When you ask why the report shows 90 inquiries, someone should be able to show you the records behind that number. Which contacts were counted? Which were excluded? Are any still missing?
Trustworthy marketing data is data whose meaning, origin, and limitations the firm can explain. Marketing operations builds the process for answering those questions; a polished dashboard cannot answer them on its own.
The work starts with definitions, continues through collection and transfer, and ends with reconciliation before the figures are used to make decisions.
Are you counting calls, inquiries, clients, or matters?
Phone calls, form submissions, people, inquiries, consultations, and engagements are different units. A person can contact the firm several times about one matter. One client can engage the firm for several matters.
Decide which unit answers the business question. If the firm wants to know how many new requests arrived, count distinct inquiries under a documented rule. If it wants to know how much contact activity staff handled, the individual events remain useful.
Keep both where appropriate. The problem is not that the totals differ; it is that a report presents them as interchangeable. Our law firm marketing analytics guide explains why connecting activity with outcomes requires a view of the whole journey.
Follow one number from the phone and form records to the report
Consider a hypothetical review of one period. The source systems contain 140 contact events. After review, the team identifies 20 additional contacts on already-counted inquiries, 10 existing-client service contacts, five spam contacts, and five controlled tests. These categories are mutually exclusive in this example.
That leaves 100 distinct eligible new inquiries: 140 − 20 − 10 − 5 − 5 = 100.
The intake system contains only 96 of them. Investigation finds four failed transfers. The dashboard shows 90 because six valid intake records use a status label omitted from its filter.
Scroll sideways to review every column.Each row is shown as a labeled card.
| Checkpoint | Observed count | Explanation |
|---|---|---|
| Source contact events | 140 | Includes repeat contacts and excluded activity |
| Reconciled eligible inquiries | 100 | Defined unit after reviewed exclusions |
| Initially received by intake | 96 | Four transfers require recovery |
| Initially shown in dashboard | 90 | Six valid intake records excluded by a filter |
| Corrected, reconciled inquiry count | 100 | Transfers recovered and reporting rule corrected |
The firm has not generated ten additional inquiries by repairing the report. It has recovered an operational gap and corrected its account of existing activity. That distinction matters when reporting the value of the work.
The example is deliberately simple. Real investigations need identifiers, timestamps, source records, and evidence for each exclusion. If a discrepancy remains unresolved, show it rather than inventing a balancing adjustment.
Preserve the path from source to report
For each important field, record where it originated, how it moved, and how the report uses it. This need not become a technical manual for every reader. A concise map can show the source, authoritative system, transformation, and reporting definition.
For example, the original campaign value may be preserved in the inquiry record while a reporting mapping groups spelling variants. Keep the raw value and the mapping version. Otherwise, a later reviewer cannot distinguish a corrected label from overwritten evidence.
The technology-stack guide discusses the wider challenge of disconnected systems. Traceability makes those connections inspectable: someone can identify whether the issue arose during collection, transfer, classification, or calculation.
Treat completeness and accuracy as separate questions
A field can be complete and wrong. Requiring a source value may encourage staff to choose “Google” when the prospect does not know. Conversely, a blank source can accurately represent missing information.
Monitor missing values, but do not reward completion at the expense of truth. Provide an explicit unknown category where appropriate and a process for correction when evidence becomes available.
Accuracy also requires checking definitions with the people doing the work. A record marked “consultation” may represent a booking, a completed meeting, or a canceled appointment depending on the user. Review examples and resolve differences before comparing rates across teams.
Choose quality checks around the decisions the data supports. An incorrect owner can interrupt service immediately. An inconsistent historical campaign label may primarily affect analysis. Both matter, but their urgency and remediation differ.
Give missing and mismatched records someone to resolve them
Failed transfers, unmatched engagements, duplicate inquiries, and invalid statuses should have visible owners and next actions. A connection that silently drops difficult records creates the appearance of completeness while reducing it.
Define how staff discover an exception, who investigates it, and how recovery is verified. A retry should not create a second inquiry or repeat an unintended communication. Preserve the original failure and resolution in the record where practical.
Do not let the reporting person become the permanent repair mechanism for every upstream problem. A documented correction can restore a report, but the source process also needs attention if the failure will recur.
Put the report's limits beside its conclusion
A report should identify the period, cohort where applicable, included channels, outcome definitions, cost boundaries, and unresolved records. Note differences in observation time when recent inquiries have not had enough time to become engagements.
Keep behavioral evidence within its limits. Our published user-journey analysis illustrates how navigation can reveal a path worth examining. A recorded sequence of pages does not establish the visitor's legal qualifications or prove that a particular link caused a retained matter.
Similarly, a tracked source is not the complete history of influence. A prospect may also have heard a referral, read reviews, or returned several times through unobserved paths. The report should support decisions without claiming to observe those missing interactions.
Release changes with a reconciliation check
When a form, status, integration, or filter changes, identify the reports that depend on it. Test representative cases and compare the result with the expected record.
Keep a dated change log. If the definition of an eligible inquiry changes, the trend line may no longer be directly comparable. Explain the break or produce a comparable view where the source evidence supports it.
Before distributing a material report, have another person reproduce a sample calculation. The purpose is not to audit every row manually forever. It is to verify that the documented method can be followed and that discrepancies have a route to resolution.
Make confidence specific
Instead of saying “the data is accurate,” say what has been checked: source events reconciled to inquiry records, duplicate treatment reviewed, engagement links verified, and remaining unknowns listed.
Then identify the decision the evidence supports. The firm may have enough confidence to investigate a campaign's service fit while still lacking enough mature outcomes to increase its budget. Confidence can differ by question.
Apply that discipline to the reconciled 100-inquiry example:
Scroll sideways to review every column.Each row is shown as a labeled card.
| Decision | Evidence after repair | Confidence and action |
|---|---|---|
| How many eligible inquiries entered this defined cohort? | Source events reconcile to 100; transfer and filter gaps are identified and recovered | High for this cohort under the documented definition; publish the restatement |
| Which inquiries came from each campaign? | Original values exist for 92; five rely on consistent self-report; three remain unknown | Moderate and uneven; show 92 observed, five self-reported, three unknown rather than allocating all 100 |
| Did intake receive every eligible inquiry? | All 100 now appear; failure queue and replay were checked | High for the repaired period; monitor the next cycle before declaring the process stable |
| Should the firm increase budget? | Qualification is complete for 72; 18 are unresolved and ten are too recent for a mature engagement decision | Low; hold the increase and set the next review by cohort maturity |
This table prevents one successful reconciliation from turning into unlimited confidence. The firm can correct the inquiry count and repair the transfer while remaining appropriately uncertain about channel influence and budget.
Marketing operations earns trust by making those distinctions visible. JurisOS is the current Juris Digital page associated with this work, but its public description does not establish a standard reconciliation deliverable. Bring the original report, source counts, transformations, unresolved records, and the decision it is supposed to support into a contextual JurisOS conversation. Ask what traceability work can be defined in writing.
Last updated:
