top of page

Job Boards Aren't Broken. They're Working as Designed (Just Not for You).

  • Writer: Susan Morrow
    Susan Morrow
  • Dec 23, 2025
  • 7 min read

Updated: Jan 1

Why is job search so hard?

You search for roles that match your background and get flooded with irrelevant results. When you do find something that looks right and apply, you hear nothing back. After enough silence, the conclusion feels obvious: you weren't qualified, someone better came along, the post was flooded by thousands of resumes or you were too late.


Job search is hard because the systems we rely on were never designed to help people find work that fits. They're designed to help platforms generate revenue while employers manage risk. Once we understand what job boards are actually built to optimize, a lot of confusing behavior starts to make sense and stops feeling personal.


How do job boards actually work?

The search systems powering most job boards became mainstream around 2004. The logic is simple: you type in keywords, the system returns documents containing those keywords. It’s about as simple and natural-sounding as ["account manager" OR "customer success") AND SaaS NOT ("call center" OR "commission only")]. Sadly, job search sites forgot to tell us that we needed to be fluent in the use of Boolean operators in order to find a job.


This kind of strict, old-school search requires you to already know exactly what you're looking for, describing it using the exact language the system expects. But when there's a gap between how you think about your work and how employers title their roles, the search breaks down. You miss relevant opportunities because you didn't use the right synonym. You waste time scrolling through irrelevant postings because the keywords matched but nothing else did.


Better technology exists. Research comparing Google to ChatGPT shows that ChatGPT users write queries that are twice as long and phrase them as questions instead of keyword strings. We already have tools that can understand "I'm looking for project management roles in healthcare where I can use data analysis skills" and return coherent results. GenAI proves this is solvable.


So why don't job boards fully leverage the power of generative AI? The constraint isn't technical. It's contractual and economic.


What do job listings actually optimize for?

Job listings are authored and owned by employers. Platforms typically only license the right to display them. Substantively rewriting or restructuring listings to improve search risks misrepresenting employer intent and violating contract terms. Platforms index what's written rather than reinterpret it.


But the bigger barrier is marketplace incentives. Job boards originated as two-sided platforms that monetize employer activity: job postings, impressions and application volume. It’s a very big business: LinkedIn alone made $7 billion from its hiring software in fiscal 2023.

Were there to be a natural language search interface, it would sharply narrow results and reduce low-fit applications, benefiting job seekers. It’s also true that using AI like this would lower engagement and shrink the perceived reach employers get for their money. And that weakens platform revenue.

That’s why friction in job search is structurally necessary. Employers favor volume and optionality. Platforms optimize for the paying side of the market. Candidates (that means you) absorb the inefficiency.


Modestly estimated, about 25-35 million Americans are actively searching for jobs at a given time, which makes this paradox even more painful: job boards have the technology to make search dramatically more efficient. Using it would expose just how much the current system depends on it being broken.


What you get instead are band-aids: filters for location and salary, "AI recommendations" that are really just pattern matching based on click history, surfacing jobs "similar" to ones you've viewed or labelling positions "easy apply." None of these fix the core search problem.


The biggest improvement these platforms could make is to basic search. They won't.


What does this broken system produce?

Start with the listings themselves. A job posting isn't a neutral description of work. It's a recruiting document designed to attract a large pool of applicants while protecting the employer from risk.


Requirements get inflated because employers would rather screen out candidates than miss the perfect hire. Research on hiring practices shows that "required" qualifications are often negotiable and "preferred" skills may actually be deal-breakers. Candidates don't know which is which, so they self-select out when they don't match perfectly.


Titles stay vague and inconsistent because they serve multiple purposes. A "Marketing Manager" at a 50-person startup might own positioning, demand generation and customer research. A "Marketing Manager" at a Fortune 500 company might execute one campaign type under three layers of approval. Same title. Completely different work. Employers use titles for internal hierarchy and search optimization, not to match the titles used by other businesses. Plus, ambiguity can benefit the employer: generic job descriptions let them attract a broad pool, then filter based on criteria that might not even appear in the posting: culture fit, internal referrals, timing or budget shifts.


Beyond confusing listings, the system produces something worse: ghost jobs—postings for roles that aren't actually open for hiring. Companies post these to build talent pipelines, signal growth to investors or make current employees feel replaceable. You can't tell the difference from the outside. (We'll examine the full scope of ghost postings, their prevalence data and their market impact in Part 3 of this series.)


Even when a posting is real, other signals you may get are unreliable, such as the apparent “competitiveness” for a position. You see a job with 493 applicants and wonder if it’s even worth it to apply. But that number might include people who clicked "Easy Apply" without reading the description, fake applications generated by AI, duplicate submissions or applicants from the last time this posting was live. It tells you nothing about how many candidates are actually competitive or how many the employer is seriously considering.


Similarly, posted duration is equally meaningless. Roles stay up because there's no penalty for leaving them. A posting that's 60 days old might mean the employer is still searching. It might mean they forgot to close it. It might mean they're keeping options open with no active hiring plan.


What are job boards actually good for?

None of this means job boards are useless. It means you need to know what question they can answer.


  • Job boards are good for scanning employer language. What terms do companies in your target industry use to describe similar work? What skills show up repeatedly across postings? What's the range of titles for roles that overlap with your background?

  • They're useful for directional market signals. Are there more postings for data analysts in healthcare or finance right now? Which roles have the most remote opportunities? What salary ranges appear for your experience level?

  • They're helpful for researching specific employers. What roles is a company hiring for? What do their descriptions emphasize? How do they frame priorities or culture?


What job boards can't tell you: whether you're actually competitive for a role, how many real candidates the employer is considering, why you didn't hear back or what the job actually entails day to day.


What's the real cost of treating job boards as neutral marketplaces?

But all their other imperfections aside, the real damage of using job boards is the assumption that they are neutral marketplaces that deliver meaningful feedback.


When you apply to 50 jobs and hear nothing, that's not evidence you're unqualified. Candidate experience surveys from CareerBuilder and similar sources consistently find that 50% to 75% of job seekers report receiving no response after applying. The system isn't built to give you feedback: silence is the default.


Job boards can be useful tools when you understand what they're measuring and what they're hiding. They can just as quickly become toxic if you let them become the basis for conclusions about your value.


But there's a problem that hits even before you start clicking "apply": what exactly are you searching for? "Marketing Manager" returns 47,000 results. Which ones actually match what you do?




The Job Search Breakdown Series

The system isn't broken. It's working exactly as designed—just not for you. This series explains why job boards optimize for volume over matching, why searching by title returns useless results, where 38% of American workers find income that never appears on job boards, how your brain turns structural dysfunction into self-blame and how to extract signal from noise without burning time or confidence.




Sources Cited

Application Volume:

AI Usage in Job Search:

  • Indeed Global AI Survey (2023): 70% of job seekers reported using generative AI for job search activities (drafting cover letters, preparing applications)

Ghost Jobs:

Platform Revenue:

Job Seeker Response Rates:

Active Job Seekers:

  • "About 25-30 million Americans actively searching for jobs at a given time" (who are either unemployed and actively job hunting or employed and expect to change jobs within the next three months)

Search Behavior (ChatGPT vs Google):

  • "Research comparing Google to ChatGPT shows ChatGPT users write queries that are twice as long"

    • Source: Multiple independent analyses of usage patterns indicate that ChatGPT prompts are substantially longer than Google search queries—often by 50–60% in word count, and in some proprietary datasets by several‑fold for informational queries—reflecting more conversational, paragraph‑style input.

    • Good summary here: “New Data Study: What Queries is ChatGPT Using Behind the Scenes”

 
 
bottom of page