs a 14-day free trial with no credit card required, and a board goes live in about an hour.