Review pay exceptions with an evidence plan

Inconsistent starting pay exceptions deserve careful review, especially when approvals live in email. A talent solutions consultant helps a compensation leader organize evidence without reaching legal conclusions. The work does not establish compliance or diagnose a problem from partial records. A sound review begins with a smaller question: which records exist, which fields have source support, and which owner must resolve every gap?

Harper, a fictional compensation director, asks Luis, a fictional talent solutions consultant, “Do these records prove we have a pay equity problem?” Luis replies, “Available records support a fact review. Compensation and counsel decide which conclusions belong in a formal assessment. Let us identify stored decisions, approval owners, and role or location fields available for verification.” Harper says, “Our band refresh is six months away, and exceptions sit across many email threads.” Luis answers, “We will inventory source records, mark gaps, and give review owners focused evidence questions.”

Luis starts with scope. He asks Harper which hiring period and employee population deserve attention, then records her answer as a working boundary rather than a final assessment design. A review of new offer exceptions from January through June differs from a review of all compensation decisions. Scope affects source systems, required fields, and responsible teams. Luis also asks who holds access to offer records, compensation ranges, approvals, and archived email. Every source needs an owner who understands completeness limits.

Harper chooses a fictional January through June review of new hires whose final approved pay differed from a range reference in an offer record. Luis writes this as a working question, not a conclusion. He does not expand scope during early collection because a manager remembers another issue from years earlier. He opens a separate parking note for questions outside current review. This protects time and gives Harper a record of subjects needing later direction.

Next, Luis proposes a simple data dictionary. Every record includes offer date, role, level, location, range in effect, proposed pay, approved pay, exception reason, approver, source location, and record status. Range in effect matters because a later range should not replace a range used during an earlier offer. Source location might read applicant system, offer file, approval email, or compensation worksheet. A field without a source stays unsupported even when someone recalls an answer.

Luis adds two control fields: source date and reviewer note. Source date identifies when a document or message was created. Reviewer note records a short explanation of an unresolved question. For example, a note might state, “Approval email lists amount but no role level.” The note does not say why a leader chose an amount. Luis keeps interpretation separate from evidence so another reviewer understands both known information and open work.

The status column keeps early review honest. Luis uses verified when a source record supports a value. He uses missing when no record has appeared. He uses owner needed when a source likely exists but no person has confirmed responsibility. He avoids correct, fair, improper, or compliant. Those words move from evidence collection into conclusions. The worksheet describes available records rather than deciding what facts mean.

Consider five fictional records. Asha’s offer file lists a March 4 date, level three, Boston location, and approved pay figure, yet no exception reason appears. Ben’s approval email names a hiring leader and an amount but lacks role level and effective range. Carla’s applicant record has job title and location, while an approval thread has a different title and no final offer document. Devon’s file contains a complete approval form. Elena’s record refers to a spreadsheet folder, but nobody knows who owns access. Luis marks Asha’s reason missing, Ben’s level and range missing, Carla’s title for reconciliation, Devon’s fields verified, and Elena’s source owner needed.

Luis does not solve mismatches through guessing. For Carla, he opens a question for compensation: “Which approved role and level match the final offer, and which record supports your answer?” For Ben, he asks a compensation operations owner, “Where are ranges effective on his offer date stored, and who verifies level?” Each question names a record and decision point. “Please clean up this file” gives a recipient no clear next move.

Email approvals need careful handling. A chain might show discussion, a proposed amount, a leader approval, or a later change. Luis preserves message dates and identifies which message supports every worksheet value. He does not move every personal detail into the sheet. A narrow evidence plan asks for fields needed for the stated review question. Private context, manager opinions, or unrelated messages do not belong in an inventory merely because they appeared in one thread.

For example, Ben’s chain starts on April 8 with a recruiter request. A hiring leader replies April 9 with a proposed amount. A compensation contact replies April 10 with approval. On April 12, a later message changes start date but says nothing about pay. Luis links proposed pay to April 9, approved pay to April 10, and offer date to final offer documentation once found. He does not assume April 12 changed an earlier approval. He asks an owner for a final offer record before marking offer date verified.

Harper asks whether the upcoming band refresh should change current work. Luis replies, “The refresh is a future compensation event. Keep source ranges tied to each offer date, then ask compensation how a separate future comparison should work.” This prevents two time frames from mixing in one review. The eventual review owner decides whether a refresh creates a new question, a new data set, or no change to present scope.

Counsel and compensation hold different responsibilities in this workflow. Counsel sets review boundaries and any protected handling approach. Compensation defines pay records, business context, and decisions for review. Luis organizes material, documents gaps, and keeps questions traceable. He does not state legal requirements, decide whether a record proves unequal treatment, or promise an outcome. These boundaries keep an evidence worksheet from becoming an unsupported verdict.

The review needs a cadence. Luis schedules a thirty minute working session with Harper and source owners after first inventory. The agenda covers fields with an identified source and fields without an owner. Before meeting, every participant receives relevant record identifiers and unanswered questions. During meeting, Luis records an owner, requested record, and due date beside every open item. A blank cell does not become a task until someone owns a next move.

Suppose an owner says, “I think Ben’s exception was approved because we needed speed.” Luis does not enter this as exception reason. He asks, “Which document records an approved reason, or should this field stay missing?” If no document exists, the worksheet says missing. A later reviewer sees a known gap instead of an explanation built from memory. A helpful lead still has value, so Luis records a follow up question for Harper rather than treating memory as evidence.

At Monday’s fictional review meeting, Luis opens rows for Asha, Ben, Carla, Devon, and Elena. He gives Carla’s title mismatch to a compensation operations owner with a Wednesday due date. He gives Elena’s folder access question to a records administrator with a Tuesday due date. After each owner names a source record or confirms a gap, Luis enters response date and status beside the relevant row, then sends Harper the updated unresolved question list for her next review.