Public Relations Strategy Case Studies
Marketing5 Case StudiesUpdated Sep 2026
These five cases come from the working life of a small agency and the in-house teams it serves. Each begins with a constraint that tidy models tend to skip: a fixed budget, a brand that has drifted from its own message, a site that still gets visitors but stops converting them, a comment thread that becomes a news story, or a library of articles nobody has reviewed. The organizations are invented, but the decisions are ordinary. Notice what evidence each team had, and who absorbed the cost when plans met reality.
Case Study 1
Spring launch of a tent line on a fixed budget
Larkspur Outdoor Co., a regional outdoor retailer with six stores and a small online shop.
The situation
Larkspur Outdoor was preparing to launch a lightweight tent line in the first week of spring. The marketing team had 48,000 dollars for the entire launch, covering paid search, paid social, two regional radio flights, sampling at trail events, and measurement. Store traffic was strong in the home metro but thin elsewhere, and the online shop had never sold more than a few dozen units of any single product in a week. Leadership wanted the tent line to carry its own weight within the season, and the category manager wanted the launch to avoid pulling attention from the stronger winter jacket sales. Nobody on the team had clean data on which channel had actually driven tent purchases before.
The challenge
The team had to choose which two or three audiences and channels deserved the money, and set conversion goals that could be measured without a long attribution argument. Each option looked reasonable on paper. Radio reached the home metro cheaply but could not be tied to online sales. Paid social promised scale, but its cost per click had been rising. Trail sampling built awareness among the right people but produced almost no direct revenue. Choosing wrongly would leave the launch underfunded in the place where it could convert.
Analysis
The first question was what the launch had to prove. The category manager cared about tent units sold in the season, while leadership cared about whether the line would justify a second production order. Those targets are different, and they point toward different channels. Teams grounded in Public Relations Strategy usually start by writing down the single conversion goal the budget must reach, then working backward to the audiences that can plausibly reach it. The evidence was thin. Tent sales in the previous season were too few to support a reliable forecast, so the team used winter jacket buyers as a proxy for outdoor shoppers who already spent with the brand.
The operational limits were tighter than the media plan suggested. Store staff could only run sampling at two events a month, the online checkout had a known slowdown on mobile, and the retailer's legal reviewer needed ten business days to approve radio copy. Each limit shaped the mix. Stakeholder impact mattered too. Store managers would absorb the extra foot traffic and any returns, and the jacket team would notice if tent promotions drew attention and shelf space away from their own range. A channel plan that ignored these people might have won the spreadsheet and lost the building.
- 1Set the goalDefine one tent-unit target before choosing any channel.
- 2Test audiencesMatch trail and camping shoppers to each channel's real reach.
- 3Fund the controlsHold one radio week back for comparison against store sales.
- 4Checkpoint and shiftMove unspent paid social money to search at the sixth week.
Actions taken
- Cut paid social to a single test audience of trail-running and camping shoppers within a 90-mile radius of each store, with a fixed daily cap and a weekly review of cost per add-to-cart.
- Fund one radio flight in the home metro only, with a unique promo code read on air, so store and online sales can be compared against a control week.
- Hold sampling to two high-traffic trail events, and have store staff log every tent conversation on a shared sheet rather than guessing at impact afterward.
- Set the conversion goal at 1,400 tent units across the season, with a go or no-go checkpoint in the sixth week that reallocates any unspent paid social money to search.
Results
By the end of the season the launch sold 1,180 tent units, short of the 1,400 target but enough to justify a reorder of the two best sellers. Paid search delivered the cheapest cost per sale, and the controlled radio week showed a lift in online orders from the home metro that the team could attribute with reasonable confidence. Paid social was the weakest performer after its test, and its remaining budget moved to search at the checkpoint. Trail sampling produced few direct sales but generated the email list the team used for the autumn relaunch. Store managers reported longer staffing hours than planned, which led to a revised event schedule for the next launch.
Questions for discussion
- Which of the four channels would you have cut first, and what evidence would have changed your mind?
- How should a team set a conversion goal when the previous season offers no reliable baseline?
- Where do store staff sit in a launch budget, and should their time be costed as a channel?
Case Study 2
Repositioning a cookware brand seen as too expensive
Brindle and Oak Kitchenware, a mid-size cookware brand sold through department stores and its own website.
The situation
Brindle and Oak had spent four years describing itself as the professional-grade choice for serious home cooks. Its marketing leaned on chef endorsements, heavy copper finishes, and product pages written in technical language. Sales were steady in the premium department store channel, but a growing share of visitors to the brand's own site left after seeing the price of a single pan. A survey of 3,200 recent buyers showed that most people saw the line as expensive and impressive, but rarely as something they would use every day. The product team believed the quality claims were accurate. The problem was that customers did not hear them the way the brand intended.
The challenge
The brand needed a proposition that matched what its engineering could actually deliver, that customers would believe after seeing it, and that did not force a price cut across the range. Leadership wanted evidence before committing to a new message across packaging, the website, and department store counters. Testing had to happen quickly, because the spring catalog was already in print and store partners wanted a decision before their floor resets.
Analysis
The business objective was narrower than a full rebrand. The brand wanted to lift conversion among shoppers who already considered it, and to stop losing them at the price point. That framing changed which evidence mattered. The survey showed that buyers valued durability and even heat distribution, but doubted they would use the pans daily. Purchase data supported the doubt. Repeat orders for the everyday skillet were low, while the sauté pan sold through quickly at department store counters with live demonstrations. The evidence pointed toward everyday use rather than professional prestige, which is a different promise and a different story about price.
The constraints were practical. Packaging for the spring range was locked, so any new claim had to fit existing printed boxes or wait for the next season. Department store partners controlled their counter signage and would accept new copy only if it matched the floor demonstrations. Engineering could not change the metal finish without a six-week supplier cycle. Stakeholder impact was uneven. Store staff who demonstrated the products would carry the new story at the counter, retail buyers would judge it by sell-through, and the chef partners who had built the old identity needed to hear about the change first rather than read it in a press note.
- First three weeksTest two product page framings against the same traffic.
- Weeks four to sixPilot everyday framing at two department store counters.
- Before public changeBrief chef partners and store staff on what is changing.
- End of seasonReview repeat orders before any packaging revision.
Actions taken
- Run a two-cell test on the product pages, one keeping the professional language and one leading with everyday meals, and compare add-to-cart rates over three weeks.
- Pilot the everyday framing at two department store locations with demonstration counters, using shelf cards that name a weeknight meal rather than a technique.
- Brief the chef partners and store staff before the change goes public, with a one-page note explaining what stays the same and what the new claim does not promise.
- Keep the price structure unchanged, and hold any packaging update for the autumn range until the supplier signs off on the finish trial.
Results
After three weeks, the everyday framing lifted add-to-cart rates on the product pages by a modest margin, enough to keep it as the primary message. The two pilot counters reported stronger sell-through on the sauté pan than comparable locations, though the sample was small and the team said so in its report. Repeat-purchase tracking would take a full season to read, so the autumn decision rests on the early signal and the counter feedback. Chef partners responded well to the advance briefing, and two of them offered to appear in the next demonstration video. Packaging stayed as printed for spring, with the change scheduled for the autumn range.
Questions for discussion
- When a survey and purchase data point in different directions, which one should set the message, and why?
- How much of a brand's identity can change before the people who built the old one start to feel left out?
- What would you need to see in repeat-purchase data before committing to a packaging change?
Case Study 3
Investigating a checkout drop with steady traffic
Mosswood Home Goods, an online retailer of bedding and small furniture with a warehouse in one region.
The situation
Mosswood Home Goods had seen visitor numbers hold steady for most of the year, with about 84,000 monthly sessions from search, email, and paid channels. Revenue, however, had fallen by roughly a fifth over the previous two quarters. The marketing director suspected that the homepage redesign launched in the third month had weakened the path to product pages, while the paid media lead blamed a shift in ad spend toward broader audiences. Finance wanted a clear answer before approving the next quarter's experiment budget, and engineering was already stretched by a warehouse software migration. Nobody had yet mapped where shoppers were actually leaving the site.
The challenge
The task was to find where conversion was failing before anyone ran an experiment. Attribution was disputed. The analytics platform credited a large share of orders to last-click search, while the email platform claimed its own share, and the two totals did not reconcile. Category page content had been changed in several places, and a payment provider update had gone live in the same window. The team needed to rank the candidate causes by evidence rather than by who argued loudest.
Analysis
The business objective was to recover revenue without buying more traffic, since acquisition costs had already risen. That made the shopping journey the first place to look. The team pulled step-by-step funnel data for the same weekly cohorts, before and after the homepage change, and split it by device. The sharpest drop sat on mobile at the shipping selection step, not on the homepage, which made the redesign less likely to be the main cause. Attribution disputes were real, but they mattered less here, because the drop appeared inside individual sessions regardless of which channel brought the visitor. Teams that start with channel totals tend to argue about credit instead of finding the leak.
The operational limits shaped what could be tested. Engineering had a freeze on the storefront until the migration finished, so the first fixes had to be configuration changes or content edits. Paid media could not pause spend without a conversation with finance, because a fixed monthly floor had already been committed. Customer service would absorb any rise in complaints about delivery estimates, and the warehouse team would face more orders if shipping prompts pushed buyers toward expedited options. Each group had a stake in the answer, so the findings had to be reported in terms each of them could act on.
Actions taken
- Rebuild the funnel report for the last eight weekly cohorts, split by device and by shipping option selected, so the drop can be located to a single step.
- Pause category page copy edits until the funnel review is complete, so that any recovery can be attributed to one change rather than several overlapping ones.
- Reconcile the two attribution totals with a shared order identifier, accepting that the reconciled view will be imperfect but consistent across finance and marketing reports.
- Schedule a mobile shipping estimate experiment for the first week after the storefront freeze lifts, with a holdout of one fifth of sessions.
Results
The funnel review located the largest drop at the shipping selection step on mobile, where a delivery estimate appeared only after the buyer had chosen a method. The payment provider update was checked and found to have no measurable effect on completion rates. The mobile shipping experiment ran after the freeze lifted and recovered a meaningful share of the lost checkouts, though not all of them. Finance accepted the reconciled attribution view for planning, and customer service reported fewer delivery complaints in the following weeks. The homepage redesign stayed in place, with a smaller follow-up change to its category links scheduled for later.
Questions for discussion
- What evidence would have been enough to rule out the homepage redesign without running a test on it?
- How should a team decide which of several overlapping changes to isolate first when engineering capacity is limited?
- When attribution totals disagree, what does a shared identifier cost, and who should pay for it?
Case Study 4
A taproom video escalates into a public dispute
Tallow Creek Brewing, a regional craft brewery with three taprooms and a distribution contract in four states.
The situation
Tallow Creek Brewing posted a short video from its main taproom showing a staff member refusing service to a regular customer. Within a day, the clip had been shared by a local food account and then by a regional news commentator, and the comment thread had split between people who said the staff member was right and people who said the brewery had a policy problem. The brewery had no crisis plan beyond a general social media guideline. Its social lead had posted a reply that was accurate in one detail but dismissive in tone, and that reply was now being screenshotted. Two of the three taprooms were open that evening, and a national distributor had asked for a statement by the next morning.
The challenge
The brewery needed to establish what had actually happened, correct the record where its reply had gone wrong, and keep staff, customers, and the distributor informed without speculating. The facts were incomplete. The owner had not yet spoken with the staff member, and the security footage was held by the landlord of one taproom. Every hour without a verified account gave the loudest voices more room to define the story.
Analysis
The business objective was to protect the taproom relationships and the distributor contract without treating a single clip as the whole story. The first step was separating verified facts from claims. The team pulled the point-of-sale record, which confirmed the time and the order that had been refused, and requested the full video from the customer who had filmed it. The staff schedule showed who was on shift, and the taproom manager confirmed the house policy on intoxication, which was written down but rarely shown to customers. Teams that handle this well sort evidence into what is confirmed, what is disputed, and what is still unknown, and they publish only the first category until the others close.
Operational constraints narrowed the options. The brewery could not release the landlord's footage without the landlord's consent, and legal wanted any statement about the staff member reviewed before it went out, which meant no personnel details could be shared publicly. Stakeholder impact was wide. The staff member faced personal risk and needed protection from harassment, regular customers felt their trust in the taproom had been tested, the distributor carried its own exposure across four states, and the other taproom managers had to answer questions from their own customers that evening. A statement that served the public audience but left the team unsupported would damage the business from the inside.
- Verify facts Confirm the sequence from point-of-sale and staff records.
- Brief stakeholders Tell staff, taproom managers, and the distributor before the public statement.
- Publish one statement Acknowledge the reply's tone, state confirmed facts, and name the next step.
- Track recovery signals Watch comment sentiment, taproom visits, and distributor orders for three weeks.
Actions taken
- Pull the point-of-sale record and the full unedited video, and write a timeline with only confirmed times before anyone drafts a public sentence.
- Hold a call with the staff member and the taproom manager before any statement, offering support and a clear account of what the brewery will and will not say.
- Post a single reply that acknowledges the tone of the earlier response, states the confirmed sequence, and commits to a review of the refusal policy without naming personnel.
- Stop replying to individual comments after the statement, and direct questions to one named contact so that new details come from one source.
Results
The statement went out on the second morning, and the earlier reply was replaced with a corrected one that acknowledged its tone. The footage released with the landlord's consent showed a shorter exchange than the original clip suggested, which reduced the intensity of the thread without ending it. The staff member returned to shift the following week, with a manager present for the first two days. Comment sentiment improved over three weeks, and taproom visits recovered in two of three locations. The distributor accepted the statement and asked for the revised policy in writing, which the team delivered by the end of the month.
Questions for discussion
- At what point does a brand's own reply become part of the story, and how would you detect that early?
- Should a company name a staff member when defending a decision, even when policy protects that person?
- Which recovery signal would you trust most three weeks after a public escalation, and why?
Case Study 5
Auditing a neglected article library for refresh priorities
Kestrel Lane Credit Union, a member-owned credit union with four branches and a public content library.
The situation
Kestrel Lane Credit Union maintained a public content library of roughly 1,240 articles, guides, and short explainers covering loans, savings, and budgeting. Traffic was concentrated in about one tenth of the pages, while a large share of the rest had not been updated in several seasons. Approvals ran through three reviewers: a product manager, a compliance officer, and a branch leader. Each review took about two weeks, and some pieces had waited in the queue far longer. Writers were producing new articles quickly, partly because new work was easier to approve than revisions, and the search team had begun to notice overlapping pages answering the same questions.
The challenge
The team had to decide which pages to refresh, which to merge, and which to retire, using evidence that was uneven. Search data existed for most pages but not for search intent, which had to be inferred from query wording. Formats ranged from long guides to single-paragraph explainers, and quality ratings had been applied inconsistently by different editors. Any change touching loan or savings content also needed compliance sign-off, which made the approval queue the real bottleneck.
Analysis
The business objective was to make the library produce member sign-ups and fewer repeat service calls, not simply to publish more. That shifted the audit toward pages that answered a real question and led members to a next step. Search performance gave only a partial picture. Pages with high impressions and low click-through often matched a query the team had not meant to target, while pages with strong clicks and weak time on page suggested the answer was present but hard to find. Teams that work this way read the query wording before deciding what a page should do, since a question about refinancing calls for a different format than a question about definitions.
The operational constraints were the approval queue and the inconsistent quality ratings. A refresh of a loan page needed the same three reviews as a new article, so a team that asked for ten refreshes at once would stall the whole queue. Branch leaders depended on the library when members walked in, and they heard first about complaints over outdated rate language. Writers had their own stake. A backlog of revisions felt like a verdict on their work, which made retirement decisions harder to accept. A fair audit had to set clear criteria, so that retiring a page read as a decision about the content rather than about the person who wrote it.
Actions taken
- Score every article on query intent clarity and traffic over the past three months, using a single rubric that two editors check on a shared sample before scoring the full library.
- Bundle refresh requests into one monthly batch of no more than twelve pages per product area, so compliance reviews one coherent set of changes rather than a stream.
- Merge overlapping loan pages into a single guide with a clear canonical version, and redirect the retired URLs so existing links and search equity carry over.
- Pause new article drafts on topics already covered by a page scoring in the top quartile, and require a short intent note before any new brief is approved.
Results
After the first quarter of the new process, the team had scored the full library, set aside 310 pages for retirement or merger, and built a refresh list of about 180 pages for the next batch. Merged loan guides reduced the number of near-duplicate pages on mortgage questions, and the consolidated pages drew more clicks than the separate ones had combined. Approval times dropped for batched refreshes, though new articles still waited longer than the team wanted. Branch leaders reported fewer questions about outdated rate language. Writers accepted the retirement list more readily once the criteria were published and tied to the rubric rather than to individual authors.
Questions for discussion
- Which signal should outweigh the others when a page has strong traffic but unclear search intent?
- How would you explain a retirement decision to the writer who created the page without making it personal?
- What review capacity does a monthly batch assume, and what happens when that capacity is missing?
Learning Resources
Marketing CASE STUDIES
China, India, United States, Indonesia, Pakistan, Nigeria, Brazil, Bangladesh, Russia, Japan, Mexico, Philippines, Vietnam, Egypt, Turkey, Germany, United Kingdom, France, Italy, Canada, South Korea, Australia, Spain.