Present project work and evidence without exaggeration

Choose work you are allowed to show

Start with completed work or a project mature enough to describe accurately. Check which pages, screenshots, photographs, names and quotations may be published and with what attribution. If permission is missing, anonymise the case or describe a general method without presenting it as a named client reference.

Choose examples that differ by the problem addressed and deliverable, rather than building a gallery of similar thumbnails. One page might explain an editorial redesign; another could show a better contact journey. Diversity should come from real work, not invented categories.

• Project, period and permissions recorded.
• Image rights confirmed.
• Distinct problem for each selected case.

Explain the starting point and scope

Describe the organisation and its need at the level of detail it has approved. What was difficult for visitors or the internal team? Which content, tools and constraints already existed? A reader should understand why the work was requested before seeing the solution.

Define your role precisely: advice, copy, design, development, migration, maintenance or coordination. Do not claim work carried out by the client or a partner. Name important dependencies, such as receiving photographs or having copy approved.

• Observable starting situation.
• Exact project scope.
• Client and partner contributions.

Show decisions and their reasons

Pick two or three decisions that demonstrate the work: reorganising navigation, clarifying a service or simplifying a form. For each, explain the alternative considered and why the final option was chosen. A before-and-after image is useful when the two states are dated and comparable.

Explain what was outside scope when that helps readers interpret the result. A faster site may depend on better prepared images; a new form alone does not prove more sales. Connect every illustration to the step being described.

• Two or three explained decisions.
• Contextualised and captioned visuals.
• Project limits visible.

Report outcomes with their measurement method

When publishing a number, give its period, metric and source. Compare reasonably similar situations and mention other changes that prevent you from attributing every effect to the website. Without reliable measurement, describe a verifiable deliverable instead: pages released, a journey tested or a procedure handed over.

Testimonials should come from the named person and be used with permission. Avoid unconfirmed ratings, percentages and client logos. A modest documented result is more useful to a prospective client than a spectacular claim that cannot be checked.

Fictional teaching example without client data: compare two 30-day periods using the same definitions of session and confirmed enquiry. Count sessions containing at least one enquiry, not every repeated trigger. In GA4, session key event rate divides sessions containing a key event by total sessions; the selected event must still be defined and checked to ensure it represents the enquiry claimed.

With 40 enquiry sessions out of 2000 sessions, the rate is 2%. With 60 out of 6000, it is 1%. The volume of enquiry sessions rises by 50%, while the rate falls by one percentage point, or 50% relatively. These calculations do not establish a redesign’s effect: campaigns, audiences, seasonality or collection may have changed. Publish the denominator and limitations alongside the outcome.

• Metric, period and source for each figure.
• Quotation and identity approved before publication.
• Unmeasured outcomes described honestly.

Scroll the table horizontally to read every column. With a keyboard, focus the table area and use the arrow keys.

• Google Analytics : Traffic acquisition

  Consulted on 22 October 2026; event count and session rate are distinct. The table uses fictional figures.

Connect the case to the next decision

End with what was delivered, what remains the client’s responsibility and which lessons could apply to a similar project. Link to the relevant service and offer a simple way to discuss a comparable need without suggesting that the same result is guaranteed.

Review case studies when the featured site changes, screenshots age or permission is withdrawn. Evidence should remain true over time. The editorial maintenance guide can include this check.

• Deliverables and limitations summarised.
• Link to the relevant service.
• Review date and approver named.

Separate delivery, use and commercial outcome

A delivered page, a usable form and an increase in enquiries are distinct outcomes. For each, state your role, period and available evidence. With only a screenshot, describe the interface; do not infer a conversion rate or financial result.