Media Buyer Portfolio Example: Case Structure and Review Checklist

A media buyer portfolio should prove how you make decisions, not merely that you have seen an advertising dashboard. A strong case lets a reviewer understand the business question, your scope, evidence quality, alternatives, risk controls, action, and learning in about ten minutes. Use this page as a build template; never copy the illustrative wording as if it described your results.

What belongs in the portfolio

Aim for two to four cases with different capabilities:

• a campaign or funnel diagnosis;
• a creative testing system;
• a measurement or reconciliation problem;
• a controlled scaling decision;
• an operating-process improvement.

Choose depth over a long gallery. If five cases all say “launched campaigns and improved ROAS,” the portfolio proves neither range nor method.

The one-page portfolio index

Begin with a scannable index:

[Name] — Media Buying Portfolio
Focus: [channels, product context, career level]
Working context: [in-house / agency / affiliate / projects]

Case	Decision	Skills shown	Evidence type
[Case 1]	[Stop, iterate, or scale question]	[diagnosis, measurement]	[real anonymized / simulated]
[Case 2]	[Creative coverage question]	[briefing, experimentation]	[real anonymized / public exercise]
[Case 3]	[Tracking or process question]	[QA, stakeholder work]	[real / simulated]

Add a confidentiality note and direct links. Do not force reviewers through a single oversized slide file if cases can be opened independently.

Annotated case template

1. Case label

State whether the case is real, anonymized, reconstructed with permission, or simulated. Add the period and your role.

Case status: simulated learning project using a public product journey. No live ads were run and all outcomes below are decision scenarios, not achieved performance.

This label prevents a polished exercise from being mistaken for client work.

2. Direct answer

Open with two sentences: what decision was needed and what you recommended.

The team needed to decide whether weak qualified acquisition came from media delivery or the post-click flow. I recommended holding audience and creative stable, repairing mobile form instrumentation, and running a bounded landing-page test before increasing spend.

The reviewer now knows the point of the case.

3. Context and economics

Describe product, customer, business model, funnel, primary event, and relevant constraint. Use ranges or indexed values when exact data is confidential. Explain whether the metric is platform-reported, analytics-reported, approved, or financial.

Include only context that changes the decision. A long company history distracts from your work.

4. Scope and collaborators

State what you owned, influenced, and could not control.

• Owned: campaign structure, pacing, creative test brief, readout.
• Collaborated: analytics event QA, landing change, claim review.
• External constraint: advertiser validation delay.
• Not my work: original product design and production execution.

This prevents accidental credit for a team's entire result.

5. Evidence map

List the evidence and reliability:

Evidence	Source	Reliability or limitation	Decision use
Delivery and click	Ad platform	Fast; attributed	Detect delivery change
Form event	Analytics	Mobile error gap found	Do not use until repaired
Approved outcome	CRM or partner	Delayed; commercial truth	Evaluate quality
User feedback	Research sample	Directional	Develop message hypothesis

Evidence maps distinguish good reasoning from confident storytelling.

6. Hypotheses and alternatives

Write your leading hypothesis and at least one plausible alternative. State what observation would separate them.

H1: mobile form friction reduced qualified completion. If true, field errors should rise on mobile while pre-form behavior remains stable.
H2: audience quality changed. If true, decline should appear before the form and vary by segment.

Do not rewrite the hypothesis after seeing the result without recording the change.

7. Test and safeguards

Specify fixed variables, changed variable, audience, sample or observation period, maximum exposure, stop conditions, policy approval, tracking QA, and maturation window. Explain why the design was proportionate.

For a simulated case, provide scenarios rather than invented outcomes:

• If qualified completion improves and downstream quality holds, extend the change.
• If form completion improves but approval falls, inspect source or expectation mismatch.
• If neither changes, deprioritize the form hypothesis.

8. Result

Separate observation, interpretation, and action.

• Observed: what the source actually recorded.
• Uncertain: what was delayed, missing, or confounded.
• Interpretation: the explanation currently best supported.
• Decision: stop, iterate, scale, or instrument.
• Review: owner and date for the next decision.

Do not decorate a weak design with a precise percentage.

9. Artifacts

Show the work behind the conclusion:

• funnel or event map;
• hypothesis card;
• creative taxonomy and sample brief;
• launch QA checklist;
• annotated chart with source and definition;
• decision log;
• forecast scenarios;
• post-test readout.

Remove tokens, account IDs, personal data, raw customer rows, and identifying client details. Screenshots should be cropped, explained, and readable.

10. Reflection

Answer: What would you do differently with today's knowledge? Which assumption was wrong? What capability or control remained after the campaign? Honest reflection is stronger than a case that pretends every decision was obvious.

Complete fictional outline

Use this outline only as a structural example:

Title: Diagnosing activation decline without increasing acquisition spend
Status: simulated; no achieved results claimed
Decision: repair measurement and test one mobile-form change before scaling
Context: subscription product; activation delayed; support capacity limited
Scope: media structure, QA, analysis, test recommendation
Evidence: platform delivery stable; analytics form errors uncertain; CRM value delayed
Alternatives: audience-quality shift versus form friction
Safeguard: spend held; approved claim unchanged; rollback defined
Scenarios: three outcome branches and actions
Artifacts: event map, QA log, annotated mock readout
Reflection: instrumentation should have been validated before the original page release

Because no real campaign ran, the outline contains decisions and scenarios, not a fake uplift.

A junior portfolio

Junior candidates can use public journeys, platform sandboxes, open datasets with licenses, or fictional briefs. Show calculations from raw inputs, QA thinking, original creative directions, and action branches. Clearly label what is observed versus assumed.

A junior case can be excellent without a large budget. Reviewers are looking for clean reasoning, basic math, responsible controls, and coachability.

An experienced portfolio

Experienced buyers should show ambiguity, downstream value, forecast or marginal economics, cross-functional decisions, and a process improvement. Replace a wall of account screenshots with the evidence chain.

If confidentiality prevents any number, add other verifiable detail: decision cadence, event map, role boundary, artifact, and an approved indexed change. Obtain permission before using employer work.

Format choices

A simple web page or accessible PDF works. Use selectable text, clear headings, high contrast, meaningful link labels, and charts with labels. Keep a PDF fallback if the site needs authentication. Test on mobile and in a private browser.

Do not use an interactive visual if it makes the decision harder to find. The portfolio itself demonstrates communication judgment.

Reviewer scorecard

Score each case from zero to two:

• decision is clear;
• scope is honest;
• metric definition is present;
• evidence and limits are visible;
• alternatives were considered;
• downside was controlled;
• result is not overstated;
• artifacts support the story;
• reflection changes future practice;
• confidentiality is protected.

Cases below 14 of 20 need revision. A missing result is acceptable in a simulated case; an invented result is not.

Final checklist

1. Every case has a status label.
2. The recommendation appears before detail.
3. Your contribution is separated from team work.
4. Numbers have period, source, and definition.
5. Simulation never looks like achieved performance.
6. Internal and personal data are removed.
7. Links work without your logged-in session.
8. Resume claims match portfolio details.
9. Each case shows a different capability.
10. A reader can explain your decision after ten minutes.

The main media buyer portfolio guide covers selection and narrative strategy. This page serves the separate need for an annotated example and buildable case structure. Pair it with the resume template so application claims lead to evidence rather than repetition.