Every EBP audit reaches a point where the audit plan moves from paper to practice. You’ve rated contributions as a significant risk. You’ve noted the plan has two payroll providers and a tricky definition of compensation. Then someone opens last year’s program, sees “select 25,” and the engagement rolls forward.
That gap between risk analysis and testing is where a lot of audits quietly lose their footing. It’s also where peer and regulatory reviewers tend to push back. Not “did you test it?” but “why did you test it this way?”
Let’s walk through what happens in between, why it matters, and how to make those decisions easy to defend.
Risk analysis should drive the audit plan. In practice, it often trails behind it.
We see risk assessments filled out after testing decisions are already made. We see control walkthroughs completed as the last step before the opinion goes out. And we see risk assessments that look eerily similar from one EBP client to the next.
But audit plans should look different from client to client. A plan with a complex definition of compensation, multiple payroll providers, or several employer match calculations needs testing designed differently than a “vanilla” 401(k). If the risk assessment changes, but the audit program rolls forward untouched, that’s a red flag.
There's a simple habit that can change everything - for every step in your audit program, ask:
What risk is this procedure responding to?
Start asking that for every audit program step, and you’ll start performing truly risk-based audits.
The hardest part of moving from risk analysis to testing is a change in altitude. Risk analysis happens at a high level. Testing happens at the transaction level.
So, we say, “contributions are a significant risk,” then pull the same number of selections we pulled last year - usually, 25.
But why 25? Is 25 truly representative of this population?
In a financial statement audit, sample sizes are driven by population value and count. An EBP audit is compliance focused, so dollar value isn’t the driver anymore. The trouble is we haven’t made attribute risk the driver either. Too often, we pick a sample size based on firm tradition and go.
There’s a second challenge hiding underneath. Risk assessment usually happens before anyone looks at this year’s data. It’s built on the prior year, high-level financial information, and client interviews. Not on what this year’s population actually shows.
That creates a ripple effect for your team. Staff get handed a sample size without knowing why. And if they don’t know why, they can’t recognize when something they find should change the approach.
Full-population analytics turn risk assessment from an informed guess into an informed judgment. When you can see every participant across every payroll, you can see where the risk actually lives before you decide what to test.
To be clear, this doesn’t replace sampling or auditor judgment - it actually makes both better! With full population analysis you can:
That last point matters because we all know clients see minimal value in their EBP audit. When you can point to something specific in their plan, the audit stops feeling like a compliance exercise and starts feeling like insight!
When testing decisions get questioned in review, it’s usually not because the decision was wrong; it’s because the reviewer can’t see the reasoning behind it.
A significant number of peer review comments come from documentation problems, just as much as testing problems. The person who made the call understood exactly why, but six months later during peer review, nobody can reconstruct it.
Some common trends we’ve heard include:
Some of these are testing gaps, and some are documentation gaps. Either way, the reviewer is left asking why.
Reviewers want to follow a clear line: risk → procedure → result → conclusion.
To keep that line intact, document:
That last one gets missed the most. It’s also what makes many auditors uneasy about changing the approach they’ve used for years. If you looked at something and concluded it wasn’t a risk, write down your thought process.
Because in audit, if you didn’t document it, it didn’t happen.
Earlier, we talked about risk assessments being built on last year’s numbers and client interviews instead of this year’s data. That’s usually blamed on a lack of data. The data exists. What’s historically been difficult and time-consuming is analyzing it in a way that truly impacts the risk assessment.
That’s the problem we built AuditMiner to solve. Auditors can see transactional risk across the full population at the time of risk assessment, not after the testing plan is already set. So the risk assessment reflects what actually happened in the plan during the year, not just assumptions about what could have happened.
That doesn’t mean testing every transaction that flags a risk. An indication of risk in the data isn’t automatically an error. There may be a perfectly valid reason behind it, like a rehire, a plan amendment, or a mid-year election change. The auditor still does what auditors do best: looks at what the data shows, assesses the risk, and sets sampling and testing based on that risk.
Because the risk assessment is tied to the data, there’s a direct line from what you found in the population, to the risk you identified, to the procedures you performed.
So when a reviewer asks, “Why did you test this and not that?” the answer lives in the audit documentation, backed by the data. Not in the head of someone who left the firm six months ago.
Risk analysis sets the direction. The decisions in between, what to test, how much, and why, are what make an audit defensible.
Start small. Pick one audit program this season and ask “what risk is this responding to?” for every procedure. Write down the calls you make, including the ones where you decide not to test something. Then look for ways to bring this year’s data into the conversation before the plan is locked.
Your reviewers will thank you. And so will the version of you sitting in peer review a year from now.
The hardest part of all this is usually the step from ‘I know this year’s data matters’ to actually having it in front of you when the plan is being set. That’s exactly what we cover in The Next Evolution of AuditMiner: A Live Platform Walkthrough. You’ll see how Audit Data Analytics surfaces risk across the full population before you pick a single sample, and how Data Traceability lets a reviewer see exactly where each result came from, all while complementing your firm’s existing workflow.