Most hiring tools pick a side. That's the problem.

Every hiring tool on the market has quietly made a choice: whose side are you on?

Applicant tracking systems work for the company. They filter, rank, and reject at scale so a recruiter's inbox stays survivable. Resume tools work for the candidate. They optimize, keyword-match, and auto-apply so one person can reach a hundred jobs before lunch.

Both are winning their half of an arms race. The arms race is the problem.

Two broken experiences, one broken loop

Talk to a hiring manager at a small company and you'll hear the same thing: too many applications, no time to read them, no confidence that the person they hired was actually the best one in the pile. So they skim the first twenty, hire someone decent, and move on. The other 180 applicants never hear back.

Talk to a job seeker and you'll hear the mirror image: I applied to sixty roles and got four replies. No feedback, no idea if I was even close. So they apply to more roles, less carefully, casting wider because the whole funnel feels like a coin flip.

Those look like two separate complaints. They're the same feedback loop, running in opposite directions:

• Employers can't read deeply, so they filter aggressively.
• Candidates know they're being filtered, so they apply more broadly.
• More applications make it even harder to read deeply.
• So the filter gets more aggressive — and around it goes.

Every tool that "wins" for one side pours fuel on this loop. A smarter auto-apply bot means more noise in the employer's pile. A more aggressive screening filter means more good candidates rejected for reasons no human ever looked at. You can sell a genuinely great product to one side and make the whole system worse.

Why picking a side fails

Choosing a side is seductive because it's simpler to build and simpler to sell. You know exactly who your user is, and you can promise them a clean win: more qualified candidates, or more interviews.

But hiring isn't a market where one side's gain is the other's loss. It's a matching problem. The employer and the candidate want the same outcome — the right person in the right role — and they're both being failed by the same broken middle. When a tool optimizes for one side's metric (applications filtered, or applications sent), it optimizes against the match itself.

The tell is what happens to information. In a healthy hiring process, signal flows both ways: the employer learns why a candidate does or doesn't fit; the candidate learns where they stand and what to fix. A side-picking tool deletes that signal. The ATS auto-rejects with a form email. The auto-apply bot fires a generic résumé into a void. Neither side learns anything — so the loop stays broken, because nobody can see it clearly enough to fix it.

What building for both sides actually means

It isn't about being neutral, or splitting the difference. It's about refusing to treat the two sides as opponents. A few things we hold to:

Make the same information legible to both sides. If a system scores a candidate against a role, the candidate should be able to see their own fit and the specific gaps — before they apply, not after they're rejected. The employer gets a ranked, explained pile; the candidate gets an honest read on where they stand. Same data, shown to both.

Optimize for the match, not the volume. The goal isn't to help employers reject faster or help candidates apply faster. It's to help the right applications find the right roles — so there's less to reject and less to send in the first place.

Keep a human in the loop. Software should surface, summarize, and explain: the strong candidate sitting at position #94, the one skill that's missing, the reason a match is a match. It shouldn't quietly make the final call with a filter nobody reads.

Refuse the arms race. We won't ship a feature that helps one side win by making the other side's experience worse. That rules out a lot of "growth" features. That's the point.

The unglamorous version

None of this is a moral crusade. It's just what you conclude when you look at hiring as one system instead of two customer segments. The employer drowning in CVs and the job seeker applying into silence aren't two markets to sell to. They're two ends of the same rope — and pulling harder on either end doesn't help anyone hold on.

That's why we build both: Spark Enterprise for the teams doing the hiring, and SparkCareers for the people going through it. Same infrastructure, same read on the data, same belief that the fix has to work for both sides or it doesn't work at all.

Most hiring tools pick a side. We think that's exactly why hiring still feels broken to everyone in it.