Method note: this brief proposes a source-led review approach. It does not claim that private client or firm data has been collected.
Question and scope
This protocol asks how often a quarterly accrual estimate differs from the invoice that follows it. The unit is a matter-level accrual line, and the measure is the absolute difference between the estimate and the invoiced amount. The study covers accruals delivered to corporate clients during the period. It does not evaluate whether the client's accounting treatment was correct.
Scope the study to matters with an accrual estimate and an invoice issued within the following two quarters. Include matters that were invoiced in full, in part, and not at all so the denominator reflects the full portfolio. Exclude matters with no accrual line and record the number excluded. State the accrual definition, the billing system, and the owner of the analysis.
Defining the event
Define a material variance as a difference above a threshold agreed with the client, such as a percentage of the estimate or a fixed amount. A variance below the threshold is still recorded but grouped separately. The estimate date is the date the packet was delivered, not the date the work was prepared. A later revision creates a new estimate and a separate comparison.
Distinguish a timing variance from a value variance. A timing variance occurs when the work is invoiced in a later quarter than estimated, while a value variance occurs when the invoiced amount itself differs. Keep the two categories separate because the client response differs. A timing variance may need only an explanation, while a value variance may need a process change.
Population and denominator
The population is every matter-level accrual line in a delivered packet that meets the inclusion rule. Report the number of packets, the number of lines, the number excluded, and the number in the final sample. Count lines rather than packets, because one packet can contain many matters. Keep the unit stable across quarters so the variance rate is comparable.
Report the share of lines with any variance and the share with a material variance. Also report the median absolute difference and the total signed difference so offsetting errors are visible. A high variance rate with a small total difference may be less concerning than the reverse, so report both. The signed difference shows whether the firm tends to estimate high or low across the portfolio.
Evidence sources
Use the delivered accrual packet, the invoice, the time and cost detail, and the client's confirmation as the primary evidence. Compare the estimate line with the invoice line and record the date each was observed. If the invoice was revised, use the final version and note the revision date. The evidence log should let a second reviewer repeat the comparison.
Rank the sources so reviewers know what to trust first. A delivered packet is stronger than a draft, and a final invoice is stronger than a pro forma. A client confirmation is stronger than an internal assumption about what was accepted. Record the evidence type for each line so the analysis can separate a communication gap from a calculation error.
Sampling
Sample across clients, matter sizes, and quarters so one large matter does not define the result. Include lines that were estimated accurately, because the clean cases are part of the distribution. Draw the sample before reading the invoice so the selection is not biased toward the variances. If the population is small, review every line and extend the period.
Stratify the sample by whether the matter had unbilled time held for review at the cutoff. A matter with held time may show a larger timing variance than one with a clean cutoff. Report the variance rate for each stratum rather than pooling them. The reader should be able to see whether the estimate or the cutoff explains the difference.
Coding the record
Use a short code list that points to a specific repair. Code each variance as timing, rate change, scope change, held time, write-down, cost estimate, or unknown. Add the amount and the quarter. Keep the code list short enough that two reviewers can apply it consistently.
Have a second reviewer code a sample and compare the results. Resolve disagreements by returning to the packet and the invoice, not by majority vote. Report the agreement rate and revise any code that is applied inconsistently. If the reviewers cannot agree on the primary cause, stop and fix the definitions before continuing.
Measuring the variable
Measure the absolute variance, the signed variance, and the variance as a share of the estimate. Report the median and the range so a few large matters are visible. Measure the time between the estimate date and the invoice date, because a long gap increases the chance of a timing difference. Keep the raw entries with the analysis file.
Do not convert the variance to a revenue or profitability figure without the approved fee and collection data. The study measures estimate accuracy, not financial performance. If the firm wants a revenue view, run it separately with the finance owner. A precise number built on an estimated rate is less useful than a clear variance distribution.
Analysis and reporting
Report the overall variance rate, the material variance rate, and the median absolute difference. Show the results by cause and by client. Compare only groups with similar accrual definitions and cutoff rules. Include the lines that matched so the report shows the full distribution.
Give the reader one table that answers the main question and put the coding detail in an appendix. Write the finding in plain language before presenting the technical definition. Include the cases that contradict the pattern rather than hiding them in an average. A report that cannot be acted on is just another record of the past.
Limits and sources of error
A variance is not proof that the estimate was careless, and a match is not proof that the method is sound. Report the completeness of the packet and invoice evidence before drawing a conclusion. Small samples, changing scopes, and reviewer disagreement can all limit interpretation. A single large matter can move the total even when most lines are accurate.
Avoid comparing clients that use different accrual definitions or billing cycles. A client with a fiscal quarter that differs from the calendar quarter will produce a different timing pattern. Describe the context beside the rate and avoid ranking individual preparers. The purpose is to improve the estimate, not to judge professional worth.
Privacy and retention
Store the accrual packets, invoices, and client confirmations in a controlled folder with access limited to the study team. Remove client names, amounts, and matter details from the analysis file when they are not needed. If an example is shared, use a role label and redact the client and the amount. Keep the unredacted evidence separate from the working analysis.
Do not circulate a client-level or preparer-level list about accrual variance outside the management context. Aggregate the results before sharing them more broadly, and check that a small group cannot be re-identified. The same rule applies to screenshots of the packet and to quoted messages. Privacy is part of the method, not an afterthought added at the end.
Applying the result
Choose one change that matches the most common cause. If held time dominates, separate held work from the estimate and explain it in the packet. If rate changes dominate, confirm the approved rate before the cutoff. If costs dominate, add a vendor estimate check. Test one change on the next quarter before applying it broadly.
A lower variance rate is not the only useful outcome. A clearer packet and fewer client questions can improve the relationship even when the rate stays similar. Report the direct result and the service result separately. Stop the change if it makes the packet longer without improving the estimate.
Review cadence
Repeat the study each quarter and after any change to the accrual definition, the billing system, or the client mix. Keep the definitions and the unit stable so periods can be compared. Report the lines that were excluded or still open beside the main rate. Review the coding guide after the first repetition and keep the changes visible.
Review the result with the billing manager and the client relationship owner, not only with the preparer. Ask which fix is easiest to sustain and which cause creates the most client questions. Set one change and a follow-up date, then check whether the change held. The next study should test the repair, not just repeat the same count.
Put the guidance to work
Choose one recurring billing workflow, identify its authoritative source records, and name the owner who reviews exceptions. We can help turn that boundary into a focused support role.
Discuss your workflow


