# What a job index holds, and what it cannot tell you

> A job index is a stored copy of postings gathered from many sources. On 24 October 2026 about one in five of the live jobs in Zorlya's index had no country recorded, and most known companies had no live job. These figures change daily, and none of them tells you how many openings are real.

Source: https://zorlya.com/blog/what-a-job-index-holds-and-what-it-cannot-tell-you

By Zorlya. Published 2026-10-26. Research and data. 8 min read.

What a job index holds decides what any job search can show you. Every job site that lets you search is searching an index: a stored copy of postings, collected from other places and kept in one database. The index is not the job market. Knowing what it holds, and what it leaves out, helps you read any search result, ours included. This post uses Zorlya's own index as the example.

## What was read, and when

These were read on 24 October 2026. They change every day, as postings arrive and leave, so treat them as a snapshot.

| Measure | On 24 October 2026 |
|---|---|
| Known companies with a live job | A small share of the companies known |
| Live jobs with no country recorded | About one in five |

"Live" means the posting is in the index and has not been closed. It does not mean the employer is still hiring for it. More on that below.

## Most known companies have no job right now

The first row shows a gap. Zorlya knows of many more companies than have a live job. The rest are known employers with no opening that we can see. The employer directory on the countries page states its own count, taken when the site was built, so it is a different measure from the snapshot above. A company with no live job has no open role that we found at the last read. It does not mean the employer has stopped hiring.

## A job is not an opening

Counting jobs counts postings. A posting is not always one position, and it is not always open.

One posting can cover several positions, and several postings can describe one position, for example the same role listed for two cities. Zorlya removes a duplicate only when the link matches one already seen, or when title, company, location and description are identical. Anything that is merely similar stays as a separate job. The post on [how Zorlya removes duplicate postings](/blog/how-zorlya-removes-duplicate-postings) has the detail.

A posting that disappears from its source is not automatically closed in the index. The platform's own notes record that postings which vanish from a board stay open, and that closing them is open work. So some live jobs have already been filled or withdrawn.

For these reasons, the total in any index is a size of the stored collection. It is not a count of people being hired.

## What does an index store for each job?

Less than the posting, often. Zorlya keeps the title, the employer, the date and the link for every job, and the place when the source gives one. The full description, the pay and the contract type are kept only where the source provides them and its terms allow it. Where a site's terms do not license republishing, the index holds only those few fields and you read the ad on the site itself. So two jobs in the same index can carry very different amounts of detail, and a job with little detail has less for a score or a filter to work with.

## How does an index learn that a job has closed?

Only in two ways: the source tells it, or it reads the source again and notices. Search engines ask employers to do the first. Google's guidance for job pages says [expired postings should be removed](https://developers.google.com/search/docs/appearance/structured-data/job-posting), by marking the end date as past, by taking the page down, or by removing the job data from it, and then asks the site to notify Google. A board that copies postings has no such notice from most employers. It finds out at its next read, if it compares what it holds with what the source still lists.

That second step is the hard part. If a read fails half way, a posting missing from the partial list may still be open, and closing it would be wrong. That is why Zorlya does not yet close vanished postings. Whatever index you use, the employer's own page is the one that knows.

## A fifth of jobs have no country

About one in five live jobs had no country recorded on 24 October 2026. The platform's notes name two cases: a remote job with no place, and a place with no country code. A search by country cannot match either, because there is nothing to match.

So when you filter by country and see a small number, the true number is somewhat larger. If you are open to remote work, search without a country filter too, and read the location line on each posting yourself.

## Every filter drops jobs that lack the field

The country filter is one case of a general rule. A filter can only keep jobs that carry the field it reads, and in Zorlya's feed each filter drops the jobs that do not.

| Filter | Reads | A job is dropped when |
|---|---|---|
| Country | The country of a parsed place | It has no place with a country |
| Remote | Whether the source said remote | The source did not say, whichever way you filter |
| Posted within | The posted date | It has no posted date |
| Minimum salary | The top of the posted pay, in US dollars | It has no pay figure |

So each filter you add narrows the list twice: once by your choice, and once by the jobs whose source left that field blank. Illustrative: the figures are invented. A search returns 400 jobs. Filter to one country and 120 remain. Add a minimum salary and 30 remain, because most of the 120 never stated pay. The 90 you lost are not below your figure. Their pay is unknown.

A good habit: apply one filter at a time and watch how far the count falls. Write down the count before and after, so you know which filter cost you the most jobs. A big fall on a filter like salary or posted date says more about the sources than about the jobs.

## What the index leans toward

An index reflects the sources it can read, and the sources are uneven. On the same date, the countries with the most live jobs were Sweden, Canada, the United States and Portugal. Countries in other regions had far fewer. That does not show there are fewer jobs elsewhere. It shows that the index holds what its sources publish, and some countries have fewer sources that allow automated reading.

That gap is why some employers are missing from every job board, and [where the jobs come from](/blog/where-the-jobs-come-from) explains how sources are chosen.

## How to use this when you search

Treat the total as a size, not a promise. A big number says nothing about the jobs that suit you. Open the posting at its source before you spend time on an application, and check that it is still listed and that the date makes sense (see [why a posted date is unreliable](/blog/why-a-jobs-posted-date-is-unreliable)). Search a second way: if you filter by country, also try without the filter, and if you search by title, also try the kind of work in plain words. For a company you want, the careers page is the one place that is current.

## Questions people ask

### Is a bigger index better?

Only if it holds the jobs you want. A large index with few postings in your field and country helps you less than a small one that covers them. Search for your own kind of work before you judge any tool by its total.

### Why does the same job appear on several sites?

Employers post on their own page, and boards copy or receive the same posting. Each site keeps its own copy, and each copy can be out of date in its own way.

### Why does a filtered search show so few jobs?

Often because the jobs that do not state the filtered field are dropped. Remove the filter and read the postings yourself to see what you lost.

### Does a live job mean the employer is still hiring?

No. It means the posting is in the index and has not been closed there. Check the [posting it was read from](/blog/three-dates-on-a-job-posted-deadline-and-closed) before you apply.

## What we know and what we do not

We know what the index stored on 24 October 2026, and how its filters treat missing fields. We have not measured what share of real openings in any country the index holds. We have not measured how many live jobs are already filled. The numbers above describe what is stored, on one day. Do not read them as coverage.