A jobs page showed 30 remote openings from the previous 30 days, even though its operator says the site had collected thousands more listings that were not reaching the page. The problem was not a sudden change in the job market: the page was reading from only one of three job sources. In a DEV Community account, Nokku.ai operator Payanai describes how they merged the feeds, corrected misleading skill tags, rendered listings into the page HTML, and added a quality threshold for generated pages. The reported result was about 200 live jobs from 150 companies—not a market-wide count or an independently verified measurement.
Why the page showed only 30 jobs
Payanai says the remote-jobs page displayed 30 jobs from the last 30 days, making the market look quiet. The collection pipeline drew listings from RemoteOK, JSearch, and Arbeitnow, but the page read from only one source. Records from the other two were stored but never shown.
As an Amazon Associate I earn from qualifying purchases.
That distinction matters: data being collected does not mean it is available to the visitor. After changing the pipeline to combine the sources and clean the records, Payanai reported approximately 200 live remote jobs from 150 companies. Those are project figures from the operator’s account, not an independently audited inventory or a measure of remote-job availability overall.
What changed in the listings pipeline
Combine the feeds, then clean the records
The revised process queried all three sources, deduplicated listings using normalized company and position information, and kept the version with pay details when available. It also rejected titles that indicated hybrid, on-site, or in-office work, and excluded links that were not HTTP(S).
These are practical filters, not guarantees of perfect data. Similar titles can belong to different roles, and a job’s title alone may not accurately describe its location policy. The account describes the operator’s heuristics but does not establish how often they misclassified jobs or left duplicates behind.
Recalculate skill tags instead of trusting stale ones
A previous substring-matching approach produced plausible-looking but false matches: “Go” could match “good,” “AI” could match “email,” and “Java” could match “JavaScript.” In Payanai’s example, 102 of 201 listings were tagged Go before the correction; afterward, 11 were. These are counts from that site’s dataset, not statistics about demand for Go developers.
For older listings, Payanai says the site recalculated skills when reading the records, using job titles and job-board tags rather than relying on stale stored tags. The broader lesson is to inspect the records behind a surprising aggregate. A large count can be a data-quality defect rather than a meaningful pattern.
Free tools Windows power users keep installed
One-click scans. No signup required.
Why the page’s HTML and URLs mattered
Before the change, the page rendered listings on the client, so the initial HTML showed a spinner while jobs loaded. Payanai says the revised /jobs page was server-rendered, refreshed every 30 minutes, and included its first 25 listings in the HTML. Filters appeared in the URL, while the canonical URL continued to point to /jobs.
This describes the implementation, not proof of a ranking improvement. The account provides no independent crawl test or search-performance results. Server-rendering listings makes them part of the initial response, but it does not by itself establish how a search engine crawls, indexes, or ranks a page.
How the operator set a floor for generated pages
The site created pages for specific roles and skills, including /jobs/remote-python-jobs, /jobs/remote-software-engineer-jobs, and /jobs/remote-jobs-with-salary. These pages address more specific browsing intents than a general remote-jobs listing; the examples are design choices, not evidence of measured keyword demand.
Rank #4
Payanai says generated pages with fewer than five listings were marked noindex and left out of the sitemap, but remained accessible to visitors. The operator also describes writing page introductions from the available data instead of swapping keywords into a generic template. The five-listing cutoff is one site’s chosen threshold, not a universal rule, and the account reports no independent search-performance outcome.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
How small samples can make a page sound more certain than it is
An early headline reported median listed pay from only three listings. Payanai says it was replaced with the more limited label “With pay listed: 3,” which tells readers how many records support the pay information without implying a robust market-wide median.
The operator also narrowed the “New” label to listings from the newest day. A two-day window had made 13 of the first rows appear new. Both examples show why a page should make the sample and time window clear wherever it presents a statistic or status label.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.A practical way to audit a job-data page
- Trace records from source to screen. Check which boards feed the collection, which records are stored, and which sources the page actually reads.
- Review cleanup rules. Look for duplicate handling, evidence that a job is genuinely remote, and checks that links are valid.
- Test tags against actual listings. Inspect the rows behind surprising skill totals, especially when matching depends on short words or substrings.
- Check what arrives in the HTML. Compare the initial page response with the fully loaded view, and verify how filters and canonical URLs are represented.
- Evaluate generated pages individually. Thin pages and generic introductions may not help visitors; a page should have enough relevant listings and information to justify its existence.
- Put sample sizes next to claims. A count of listings with pay information may be more honest than a median based on a handful of records.
Payanai’s concise advice is: “Treat surprising numbers as bugs until proven otherwise.” In the account, an implausibly high Go count led back to faulty matching rather than a market trend.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →
