Why we removed the worker marketplace, and what Koora is becoming

We have removed the part of Koora where a provider could browse care workers, and we are not building the job posting that was meant to follow it. This is a note about why, because the reasoning changed what we think the company is.

What we built, and why it looked right

The original shape was reasonable. Care has a workforce shortage, providers struggle to fill shifts, and Koora was already assembling verified, current credentials for workers. If you hold a set of people whose checks are current, letting a provider search that set looks like the obvious next move. Post a job, find someone already cleared, onboard them in a day instead of three weeks. We built the browse surface and had job posting queued behind it.

The logic was that compliance gets you in the door and hiring is where the value compounds. Compliance is a burden, and burdens are a hard thing to build a company on. Nobody wakes up excited about credential expiry. So we wanted to be more than that, and hiring was the answer we reached for first.

What providers actually said

We launched in July. Since then we have spent most of our time in front of providers, and one thing came up repeatedly. It was not about features, and it came up at the same point every time: when a provider was working out whether to put their workforce in.

Providers were uneasy about putting their workforce into a system where other providers could see it. Not their compliance data, which nobody could see anyway without the worker granting access. The plain fact of the roster. In a sector where everyone is short-staffed, a searchable list of care workers with current credentials, sorted by location, is a recruitment list. We had built one and then asked providers to fill it with their own staff.

Job posting would have made that sharper, not softer. The clearest signal that a service is worth approaching is a service advertising that it is short.

The part we got wrong

The mistake was not the marketplace on its own. It was not noticing what the marketplace did to everything underneath it.

The value of a shared credential record comes from it being complete. The entire proposition is that the workforce is in one place with current checks, so nobody has to assemble it again from scratch. Anything that gives an organisation a reason to hold part of its workforce back does not reduce that value at the margin. It removes the reason the thing exists at all.

So a feature we added to make the product worth more was quietly making the core of it harder to adopt. We were asking for the whole workforce while running the one feature that made handing over the whole workforce feel risky.

What it makes possible

Two things open up once nobody is shopping for staff.

A provider can put their entire workforce in without it being a recruitment risk. That is the difference between a compliance tool used for the awkward cases and a record that actually covers everyone, which is the only version that does the job.

And staffing agencies can use it properly. An agency's book is its business, and a platform where client organisations could browse workers is close to the last place an agency would ever upload it. With browsing gone, an agency can put its whole book in and use Koora for the thing it is genuinely good at, which is keeping credentials current across people who work across many sites for many clients. Agencies do not pay for passport access here. Placements work the way they always have, with an agency sharing a specific worker with a specific client, which is a deliberate act rather than a shop window.

Being listed does not depend on being a customer

The other direction of this deserves saying plainly, because providers will find it out anyway and it is better heard from us. A workplace can appear in Koora whether or not that provider has an account, and a worker can write about somewhere they have worked whether or not the employer is a customer of ours. The directory covers the sector as it exists, including workplaces that have never heard of us.

That will sit uncomfortably with some operators, so here is the reasoning rather than just the policy. Plenty is already published about a care service. Childcare services carry an NQS rating from their last assessment, residential aged care homes carry Star Ratings on My Aged Care, and anyone can leave a Google review. What almost none of it covers is what the place is like to work at. Those ratings measure the care delivered to a child or a resident, and the reviews are written by families. So a worker deciding whether to pick up a shift somewhere they have never set foot still has very little to go on, while the same worker has been screened and rated from every direction. If that picture existed only for providers who had signed up, the places a worker would most want to know about would be exactly the ones missing from it.

Workers reviewing workplaces already exists, too. Glassdoor and Indeed both carry employer reviews and both are open to anyone who has worked somewhere, so this is not a capability Koora is introducing to the sector. What changes is where it sits. On a general review site a worker has to know the site exists, go to it separately, and search for a specific service before they take the job. Here it sits beside the credentials they already maintain and the workplaces they are already looking at, which is one fewer place to have to know about.

The limits on it matter as much as the openness. Feedback comes from workers describing somewhere they actually worked, and reading the detail requires having your own Career Passport submitted, so it is a closed room rather than an open comments section. Serious concerns are held for moderation before they appear anywhere. A provider can dispute what has been written about them. And any workplace can claim its profile, see what is being said, and take part in the conversation.

For most providers this is an opportunity rather than a risk. A service that treats people well has had no way to prove that to a worker who has not worked there yet, and word of mouth has never been evenly distributed. The information already moves. It moves through group chats and Facebook groups that no provider can see into, respond to, or correct. Bringing it somewhere with verification, moderation and a right of reply attached is better for a good employer than leaving it where it is.

What we think Koora is now

We have started describing Koora as a workforce bureau for care, and the word needs defining because it can be read two unhelpful ways.

It does not mean a government body. And it does not mean a credit bureau, which is worth being precise about, because a credit bureau holds a file you do not own and cannot see without asking. This is the other way around. The worker holds their Career Passport and decides who sees it. The sources are government registers rather than employers reporting on each other.

What we take from the bureau idea is narrower. One maintained record that many organisations can rely on, instead of the same person being checked from scratch by every service they work for. Checked once, kept current, and readable by whoever the worker allows.

That reframes the burden problem too. Compliance is genuinely a burden, and we were right that a compliance tool alone is thin. We were wrong about where the extra value comes from. It comes from the same work stopping being repeated across the sector: one worker, checked once, current everywhere they work, rather than every service they pick up shifts with running the same checks over again in the same year.

What we are not claiming

Much of what "bureau" implies is not built yet, and the distance is worth naming rather than glossing. We are early and we have no track record. This is a decision and the reasoning behind it, not evidence that the decision was right.

We also gave something up, though it was never a customer request. Nobody asked us for job posting. It was our own bet on where the value would come from, and we had built the browse surface and queued the posting behind it before we tested that bet. We removed the code rather than hiding it behind a setting, because a feature kept alive on the off chance is a feature that keeps shaping the product. If we are wrong about this we will rebuild it deliberately, and we would rather make that decision again from scratch than drift back into it.

The limit on the rest is the same as it has always been. Where an authoritative register exists we verify against it: identity through the Document Verification Service, AHPRA, state working with children check portals, ban registers, and the police checks we will run ourselves soon. Everything else we review, which is a different and weaker claim, and we will keep saying which is which. Koora pre-clears. The provider keeps the legal responsibility for who they engage.

What this commits us to

• Providers do not browse workers. Not as a setting, not as an enterprise feature, not later.
• The worker holds the Passport and controls who sees it. The bureau holds the verification history.
• Agencies do not pay for passport access.
• We keep "verified at source" and "reviewed" distinct, including when the looser word would read better.

None of this makes compliance exciting. It does mean the work only has to happen once, which is the closest thing to a real answer we have found.