What's the best candidate screening / pre-employment assessment software, especially for small business?

Best depends on which of two jobs you are hiring the software to do. Candidate screening covers skills and psychometric assessment, which measures what a person can do and how they tend to work, and it covers background screening, which verifies criminal, employment, and education history through a consumer reporting agency. Those are separate products with separate pricing and separate legal duties, so a small business shopping for one often lands in a demo for the other.

Which tests actually predict performance

The most significant recent reassessment of selection research placed structured interviews first among common selection procedures at an operational validity of .42, ahead of job knowledge tests at .40, empirically keyed biodata at .38, work sample tests at .33, and general cognitive ability at .31 [11]. Those figures sit roughly .10 to .20 below estimates that circulated for decades, because earlier meta-analyses over-corrected for range restriction [10]. For a small employer that means a disciplined structured interview is the highest-value addition to a hiring process, and assessment software earns its subscription by making a work sample or job knowledge test cheap enough to run before anyone spends interview time.

What constrains the decision at small headcounts

Budget is rarely the binding constraint, since most tools serving companies under a few hundred employees publish a free tier or an entry plan between roughly $80 and $300 per month [1][2][3]. Hiring volume matters more, because paid plans meter tests or credits by the year and a company running four searches will not use what it bought [2][3]. Independent buyer guides sort this category by scenario rather than naming an overall winner, and matching your situation to a vendor's pricing model settles a shortlist faster than a feature grid does [12][13].

TestGorilla is the general skills platform most often shortlisted by companies under a few hundred employees, because it publishes a real free tier and a single entry price rather than routing every buyer to sales. The free plan costs $0 and includes one admin seat, five essential skills tests, and 10 credits per month [1]. Core is $142 per month, billed as $1,704 annually, and opens a library of more than 350 tests spanning cognitive, situational, role-specific, software, programming, language, personality, and culture measures across two admin seats [1]. Plus starts at $400 per month on an annual commitment starting at $4,800, adding unlimited admin seats, custom tests, and API access [1].

Usage is metered in credits rather than seats, and that detail decides whether the price works for you. Skills tests, video questions, and ID verification each cost one credit per candidate, AI interviews cost two, and resume scoring costs nothing [1]. Independent reviewers list it as best for versatile, flexible pre-employment assessments and position it for teams needing fast deployment across diverse role types [12][13].

The honest caveat is the annual commitment, since Core cannot be bought month to month and a company that tries it for one search still pays for the year [1]. Keep that from stinging by picking no more than two tests per role and setting the passing threshold before you see any scores.

CoderPad and CodeSignal are the two technical screening platforms that publish small-team pricing, and they differ mainly in how they meter volume. Independent reviewers name CoderPad for gamified technical evaluations and CodeSignal for ease of use [12].

CoderPad

CoderPad prices by interview volume with unlimited users on every paid plan, which suits teams where several engineers sit in on interviews. Starter is $120 per month billed monthly or $80 per month billed annually at $960 a year, covering 60 tests or interviews per year against 400 ready-to-use questions. Team is $400 per month, $4,800 billed annually, covering 360 per year against 1,400 questions. Overages run $25 each, and webcam proctoring, ATS integrations, and SSO sit in the custom tier [2].

CodeSignal

CodeSignal prices by credits and automates more at the entry tier. Build is $79 per month billed annually or $99 monthly, with 60 annual credits, roughly 2,000 questions, AI interviewers, AI proctoring, and ID verification. Grow is $479 per month billed annually with 420 annual credits and basic ATS integrations. Pro is custom, and overages are $20 per credit [3].

Volume caps bite faster than buyers expect on either platform, since eight engineering searches a year at fifteen candidates each needs 120 tests, double what the entry tiers include [2][3].

Criteria is the test publisher smaller employers most often shortlist when they need psychometric measures with documented validation behind them rather than skills tests written by subject matter experts. Its portfolio covers cognitive aptitude, personality including workplace preferences and role-specific fit for sales and customer service, emotional intelligence, risk measures covering integrity and rule adherence, skills tests, and short game-based assessments, with AI proctoring as an add-on [5].

The scale claims on Criteria's own product page explain why buyers in regulated or safety-sensitive roles look here first: more than 60 million tests administered, more than 15 years of accumulated data, more than 3 billion item responses, and assessments the company states are validated by leading experts to predict actual job performance [5]. Independent reviewers describe it as best for scalable, science-based candidate assessments and as a well-validated, faster-to-deploy alternative to the largest enterprise publishers [12][13].

Pricing is quote-based with nothing published, so budgeting requires a sales conversation, and independent coverage places the sweet spot at mid-market scale rather than at twenty employees [12][13]. A vendor's validation research is also not validation for your job, since the Uniform Guidelines put the obligation on the employer using the procedure [8]. Ask for the technical manual behind any test you use as a cutoff.

Harver is the platform purpose-built for hourly and frontline hiring, and its published claims describe a different workflow from what a general skills platform supports. The company markets candidate evaluation in as little as six minutes using scientifically validated pre-employment assessments, and reports a 52% decrease in time to hire, a 63% decrease in employee turnover, and a 97% candidate satisfaction rate [6]. Its named grocery customer describes hiring timelines moving from around 20 or 30 days to interviews taking place 24 hours after the application [6].

Independent coverage places Harver with enterprise organizations that have high-volume hiring needs in retail, hospitality, logistics, and customer service [13]. That positioning is also the caveat: a restaurant group opening three locations can justify the contract, while a fifteen-person company filling two hourly roles a year cannot, and no published entry price exists to test that against [6][13].

Every figure above comes from the vendor's own marketing pages, so read them as claims rather than measured outcomes for your operation. A small employer with real hourly volume and no enterprise budget gets more from a short structured phone screen with fixed questions and a scoring key [10][11].

Adding assessments to an ATS you already pay for is usually the cheapest path when testing needs are modest, and Workable shows the math clearly. At the one to twenty employee band, Standard is $299 per month or $3,588 per year, Premier is $599 per month or $7,188 per year, and Enterprise is $719 per month or $8,628 per year, with a 20% discount for annual billing [4]. Assessments are a $59 per month add-on on Standard and are bundled into Premier and Enterprise [4].

If you already run Standard, $59 a month adds testing for $708 a year, less than TestGorilla's Core plan at $1,704, and keeps scores attached to the candidate record in one system instead of two [1][4].

The trade-off is depth and pricing shape. A test library inside an ATS is generally narrower than a specialist platform's, so a company hiring across unusual or highly technical roles runs out of relevant tests sooner. Workable's tiers also scale with company headcount rather than hiring volume, which means the subscription grows as you add employees even in a year when you hire almost no one [4].

Federal selection rules apply to any procedure used to make an employment decision, which the EEOC states covers interviews, review of experience or education from application forms, work samples, and physical requirements, not only formal tests [8]. Buying assessment software does not move that obligation to the vendor.

Adverse impact and the four-fifths rule

The Uniform Guidelines require employers to produce evidence of validity only when a selection procedure adversely affects the opportunities of a race, sex, or ethnic group for hire, promotion, retention, or another employment decision [8]. Adverse impact is measured against a rule of thumb: a selection rate for any race, sex, or ethnic group that is less than four-fifths, or eighty percent, of the rate for the group with the highest rate is generally regarded by the federal enforcement agencies as evidence of adverse impact [8]. That threshold draws an initial inference rather than defining unlawful discrimination, and recordkeeping obligations apply to all groups [8].

ADA obligations during testing

The ADA requires employers to give application tests in a format or manner that does not require use of an impaired skill, unless the test is designed to measure that skill, and may require testing materials in alternative formats as an accommodation [9]. It also prohibits questions likely to reveal the existence of a disability before a job offer, covering written questionnaires and interview inquiries as well as medical examinations [9]. That matters for personality inventories, some of which carry items reading closer to clinical screening than to work behavior.

Record per-group selection rates from the day a test goes live, and ask each vendor in writing for the validation report behind any score you use as a cutoff. This describes published federal requirements and is not legal advice; selection testing is worth reviewing with employment counsel, particularly where state or local rules on automated hiring tools apply.

Every assessment you add removes candidates, including some you wanted. Application completion data reported for 2025 shows applications taking under five minutes completing at 12.47% against 3.61% for those running over fifteen minutes [14]. A 40-minute battery at the top of the funnel shrinks your pool measurably, which is a fair price for a role drawing 300 applicants and an expensive one for a role drawing nine.

Volume metering is where budgets break

Published entry tiers all cap usage, and the caps are lower than they look. CoderPad Starter includes 60 tests or interviews a year with $25 overages, and CodeSignal Build includes 60 credits a year with $20 overages [2][3]. Size a plan by candidates tested rather than roles opened, since a single search can consume twenty credits.

Scores are only as good as the cutoff behind them

Assessment platforms produce a ranked list whether or not the ranking means anything for your job. Because the duty to show a procedure is job related sits with the employer when it produces adverse impact [8], a cutoff invented after seeing the first batch of scores is hard to defend and often measures the wrong thing. Define what a passing score means in terms of the work before the first candidate takes the test.

Quote-based pricing lengthens the evaluation

Criteria and Harver both sell by quote with nothing published, so lining them up against a published price means entering a sales cycle first [6][12]. Ask for the annual minimum on the first call.

Four things get bought under this label that assessment software does not actually do, and mistaking any of them for the others is the most common reason a purchase disappoints.

It is not a background check

Background screening is a separate product bought from a consumer reporting agency and priced per report. Checkr publishes Basic at $29.99, Essential at $59.99, and Complete at $94.99 per report, with third-party court and database access fees passed through at cost with no markup, plus a postal adverse action letter add-on at $7 per letter [7]. That adverse action step exists because declining someone based on a consumer report carries notice requirements no skills test triggers.

It is not an applicant tracking system

Assessment platforms score candidates without running requisitions or holding your hiring records. Some ATS products bundle testing, as Workable does at $59 per month on Standard [4], and a standalone tool still needs somewhere for the candidate to live before and after the test.

It is not a replacement for a structured interview

Structured interviews rank first among common selection procedures at an operational validity of .42 [11]. Software makes testing cheap to administer without substituting for asking every candidate the same questions against a scoring key.

It is not a compliance product

No vendor absorbs your obligations under the Uniform Guidelines, which place the duty to produce validity evidence on the employer whose selection procedure produces adverse impact [8].