Demonstrate Experience Without Compromising Confidentiality

12.1 Match Capability and Outcome Claims to Evidence That Can Support Them

Start with the exact sentence the firm proposes to publish. “We advise on procurement,” “we redesigned this client’s procurement process,” and “our work reduced procurement costs by 15%” make different claims. They require different support.

The US Federal Trade Commission’s advertising-substantiation policy calls for a reasonable basis for objective express and implied claims before publication. The required support depends on the claim and the circumstances. Use this as one authoritative reference for a broader editorial discipline: establish the evidence before writing the promise. 

A signed scope supports what the firm was engaged to do. It does not establish that all the work was completed or that the intended benefits occurred. Delivery records, accepted outputs, measurement records, and client confirmation answer different parts of the evidence question.

Use precise status words. Planned work, completed work, an estimated benefit, an observed change, and realized savings should remain distinguishable throughout the draft. A client accepting a recommendation is different from implementing it.

Keep the internal evidence separate from the public explanation

Maintain the original records in an appropriate restricted location. The public page can explain the source, method, period, and limitations without exposing underlying client data. If a figure comes from the client’s management reporting and was not independently verified, describe its source accurately.

Do not label a result independently verified because the engagement team checked its own calculation. Equally, a firm-authored case repeated in a press release and a directory remains the same underlying account; repetition does not provide independent corroboration.

When the available support is narrower than the proposed claim, narrow the wording. A documented process redesign can be useful evidence of how the firm works even when financial outcomes were not measured or cannot be disclosed.

Prioritize the claims that matter most to client selection: experience in the relevant problem, the work actually delivered, the qualifications required, and any results prominently used in marketing. Assign an accountable reviewer to each material claim.

12.2 Develop Case Studies With Context, the Firm’s Role, Methods, Results, and Limitations

A case study should help a reader understand what happened and whether the experience is relevant to their situation. Use a consistent structure, but allow the evidence to determine what the story can say.

• Context and problem. Describe the client’s situation and the decision or difficulty addressed. Include only the characteristics needed to assess relevance.
• Assignment and role. Explain the firm’s agreed work and its actual contribution, including material boundaries and the roles of other participants.
• Methods and work performed. Describe the analysis, professional judgment, consultation, design, testing, or implementation that occurred.
• Outputs and results. Identify what was delivered, what changed, and how any reported outcome was measured.
• Limitations and implications. State what the case cannot establish, relevant dependencies, and circumstances that could affect its application elsewhere.

Avoid forcing every engagement into a financial success story. A completed assessment, an identified control weakness, a documented decision, or an agreed implementation plan can be a substantive output. Explain its purpose and status without inventing a subsequent business benefit.

Attribute the work precisely

Separate the contribution of the firm from the work of the client, other advisers, subcontractors, and technology suppliers. If the firm supplied a specialist to a larger program, describe that assignment. If it introduces an independent consultant, distinguish the introduction from directing or delivering the project.

An individual’s experience at a former employer should be identified as that individual’s earlier work. It should not silently become a project delivered by the current firm. Verify both the role description and the right to discuss the assignment.

Engineering project descriptions need particular care about roles. A firm that assessed selected structural elements should not appear to have designed the entire facility. NSPE’s Code of Ethics expressly addresses exaggerated responsibility for prior assignments, proper credit, and recognition of proprietary interests. Governing licensing rules and contractual obligations also require attention. 

Accounting and legal examples

In an accounting process-redesign example, the firm mapped the monthly close, introduced reconciliation templates, and clarified responsibilities. Client managers implemented the revised workflow. A useful account would identify the delivered materials and any observed changes, while preserving unresolved reconciliation issues. It should describe the advisory assignment accurately and avoid implying that it provided an audit opinion.

In a legal matter example, a firm analyzed contractual notice requirements and supported negotiations leading to a revised performance timetable. The account should distinguish the firm’s work from the client’s commercial decisions and the parties’ agreement. If no court decided the merits, the case should not be presented as a court victory.

Keep context alongside the result in the headline, opening summary, captions, and related service-page excerpts. A careful qualification buried at the end cannot help a reader who encounters only an overstated summary.

12.3 Distinguish Named Cases, Anonymized Accounts, Aggregated Findings, and Labelled Illustrations

Choose the publication format after reviewing the available evidence and restrictions. Each format communicates something different about the work and its support.

For a real case, use a description such as “client identity withheld” when that is the fact you can establish. Treat a stronger claim of anonymity as something requiring assessment. The label does not remove the need to consider whether the information itself may be disclosed.

The UK’s Information Commissioner’s Office explains that removing direct identifiers may leave people identifiable through combinations of details or linkage with other information. Its guidance concerns personal data; confidentiality of a corporate client also requires its own contractual and professional review. 

Make aggregation intelligible

State what is being counted. Ten projects for two clients are ten projects and two clients. A percentage based on respondents differs from one based on all engagements. Identify missing observations and whether the published set includes only cases for which clients permitted disclosure.

Keep success-case selection separate from claims about typical performance. A collection chosen because its outcomes were favorable can illustrate those outcomes, but cannot by itself establish the frequency with which all clients achieve them.

Aggregation also needs a disclosure review. UK government guidance on statistical tables describes risks from small groups, dominant contributors, and combinations of overlapping releases. The implication for a firm’s portfolio statistics is to inspect what could be inferred from the published combinations, rather than assume a total or average is automatically safe. 

Label constructed examples at the point of use

A composite assembled from several engagements should be labelled as a constructed illustration. Do not present its combined circumstances and outcomes as one completed client project. Check that source details remain permitted even after they have been combined.

For fully invented examples, state that the data and scenario are fictional. Preserve that label in the title or opening, captions, downloadable files, and excerpts. A worked example can demonstrate analysis or communication; it should never enter the firm’s count of client engagements or realized results.

12.4 Review Permissions, Confidentiality, Indirect Identification Risks, and Work-Product Rights

Apply the publication policy from Chapter 4 to the exact proposed release. Give reviewers the draft, headline, graphics, quotations, proposed client attribution, and intended channels. Approval of a general idea is insufficient to establish what may appear in a finished case study.

The review should resolve five practical questions before publication:

• Authority. Who can approve this use for the firm, client, and any other relevant rights holder? Confirm authority rather than assuming an enthusiastic project contact can authorize publication.
• Obligations. What do engagement terms, confidentiality commitments, professional rules, and any matter-specific restrictions permit?
• Information. Does the draft reveal protected facts about the client, individuals, counterparties, employees, or other participants?
• Rights and attribution. Who owns or controls the proposed quotations, photographs, drawings, templates, reports, and other work products? What credit and permissions are required?
• Extent of approval. Which version, assets, channels, adaptations, and future uses are authorized? What conditions require another review?

Apply professional obligations to the actual facts

Under the ABA commentary to Model Rule 1.6, confidentiality extends beyond communications made privately by a client. Disclosures that could reasonably lead to protected information also matter. Without a separate basis authorizing disclosure, a hypothetical should not create a reasonable likelihood of identifying the client or situation. These are model provisions; the responsible lawyer must apply the governing rules. 

Public availability is a separate question. ABA Formal Opinion 480 explains that information in a public record is not automatically exempt from the Model Rules’ confidentiality obligations for lawyer commentary. A court filing or published result therefore does not, by itself, settle whether the representing firm may use it in a promotional case. 

For AICPA members in public practice, the Code addresses publication of benchmarking or studies even where clients are not specifically identified. When the information is confidential, the relevant interpretation requires specific consent and prefers it in writing, identifying the information, recipient types, and intended use. Other applicable requirements may be more restrictive. 

Review the whole combination of identifying details

Ask a reviewer familiar with the sector to examine the proposed case alongside the firm’s other public material. An unusual problem, a location, an executive role, a transaction date, and a project photograph may collectively reveal far more than each item alone.

Use that review to remove unnecessary specificity or select a different format. Generalization must remain truthful and preserve what the reader needs to interpret the result. Do not change material facts and continue presenting the account as an actual engagement.

Review the assets as well as the prose. Check filenames, document properties, tracked changes, comments, screenshots, chart labels, and embedded data. An approved paragraph does not clear the unreviewed client information attached to it.

Give outside writers and AI drafting tools only material approved for that tool, recipient, and purpose. Preparing a public version should not expose confidential source material through an unapproved drafting or anonymity-testing process.

Retain the final approved version, supporting permissions, restrictions, approvers, and publication locations. Review substantive adaptations and changes in facts or permissions. If a correction or withdrawal becomes necessary, update the firm’s controlled copies and record where further follow-up is needed.

12.5 Describe Results Without Overstating Causality, Representativeness, or Future Outcomes

Separate the observed change from the explanation for it. HM Treasury’s impact-evaluation guidance treats a simple before-and-after comparison without a credible comparison group as weak evidence for attribution because other influences can explain the change. This is a methodological reference, not a requirement to run an experiment for every case study. 

A consulting improvement example

In a fictional order-approval pilot, the median turnaround fell from 8.0 to 6.0 business days. The observed median was 25% lower: (8.0 − 6.0) ÷ 8.0. The consultancy mapped handoffs and helped pilot revised routing. The client also added a coordinator.

The supporting record should identify the two observation windows, eligible requests, counts, start and end events, treatment of cancellations and unfinished requests, and the business-day calendar. Apply consistent definitions and explain material differences in the work observed.

The figures can support a statement about the observed medians if the records are sound. They do not establish that the consultancy caused the entire reduction. A defensible account describes the work, the change, and the concurrent staffing decision. Stronger causal wording requires stronger evidence.

Use client explanations carefully. A manager may believe the project contributed substantially; attribute that view as a view and assess whether the surrounding wording turns it into an unsupported objective claim.

Define financial and operating results

For financial claims, distinguish gross benefits from benefits after implementation costs, projected savings from realized savings, and an annualized estimate from results observed over a stated period. Explain material exclusions and the basis of the calculation.

Time released for other work is not automatically cash saved. An avoided future expense differs from a reduction in current expenditure. A percentage reduction also differs from a percentage-point change: a rate falling from 5% to 4% falls by one percentage point, or 20% relative to its starting value.

Avoid mechanically extending a short favorable period into a full-year claim. Where an estimate is useful and supportable, label it as an estimate and disclose the assumptions that drive it.

Review expectations created by the whole presentation

A selected case establishes what occurred under its circumstances. To claim that clients typically achieve an outcome, examine the relevant population, comparability, missing results, selection, and distribution. One exceptional result cannot supply those answers.

ABA Model Rule 7.1’s commentary explains that an accurate account of a lawyer’s achievements can mislead if it creates unjustified expectations without the relevant factual and legal context. Keep that concern in view when drafting legal matter headlines and summaries, subject to the governing jurisdiction’s requirements. 

Testimonials need the same scrutiny. FTC staff guidance explains that exceptional-result endorsements can imply what others should expect, and a generic “results not typical” statement does not resolve that implication. The guidance calls for support for typicality or clear disclosure of generally expected performance in the circumstances. 

Permission to quote a client is one part of this review. Check the quote’s accuracy, any material relationship requiring disclosure, the claims it implies, and the impression created by its placement. Narrow or omit an endorsement when the necessary support is unavailable.

12.6 Use Approved Work Samples, Methodologies, and Other Evidence When Case Details Cannot Be Published 

When a real engagement cannot be discussed publicly, identify the useful question that remains answerable. A prospect may want to understand the firm’s method, the decisions it examines, the form of its deliverable, or how it communicates uncertainty.

• Methodology explanation. Describe the actual stages, inputs, analytical choices, and limitations of the firm’s approach, using material the firm has the right to publish.
• Annotated sample deliverable. Create a report, model, presentation, or process map with clearly fictional data. Explain what each section enables a client to decide.
• Approved template. Publish a blank checklist or planning document with enough explanation to show how it is used and where judgment is required.
• Process demonstration. Work through an example using invented data or public material cleared for the proposed use. Identify the sources and assumptions.
• Permitted assignment description. Where authorized, explain the scope and role without disclosing restricted results or client circumstances.

State what the asset demonstrates. A fictional board presentation may demonstrate a way of organizing analysis. It cannot establish that a client adopted the recommendations. A blank assessment template can explain a process but does not prove that the process has produced a particular outcome.

An approved excerpt should retain its purpose, status, and limitations. Removing the page that explains assumptions can make the surviving work sample misleading. Where meaningful interpretation requires information that cannot be published, create a separate illustration.

Publish an accessible explanation beside the asset: what it is, who prepared or reviewed it, the relevant date, the question it addresses, and its limitations. Link it to the appropriate service page. Follow Chapter 10’s access practices without weakening the protections on source files.

For a small firm, one well-explained method and one useful sample can be a sensible first publication batch. Expand when the firm has additional evidence that answers a distinct client question.

Practical Output: A Claims-and-Evidence Register and a Case-Study Brief

Select one priority service and examine the material claims already used on its website, proposals, and professional profiles. Start with claims about experience, outcomes, or comparative performance that could materially affect a client’s decision.

Build the claims-and-evidence register

Use one record per material claim. Keep its evidence status separate from its publication clearance. The following fictional entries illustrate the decisions the register should support.

For each entry, record the claim owner, source location and version, calculation or method, evidence reviewer, proposed publication format, necessary qualifications, and permission requirements. Link to the exact approved wording and assets, permitted channels, publication locations, review date, and change triggers.

Use separate status fields for evidence and clearance. For example, evidence may be incomplete, verified within stated limits, or insufficient; clearance may be pending, approved for specified use, or unavailable. Record publication and correction status separately. One general “approved” field can hide an unresolved question.

Complete a detailed case-study brief

The worked brief below uses a fully fictional consulting scenario. It can be published as an illustration once its stated checks are complete. A real engagement would require verified project records and the applicable client, professional, and rights clearances.

Purpose and audience: Help an operations leader assess the firm’s approach to diagnosing delays in an order-approval process.

Proposed title and format: “Illustrative case: assessing delays in order approval.” Label the page, measurement table, and any downloadable sample as fictional. This example is not evidence of a completed client engagement.

Context: An invented distribution business wants to understand why routine orders take too long to receive approval. The demonstration covers one approval workflow, rather than the business’s entire order-to-delivery process.

Assignment and roles: The consulting team maps handoffs, analyzes eligible records, interviews participants, and helps pilot revised routing. Client managers supply information, approve operating changes, and staff the process. The client adds a coordinator during the pilot.

Measurement specification: The invented baseline includes 180 eligible requests submitted during an eight-week window, and the follow-up includes 190 submitted during a separate eight-week window. The baseline cohort completes before the routing pilot begins. The follow-up cohort enters the process after the revised routing is in operation. Each cohort is followed to completion. Cancelled requests are excluded under the same rule. Turnaround runs from receipt of a complete request to final approval, measured using a consistent business-day calendar.

Result and calculation: The fictional median is 8.0 business days in the baseline and 6.0 in the follow-up, a 25% lower observed median. There is no comparison group, and the calculation does not isolate the routing change from the coordinator’s contribution or other differences between periods.

Proposed opening copy: “This fictional example shows how a consulting team might investigate delays in order approval. The team examines handoffs and pilots revised routing with client managers. In the invented data, median turnaround is lower in the follow-up period, while a concurrent staffing change limits what can be attributed to the routing intervention.”

Supporting assets: A synthetic process map, a table showing the measurement definitions and results, and an annotated recommendations page. The assets should demonstrate the reasoning, including unresolved questions.

Review and release: The practice lead checks plausibility, arithmetic, attribution, and consistency with the real service. The editor checks that the illustration label and limitations survive in summaries and downloads. Confirm that the demonstration uses no client-confidential or restricted third-party material.

For an actual engagement: Link each material factual or outcome claim to its register ID and supporting records. Identify the evidence reviewer, exact proposed wording and assets, intended channels, required approvers, and current clearance status. Keep the case pending until the records support the wording and the specified use is authorized.

Maintenance and next step: Assign a page owner and review date. Link to the relevant diagnostic service and its explanation of scope. Recheck the illustration if the methodology or offer changes.

Use the brief to commission an actual case

For a real project, gather the relevant records and determine the permissible format before commissioning the narrative. Put unresolved claims and permissions into the register. Send the complete proposed publication to the appropriate reviewers, including any client approver whose authorization is required.

Before release, ask a colleague to identify what the firm did, what the evidence shows, who else contributed, and what the case does not establish. Check the headline and short excerpts with the same questions. Resolve any material difference between that reading and the approved claim.

After publication, add the page to the existing monitoring program. If an assistant omits a limitation or attributes another participant’s work to the firm, inspect the source passages and strengthen the explanation where needed. Record the incident and the correction without treating one revised answer as proof that every future answer will be accurate.