Case study presentation: one structure for customer, interview and class cases, with three worked examples

Disclosure: NOXI publishes this guide and is one of the tools it mentions; the method works in any presentation software.

Three different people search for "case study presentation": a marketer turning a customer win into proof, a candidate given a take-home case before a job interview, and a student presenting a business case in class or at a competition. This guide gives all three one structure, then shows what changes for each, with a worked example per type. All three examples are fictional: Maple & Finch Dental Group, Larkspur Grocers and Harrow & Vale Coffee Roasters do not exist, and every calculation is shown so you can rerun it with your own numbers.

What is a case study presentation?

A case study presentation is a short talk with slides that walks an audience through one specific situation, the problem in it, the action taken or recommended, and the measured or expected result. Its purpose is proof: that a product works, that a candidate can think, or that an analysis holds up.

The same words hide three jobs. The audience decides something different in each, so the same five beats get different weight and different proof.

A case study presentation is not a company profile and not a business case. A company profile proves a company can deliver by summing up many projects; our company profile guide covers that deck and its one-slide project summaries. A business case presentation argues for funding a future investment. A case study goes deep on one situation, and in a customer case it has already happened.

The 8-slide case study presentation template

A complete case study presentation needs eight slides: a title that states the result, context, challenge, approach, results, what made it work, the lesson or risks, and the next step. The order stays the same for all three types; what goes on each slide changes.

Eight slides fit a 10 to 20-minute slot. For a 5-minute sales slot, merge slides 2 and 3, and 6 and 7; for a 30-minute class presentation, give the analysis two or three slides and move detail to an appendix. The general rules for ordering slides around one message are in our guide to how to structure a presentation; this template applies them to cases.

Three kinds of case study presentation: same beats, different weight

A customer case study spends most of its time on results, an interview case on analysis, and a class case on analysis and the recommendation. The five beats are the same; the share of slides and minutes given to each beat changes with what the audience is judging.

These weights are our guidance for allocating slides and minutes, not a published standard, and each row sums to 100%. Use them as a budget: in a 20-minute interview case, 35% is 7 minutes of analysis; in a 10-minute customer story, 35% is three and a half minutes on results. If your draft spends half its time on context, the audience is waiting for the part they came for.

Write the title as the result

The title slide of a case study should state the result or the recommendation as a sentence, with a number where there is one. "Maple & Finch cut no-shows by 37% in six months" tells the audience why to listen; "Case study: Maple & Finch" makes them wait.

The same applies to every slide title. Read the eight titles in a row: they should tell the whole story without the slide bodies. Consultants call these action titles, and our guide to consulting-style slides has more before-and-after pairs.

Two cautions. Round only once and the same way everywhere: if the title says 37%, the results slide says 37.3% and shows the arithmetic. And if the result is relative, say so somewhere on the slide, because a reader who sees "37%" next to "14.2% to 8.9%" will check the subtraction and find 5.3.

Context and challenge: two slides, not a company history

The context slide answers who, where and what was at stake in three or four facts; the challenge slide states the problem in numbers before anything changed. Together they should take no more than a quarter of the talk.

Context is the part presenters overfill, because it is the part they know best. Keep only the facts the audience needs to judge the result: size (9 clinics, 48,600 appointments in six months), the setting that makes the result relevant to them (multi-site, mixed private and insured patients), and the stake (empty chairs cost money and staff time). The founding story, the office photos and the full product list go elsewhere.

The challenge slide needs a number and its denominator: "6,901 of 48,600 booked appointments were no-shows, 14.2%". A challenge without a number cannot be compared with the result, and the audience will assume the result is soft too. In an interview case the challenge is what the data shows; in a class case it is the decision the protagonist faces, written as one question ("Which growth path should Harrow & Vale fund with $1.0 million over three years?").

Approach: show the decisions and the options you rejected

The approach slide shows what was done, in the order it was done, and at least one alternative that was considered and rejected with a reason. Rejected options are what make an approach look like judgement rather than a sales pitch.

• Customer case: what changed in the customer's process, who did it and over what dates. Maple & Finch rolled out two-way text reminders at 72 and 24 hours, one-tap rebooking and a waitlist that offers freed slots to other patients, live in all 9 clinics on 1 January 2026. It ruled out a cancellation fee because of patient complaints at a pilot clinic.
• Interview case: how you broke the problem down. Say what you checked first and why, then what you found. The panel scores the path as much as the answer.
• Class case: the options and the criteria you will judge them by, before any scoring. Naming the criteria first shows you did not pick them to suit your favourite option.

One slide is enough for most approaches. If the method has more than four steps, show it as a simple timeline and keep the detail in the appendix.

Results that hold up: the results ledger

A result holds up when it shows six things: the metric, the baseline, the after figure, the period with its denominator, the source of the data and, for a customer, the permission to publish it. A results ledger is a one-page table that holds those six fields for every number on your slides.

Six rules keep the numbers honest:

1. Show the baseline. "8.9% no-shows" means nothing until the audience sees 14.2% before.
2. Match the periods. Compare January–June with January–June, not winter with summer, so seasonality does not do the work.
3. Give the denominator. 4,557 of 51,200 appointments, not just a percentage.
4. Separate percent from percentage points. 14.2% to 8.9% is a fall of 5.3 percentage points and a relative reduction of 37.3%. Show both.
5. Check what else changed. Two Maple & Finch clinics added Saturday hours in March 2026. The seven clinics with no other change went from 14.0% to 8.9%, a relative reduction of (14.0 − 8.9) ÷ 14.0 = 36.4%, so the Saturday hours do not explain the result. Saying so earns more trust than the headline.
6. Name the source. A system export, a finance report, a survey with its sample size. "Internal data" alone is not a source.

A baseline is the value of a metric before the change being described, measured the same way and over a comparable period. A result is the difference between the baseline and the after value, reported as an absolute change (percentage points or units) and, where useful, as a relative change (percent of the baseline).

Worked example 1: a customer case study with full arithmetic

A customer case study presentation proves one claim with one customer's numbers: here, that two-way appointment reminders cut no-shows at a multi-site dental group. Every figure below appears in the results ledger above, so each slide can be traced to a source.

The arithmetic

Note the denominator trap in the "appointments kept" row. Bookings rose from 48,600 to 51,200, so comparing raw no-show counts (6,901 against 4,557 = 2,344 fewer) understates the effect. The fair comparison applies the old rate to the new volume: 7,270 no-shows were expected, 4,557 happened, 2,713 were avoided.

The slides (12 minutes)

The deck uses the eight-slide template with the results split over two slides, because results are what this audience came for.

1. Maple & Finch cut no-shows by 37% in six months (0:30): the result, the customer, the period.
2. Nine clinics, 51,200 booked appointments in six months (0:45): who they are and why a group of their size is a fair comparison for the room.
3. One appointment in seven was a no-show (1:15): 14.2% in January–June 2025, 212 front-desk calls a day.
4. Two-way reminders at 72 and 24 hours, with one-tap rebooking (2:00): what changed, rollout dates, the cancellation fee they ruled out.
5. No-shows fell from 14.2% to 8.9% (2:30): before/after bars with both numbers, 5.3 points and 37.3%, and the typical-result line.
6. 2,713 appointments kept and 64 fewer calls a day (2:00): the denominator-fair calculation and about $320,000 of chair time.
7. The result holds in the seven clinics where nothing else changed (1:30): three reasons it worked, and the 36.4% check.
8. "We got our Friday afternoons back." (0:45): the approved quote and the disclosure that Maple & Finch received three months of service free for taking part.
9. Run a 90-day pilot in three of your clinics (0:45): the next step, measured the same way.

The times add up to 0.5 + 0.75 + 1.25 + 2 + 2.5 + 2 + 1.5 + 0.75 + 0.75 = 12 minutes. Results (slides 5 and 6) take 4.5 of 12 minutes, 37.5%, close to the 35% budget for a customer case.

Here are two slides from this article's worked example, as NOXI built them, unedited.

Customer case studies: permission, logos, quotes and the typical-results rule

A customer case study used to sell is advertising, so it needs the customer's permission, evidence for every claim and, in the United States, a statement of the generally expected results when the featured customer did better than most. Put the typical result and any material connection on the results slide, not in a footnote nobody reads.

The U.S. Federal Trade Commission's guidance on endorsements is direct about exceptional results. An ad featuring an endorser who achieved exceptional results must make clear to the audience the generally expected results, and statements like "Results not typical" or "Individual results may vary" do not change how the claim is read. The advertiser must either hold adequate proof that the results shown are typical, or disclose clearly and conspicuously the generally expected performance in the circumstances shown (FTC, Endorsement Guides: What People Are Asking). The same guidance says an endorsement must reflect the honest opinion of the endorser, and that a connection the audience would not expect, such as payment or something of value given to the endorser, should be disclosed.

For Tidewell, that means one line on slide 5: "Typical result: clinics using Tidewell for at least six months cut no-shows by a median 22% (140 clinics, to 30 June 2026)." Maple & Finch's 37.3% is a strong result, and the line shows the audience what to expect if they are average. Because Maple & Finch received three months of service free for taking part, slide 8, the quote slide, says that too.

Other markets point the same way. The UK CAP Code requires marketers to hold documentary evidence that a testimonial is genuine, with the contact details of the person giving it (rule 3.47), and says marketing must not feature a testimonial without permission (rule 3.50) (ASA, CAP Code section 03). In the U.S., the FTC's rule on consumer reviews and testimonials, in effect since 21 October 2024, prohibits writing or creating fake or false consumer testimonials and authorises civil penalties for knowing violations (FTC, Consumer Reviews and Testimonials Rule: Questions and Answers). So never polish a quote into words the customer did not say, and never present an invented customer as real; this guide's examples are labelled fictional for that reason.

The permission pack

When a customer agrees to share results but not its name, describe it precisely ("a 9-clinic dental group"). A precise description is nearly as persuasive as a logo. The same evidence discipline applies to every claim in a company deck; the claims ledger in our company profile guide is the companion to the results ledger here.

Interview case presentations: the 20-minute plan

An interview case presentation should state the recommendation in the first minute, spend about a third of the time on analysis, and end with risks and a first step. For a 20-minute slot that is 1 minute for the answer, 3 for context and assumptions, 2 for the challenge, 7 for the analysis, 4 for options and the recommendation, and 3 for risks and next steps.

Our job interview presentation guide covers the 10-minute task presentation, the "present yourself" brief and the 30-60-90 day plan. This section covers the longer take-home case, where you receive a brief and a data pack days ahead and present for 20 minutes plus questions.

The minutes add up: 1 + 3 + 2 + 7 + 4 + 3 = 20, and the shares match the interview weights above (context 15%, challenge 10%, analysis 35%, recommendation 25%, risks 15%). Rehearse to 18 minutes so nerves and one question mid-talk do not push you over; if you run long, cut from the analysis, never from the recommendation or the risks.

Worked example 2: Larkspur Grocers (fictional)

The analysis splits the fall into its parts. New-customer orders held at 9,000 a month, so all of the fall is in repeat orders: 33,000 down to 29,640, a loss of 3,360. Repeat orders are customers times frequency.

The pack also shows that late deliveries rose from 4% to 11% of orders after a move to a new depot in July. Of the 750 repeat customers who lapsed, 480 (64%) had at least one late delivery in July–August, against 2,850 of the 14,250 who stayed (20%). That is an association, not proof of cause, and the slide says so; it is still the strongest lead in the pack.

The options slide compares three moves. A 15% win-back discount to lapsed customers treats the symptom. Two extra vans at the new depot, at $9,500 a month each, treat the likely cause: if frequency returns to 2.20, that recovers 14,250 × 0.12 = 1,710 orders a month, worth 1,710 × $14 gross margin = $23,940 a month against $19,000 of cost. Break-even needs 19,000 ÷ 14 = 1,357.1, so 1,358 orders a month, 79% of the 1,710-order frequency gap. The recommendation: add the vans now, measure on-time delivery weekly, and send a targeted credit only to customers who had a late delivery.

The assumptions slide names three things the candidate could not check: that the $14 margin applies to recovered orders, that new-customer orders are unaffected by the depot, and that late deliveries drive lapsing rather than both having a third cause. The risks slide says what would change the answer: if on-time delivery recovers and frequency does not, the problem is elsewhere, and the week-1 action is to read 50 lapsed customers' support tickets.

Two rules specific to interview cases. Use only the data the employer gave you plus public information, and never bring confidential material from a current or past employer. And if you use AI tools to prepare, follow the employer's rules; our job interview guide has a section on what is fair when using AI.

Class and competition cases: options, criteria and a weighted decision matrix

A class or competition case presentation should name the decision, set out two to four options, judge them against explicit weighted criteria using numbers from the case, recommend one, and test whether the recommendation survives a change in the weights. Judges reward a defensible recommendation more than a confident one.

Business-school cases are written to end at a decision. Harvard Business School describes each case as a 10 to 20-page document written from the viewpoint of a real person leading a real organization, ending in a key decision to be made, and its MBA students read about 500 cases over two years (Harvard Business School, The HBS Case Method). Your presentation takes the protagonist's chair: what should they decide, and why?

Worked example 3: Harrow & Vale Coffee Roasters (fictional)

The team sets four criteria and their weights before scoring, then writes a scoring rule for each so that every option is scored the same way:

• Year-3 operating profit (40%): 5 = $400,000 or more; 4 = $300,000–399,000; 3 = $200,000–299,000; 2 = $100,000–199,000; 1 = under $100,000. Case figures: A $310,000, B $240,000, C $420,000.
• Fit with current capabilities (30%): 5 = uses what the company already does every day; 1 = needs mostly new skills. B uses the roastery as it is (5); A needs subscription marketing (3); C needs site selection and more café managers (2).
• Capital required (20%): 5 = under $100,000; 4 = $100,000–249,000; 3 = $250,000–499,000; 2 = $500,000–999,000; 1 = $1.0 million or more. A $350,000, B $180,000, C $900,000.
• Risk (10%): 5 = no single factor could cut year-3 profit by more than a quarter; 1 = one factor could wipe it out. A 5; B 2 (three office clients would take 60% of wholesale volume); C 3.

Each cell is score × weight: for B, 3 × 0.4 + 5 × 0.3 + 4 × 0.2 + 2 × 0.1 = 1.2 + 1.5 + 0.8 + 0.2 = 3.7. B wins by 0.1, which is close enough that the presentation must say so.

The sensitivity check shows why. Move 10 points of weight from fit to profit (50/20/20/10) and the totals become A 4 × 0.5 + 3 × 0.2 + 3 × 0.2 + 5 × 0.1 = 3.7, B 1.5 + 1.0 + 0.8 + 0.2 = 3.5 and C 2.5 + 0.4 + 0.4 + 0.3 = 3.6. A now wins. So the recommendation slide reads: "Start with office wholesale: it scores highest at our weights, needs only $180,000 and uses the roastery as it is. If the founder values year-3 profit above fit, the subscription is the better choice." The judges see that the team knows exactly which judgement decides the case.

For a competition, read the rulebook first: it sets the time, the format, whether an appendix is allowed and how judges score. A common shape is the eight slides above plus an implementation timeline and a financial summary, with backup slides for the questions you expect. If the case asks you to compare competitors, our competitive analysis presentation guide covers the comparison slides; if it asks for an investment decision, the business case guide covers payback and NPV.

UX and portfolio case studies: process, with an outcome

A UX or portfolio case study presentation shows how you worked through a design problem and what changed because of it: the problem, your role, the research, the key decisions with the options you rejected, the outcome, and what you would do differently. Hiring panels want to see your judgement, so decisions matter more than deliverables.

The five beats still apply, with two adjustments. State your role precisely ("I led the research and the checkout redesign; a second designer built the components"), because panels ask, and vague ownership reads as inflation. And end the process with an outcome, even a modest one ("checkout completion rose from 61% to 66% over the four weeks after launch, measured on 18,400 sessions"); if there is no measured outcome, say what you would have measured. Show two or three decisions in depth rather than every wireframe. Client work under a non-disclosure agreement needs the client's permission or anonymised screens, the same rule as a customer case.

From case study to a one-slide proof point in a sales deck

A one-slide proof point compresses a case study into four lines: the customer (or an exact description), the result with baseline and period, how it was achieved in one sentence, and the typical result. Sales decks use it to answer "has this worked for someone like us?" without a full story.

• Customer: Maple & Finch Dental Group, 9 clinics.
• Result: no-shows 14.2% → 8.9% (−37.3%), January–June 2026 against the same months of 2025.
• How: two-way reminders at 72 and 24 hours with one-tap rebooking.
• Typical result: median 22% fewer no-shows across 140 clinics.

Keep a library of these slides by industry and company size so a salesperson can pick the two closest to each prospect. Our guide to building a sales deck shows where proof slides sit in the sequence.

Visuals that make a case study believable

Three visuals carry most case studies: a before/after bar chart with both values labelled, a short timeline of what changed and when, and one quote set large with the speaker's name and title. Each should show one point, and the slide title should state it.

A bar chart of 14.2% against 8.9% with the axis cut at 8% would show bars of 6.2 and 0.9 points above the axis, so the drop would look like 85%, and an attentive prospect will notice. Our guide to presenting data on slides covers axis rules, labels and chart choice in more depth.

Q&A: the questions each audience asks

Each case type draws predictable questions, so prepare an answer, and ideally a backup slide, for each one before you present. The table lists the questions most likely to come up and what to have ready.

Common case study presentation mistakes

Most case study presentations fail on proof and order, not on design: results without a baseline, a favourable period picked after the fact, a long company introduction, and the result held back until the last slide. Each has a quick fix.

Pass/fail checklist before you present a case study

Run this check before every case study presentation. Each line passes only if the stated condition is true.

3 AI prompts for a case study presentation

Use them in NOXI after attaching your source material, or in any AI tool that accepts files. Replace the brackets and keep the instruction to use your numbers unchanged: an AI tool should arrange your results, never supply them.

How to build a case study presentation in NOXI

NOXI turns the material behind a case study (interview notes, a results export, a project report or the case PDF) into a structured deck in your own style. It is free, with no card and no watermark; uploading files needs a free account, and a full deck uses about 25 of the 1,000 monthly credits.

Frequently asked questions

How do you present a case study?

Present a case study in five beats: the context, the challenge in numbers, what was done or what you analysed, the result or recommendation with its baseline and period, and the lesson or next step. State the result in the title of slide 1, prove it on the results slide, and give the audience time for questions. Spend the most time on the beat your audience is judging.

What is the format of a case study presentation?

The usual format is eight slides: a title that states the result, context, challenge, approach, results, what made it work, the lesson or risks, and the next step. Customer cases put about 35% of the time on results; interview and class cases put about 35% on analysis. Each slide title should be a full-sentence finding, so the titles alone tell the story.

How many slides should a case study presentation have?

Eight slides suit most case study presentations of 10 to 20 minutes. A five-minute sales version can merge them into four or five slides, and a one-slide proof point works inside a sales deck. A 30-minute class or competition case can run 10 to 12 slides with an appendix of backup slides for questions.

How do you present a case study in a job interview?

State your recommendation in the first minute, then give context and your assumptions, the problem in numbers, your analysis, two or three options compared with numbers, and risks with first-week actions. For a 20-minute slot that is 1, 3, 2, 7, 4 and 3 minutes. Use only the data the employer gave you plus public information, and rehearse to 18 minutes.

How long should a case study presentation be?

Most case study presentations run 10 to 20 minutes plus questions. Interview take-home cases are often 20 minutes with 10 minutes of questions, class presentations follow the professor's brief, and competitions follow the rulebook. A customer case at an event usually fits 10 to 15 minutes; inside a sales meeting it shrinks to one or two slides.

What should be included in a customer case study?

A customer case study should include who the customer is, the problem in numbers before the change, what changed and when, the result with its baseline, period and denominator, what made it work, one approved quote, and a next step. When the result is better than typical, add the generally expected result, and disclose anything of value the customer received.

Can I use a customer's logo or quote in a case study?

Only with the customer's permission for that use, ideally in writing, and with the exact approved words for a quote. The UK CAP Code says marketing must not feature a testimonial without permission and requires evidence that it is genuine. Without permission, describe the customer precisely instead, for example "a 9-clinic dental group".

How do you structure a case competition presentation?

Open with your recommendation, name the decision as one question, set out two to four options, compare them in a weighted decision matrix with criteria and weights stated, then show implementation, financials and risks. Add a sensitivity check that shows whether the answer survives a change in the weights. Follow the competition's rulebook for time, format and appendix rules.

Can AI make a case study presentation?

Yes. AI presentation tools can turn interview notes, a results spreadsheet or a case PDF into a structured deck, write action titles and lay out the charts. Give the tool your own numbers and tell it to invent none, then check every figure against your results ledger. For interview cases, follow the employer's rules on AI use.

Sources & methodology

How this guide was made: it combines the U.S. Federal Trade Commission's guidance on endorsements, testimonials and advertising substantiation, the UK CAP Code's rules on testimonials and Harvard Business School's description of its case method with practitioner method for the slides. The five-beat structure, the beat weights, the eight-slide template, the results ledger, the 20-minute interview plan, the permission pack and the checklist are editorial recommendations, not published standards. Maple & Finch Dental Group, Tidewell, Larkspur Grocers and Harrow & Vale Coffee Roasters are fictional, as are all their figures; every calculation was checked. All sources were checked on 16 October 2026.

• [1] Ads featuring exceptional results must make clear the generally expected results; "Results not typical" does not change the interpretation; endorsements must reflect the endorser's honest opinion; unexpected connections such as payment should be disclosed; the Guides were revised in 2023 — U.S. Federal Trade Commission, FTC's Endorsement Guides: What People Are Asking (June 2023).
• [2] The rule took effect on 21 October 2024, prohibits writing or creating fake or false consumer testimonials, and authorises civil penalties for knowing violations — U.S. Federal Trade Commission, Consumer Reviews and Testimonials Rule: Questions and Answers (November 2024).
• [3] Advertisers need a reasonable basis, meaning objective evidence, for claims before an ad runs; consumer endorsements must reflect typical experience or disclose it; material connections must be disclosed — U.S. Federal Trade Commission, Advertising FAQs: A Guide for Small Business.
• [4] Rule 3.47 (documentary evidence that a testimonial is genuine, with contact details) and rule 3.50 (no testimonial without permission) — Advertising Standards Authority, CAP Code section 03: Misleading advertising.
• [5] Each case is a 10 to 20-page document written from the viewpoint of a real person leading a real organization and ends in a key decision; about 500 cases read in the two-year MBA — Harvard Business School, The HBS Case Method.

Build your case study presentation in NOXI, free: attach your notes, results or the case PDF, review the eight titles in Plan mode, check every number against your results ledger, and export an editable PowerPoint or a PDF.