Why Background Check Results Vary So Much by Provider

Run the same candidate through two screening providers and you can get two different reports, two different turnaround times, and two different levels of confidence about whether the result is complete. Both vendors described themselves the same way in the sales cycle.

The divergence is not marketing spin. It comes from architecture — where the data is sourced, how the workflow is built, and who answers the phone. Here is where the differences actually live.

Speed: the number nobody defines the same way

Almost every provider claims to be fast. Very few say what they are measuring. A median across all searches looks very different from a median across county criminal searches only, and an average that excludes anything requiring a court clerk looks better still.

Ask for the definition, not the adjective:

• Is the figure a median or an average? Averages hide the long tail that actually frustrates recruiters.
• Does the clock start at order submission or at candidate consent completion?
• Are delayed and manual-research files included in the number, or excluded as exceptions?
• What percentage of orders complete same day, and over what search mix?

Real speed comes from direct court integrations and automated workflows that clear straightforward candidates immediately, so human researchers spend their hours on the files that genuinely need them. For reference, we report a trailing 12-month median of 11 hours and 37 minutes and publish how it is calculated.

Compliance: built in, or bolted on afterward

Screening law varies by state, county, and city, and the differences are not cosmetic. Whether a record is reportable, how far back it can be reported, what a disclosure form must say, and how long a candidate has to respond before final adverse action can all change when a candidate lives one county over.

Some providers encode those rules in the platform, so the wrong disclosure or the wrong timeline cannot easily be produced. Others handle compliance as guidance — a PDF, a webinar, a support ticket — and leave interpretation to the employer. Both approaches are described as compliant in a proposal. Only one of them prevents the mistake at the moment it would happen.

Data coverage: where 'national' can mean less than it sounds

There are more than 3,000 counties in the United States, many operating multiple court systems with their own records, formats, and access rules. Reaching them directly requires integrations, court runners, and continuous maintenance as clerks change systems.

A cheaper alternative is to lean on aggregated multi-jurisdictional databases. Those databases are useful as a pointer — they surface leads worth confirming — but they are compiled from contributed data, they lag, and they can miss or misreport dispositions. A program that treats a database hit as a finished answer produces both false positives that harm candidates and false negatives that harm employers.

The questions that separate providers here: how many courts do you access directly versus through a database, how quickly is a database hit confirmed at the source, and what is your quality-audit process on returned records? We maintain direct digital access to 3,000+ courts, which is why same-day clears are possible on a large share of orders.

Integration: every re-keyed field is a delay and an error

When screening lives outside the applicant tracking system, recruiters re-enter candidate data, chase status in a second portal, and relay updates by email. That friction is invisible on a rate sheet and expensive in practice — it slows time-to-fill and introduces typos into the identifiers a search depends on.

A real integration passes the order automatically, writes status back into the ATS, and keeps the candidate's experience inside one flow. When you evaluate, ask whether the integration is certified and maintained by the provider or a one-off build the last customer paid for.

Support: the variable that surfaces after implementation

Support quality is almost impossible to evaluate from a demo and immediately obvious in week three. An ambiguous record, a candidate dispute, or a jurisdiction question either gets resolved by someone who understands screening, or it sits in a queue while the offer cools.

Chatbots and tiered ticketing improve provider economics; they do not always improve resolution. The practical questions are who answers, where they sit, what their screening background is, whether candidates get support too, and what the escalation path looks like when a hire is on the line.

How to compare providers on the things that matter

1. Ask for turnaround as a median, with the measurement definition and the search mix behind it.
2. Ask how many courts are accessed directly and how database hits are confirmed at the source.
3. Ask to see the disclosure and adverse action workflow inside the platform, not in a slide.
4. Confirm the ATS integration is certified, maintained, and writes status back.
5. Ask who supports your team and your candidates, during what hours, with what screening expertise.
6. Run the same three real candidate profiles through each finalist and compare the reports side by side.

Frequently asked questions

Why did two providers return different records for the same person?

Usually source depth. One searched the court of record directly while the other relied on an aggregated database, or the two used different name and identifier matching logic and different jurisdictional scopes.

Is a national criminal database search enough on its own?

No. It is a lead-generation tool. Responsible programs confirm every database hit at the court of record before it appears on a report used for an employment decision.

Does a lower per-check price mean a worse report?

Not automatically, but price pressure has to come from somewhere. If it comes from automation and direct court access, quality holds. If it comes from substituting databases for court research or from thinning support, it does not.

Key takeaways

• Turnaround claims are meaningless without the measurement definition behind them
• Compliance encoded in the platform prevents mistakes that guidance documents only describe
• Direct court access is the difference between a lead and a confirmed record
• Integration and human support decide how the program feels once it is live