Suppose 44 out of 100 tax returns went back to preparers after first review last period. This period, only 26 out of 100 did. Before you move managers off review or expand an AI check, compare the same types of returns across both periods. Then apply this period's rates to last period's mix. If the two types are coming back just as often as before, a lighter mix can explain the entire drop.
In one tax practice we worked with, leadership confirmed a $1.5M increase in recurring profit at the follow-up eight weeks after the engagement ended. We traced why returns came back, moved missing-document and recurring-error checks earlier, and gave the firm an effort record by return type. The team also agreed that more returns had to clear first review without lowering quality. The technical case explains how the firm measured the work and tested the changes. You can use the same kind of review records to check whether an improvement in your own report survives a change in the work coming in.
Count each return at its first review
Start with two closed reporting windows. Use the same definition of first review in both. Have the quality lead pull one row per eligible return from the workflow log, using the date its first review began to assign the period. Record the first disposition, including whether the reviewer sent the work back for another pass. Keep pending dispositions visible and wait for those reviews to resolve before you compare the full windows. That way, a return that goes through three rounds of corrections still counts as one send-back out of one first review.
Ask your tax lead to define the return groups using facts available before review, like whether the work covers one state or several. If you label returns complex after you see the review notes, the outcome starts deciding which group it belongs in. Apply the same definitions in both periods, and record which offices and reviewer roles handled each group. That gives you a way to spot a change in who did the work alongside a change in the work itself.
Agree on what counts as a material finding and keep it separate from minor notes. A return can clear first review with a minor note, while a material issue can still require attention even if the workflow log says the return cleared. Record whether each return had a material finding and how many minor notes it had. Have the quality lead arrange a second review of a comparable sample from both periods using the same material-finding standard. That extra check matters because a lower count of recorded findings can also mean reviewers caught fewer problems.
Log active minutes by role for the first and repeat passes through the same endpoint, such as reviewer sign-off. If later returns are still open, wait for their effort totals before comparing staffing needs. Recording time by role helps you see whether managers gained hours or spent the time on a longer first pass or passed the work to another reviewer.
If you want to finish more client work without asking managers to pick up another round of corrections, start by finding out which work keeps coming back. That gives you something concrete to bring to a broader growth discussion.
If you’re a tax practice COO and want to grow your firm without another overloaded busy season, we’ll identify the biggest bottleneck standing in your way, build the plan to remove it, and implement it in eight weeks* or less - guaranteed or we continue to work for free and you don’t pay.
Find my bottleneck* Eight-week delivery, or twelve weeks when substantial data foundations are required, as agreed before work begins. If a covered stage fails its agreed benchmarks, we provide up to four additional weeks of remedy work at no charge, then waive or refund that stage’s fee if it still fails. Successfully delivered stages remain payable. Agreed client-caused delays pause the clock. Delivery guarantee terms.
Give both periods the same mix
Now take the return counts from the opening and put both periods on the first period's mix. In our example, both windows had 100 returns. The first window was mostly complex work. The second had more straightforward returns, which were less likely to come back in either period.
| Return group | First period: sent back / reviewed | Later period: sent back / reviewed | Later rate applied to first-period volume |
|---|---|---|---|
| Straightforward | 4 / 20 (20%) | 16 / 80 (20%) | 20% × 20 = 4 |
| Complex | 40 / 80 (50%) | 10 / 20 (50%) | 50% × 80 = 40 |
| All returns | 44 / 100 (44%) | 26 / 100 (26%) | 44 / 100 (44%) |
Divide later send-backs by later first reviews for each type. Multiply each rate by the number of returns of that type in the first period. Add those expected send-backs and divide by the first period's total first reviews. This gives you a later rate at the first period's mix, which you can compare directly with the first period's observed rate.
Here the adjusted rate stays at 44%, because the send-back rate stayed the same within each type. Keep the observed 26% in your report. It describes what the team handled that period. For the next schedule, check how much complex work is due. If the heavier mix comes back, the two group rates give you a better basis for planning review coverage than the 26% headline alone.
Now suppose 8 of the later complex returns had come back instead of 10. The complex rate would be 40%, and applying it to the first period's 80 complex returns would give you 32 expected send-backs. Add the 4 straightforward returns and the adjusted rate falls to 36%. That's when you'd look at whether quality and review effort held up alongside the drop.
Decide what the evidence supports
Copy this worksheet for one comparison. Have the quality lead enter counts from the review log and ask the tax lead to confirm the groups, coverage, and quality rules before taking the result to the staffing meeting.
| Field | Fill in for your practice | Example |
|---|---|---|
| Scope | First and later date windows, offices, reviewer roles, coverage changes, unresolved first reviews, and shared endpoint for effort | Two closed windows, same office and reviewer roles, all first reviews resolved and time captured through sign-off |
| Repeat for each return type | Period, type, eligible first reviews, sent back, returns with material findings, minor-note count, active minutes by role on first and repeat passes | Straightforward, later period: 80 reviews, 16 sent back, 2 returns with material findings, 9 minor notes, 1,600 first-review minutes plus 320 repeat-review minutes, all manager time |
| Compare the periods | Actual rate in each period, later rate using the first-period mix, counts behind each rate | 44% first, 26% later as observed, 44% later at the first-period mix |
| Check quality and effort | Divide returns with material findings by reviewed returns, and recheck a comparable sample under the same standard. Compare first plus repeat minutes per return by type and reviewer role. | Only the later straightforward row is shown. Pull the other rows and repeat-review time before changing coverage. |
| Record the decision | Comparability issues, supported conclusion, coverage decision, owner, and review date | Hold the improvement claim. Keep coverage for complex work pending the quality and effort comparison. |
If a type is missing from either window, you have no pair to compare. If a few reviews drive a type's rate, show the counts and collect more comparable work before drawing a conclusion. A change in which offices or reviewers handled a group can also move the result, so compare within like-for-like groups where you have enough records. Where you can't, report the gap and hold the claim.
If the adjusted rate falls, identify which types improved. When the comparable quality sample holds up and the type takes fewer minutes per return for the role you're moving, test a small coverage change. Keep enough review time for the next batch's mix, and check whether the review queue or quality measures worsen before making the change permanent. If the result will also be used to credit an AI check, compare similar work with and without the check while keeping review rules and coverage stable. The mix calculation controls for the groups you counted. A comparison designed around the intervention helps you find out whether the check changed the work. If the comparable returns are still coming back too often, use the tax review queue guide to decide whether to investigate readiness or reviewer capacity first.
Before the next staffing meeting, have your quality lead fill in the worksheet and bring both rates with the underlying counts. In the example, there's enough evidence to keep the 26% on the report and enough reason to hold the improvement claim. That gives the COO time to check what the next batch will need before moving review hours elsewhere.
In the tax practice we worked with, repeated manager review was taking time the team needed to finish client work.
If you’re trying to grow your own firm, start with your goal, look at what it'll take to achieve it, and then identify what's keeping you from taking action. Just remember that every firm is different and that repetitive manager review was just their bottleneck and it might not be yours.
If you’re a tax practice COO and want to grow your firm without another overloaded busy season, we’ll identify the biggest bottleneck standing in your way, build the plan to remove it, and implement it in eight weeks* or less - guaranteed or we continue to work for free and you don’t pay.
Find my bottleneck* Eight-week delivery, or twelve weeks when substantial data foundations are required, as agreed before work begins. If a covered stage fails its agreed benchmarks, we provide up to four additional weeks of remedy work at no charge, then waive or refund that stage’s fee if it still fails. Successfully delivered stages remain payable. Agreed client-caused delays pause the clock. Delivery guarantee terms.