Manual review puts a real person on the orders that fall in between auto-accept and auto-decline, to make the call automation couldn't make on its own.
Why it's not actually a good primary strategy
Manual review demands real time and headcount, doesn't scale, and slows down the customer's experience while an order sits in a queue. Quality depends entirely on the reviewer's expertise and tools — and high staff turnover makes consistent performance genuinely hard to maintain.
The scaling problem, in numbers
A 2013 CyberSource report found merchants under $5M in annual revenue manually reviewed 42% of orders, while merchants over $100M reviewed just 11%. The gap isn't because bigger merchants have less fraud — it's that manual review simply can't scale with volume the way automated screening can. The real question is what happens when a growing business hits its own peak season with a review process built for its smaller self.
Main drawbacks
- No real scaling path beyond hiring more people
- Inconsistent results across reviewers with different experience levels
- Rarely backed by a formal training program
Where it's actually worth keeping
Reserve manual review specifically for orders that would otherwise be auto-declined — using it to convert borderline sales rather than as a first-line fraud filter. That's where it earns its cost, both in fraud caught and revenue saved.