I have been writing about institutional marketing strategy for long enough that I have learned to distrust framework pieces that never touch a real campaign. Frameworks compress experience into shape, but the shape is not the experience. Anyone who has actually run an enrollment cycle knows that a plan survives contact with the market for about three weeks before the surprises start arriving.
What follows is one season. A single K-12 international school in a tourist region of the Caribbean. We will not name it — anonymity is a condition of the work — but we will describe it specifically enough that the texture is real. The school is real, the numbers are real, and the surprises are real. The framework laid out in the first three articles of this series held up. It also got bent in specific places that I want to walk through honestly, because the bending is where the learning happened.
I am writing some of this in the first person because I was in the room for most of the decisions. The rest of the team — the head of school, the admissions director, the small marketing pod, the two faculty members who quietly became our best on-camera assets — deserves more credit than this piece can give them. The framework is mine to defend. The execution was theirs.
1. The school, the market, the goal
The institution is a K-12 international school in a coastal region of the Dominican Republic that has been changing rapidly for a decade. Tourism brought money, money brought construction, construction brought families — including a large and growing population of European and North American extranjeros who settled in the area and now needed schooling that would not strand their children if they moved again in five years. The local Dominican professional class, simultaneously, had become wealthier and more demanding about what they expected from a private school. The market was no longer captive.
The specifics: roughly 800 students, 31 nationalities represented, full K through 12, Cognia accreditation, integrated arts and robotics in the regular curriculum (not as electives), graduates accepted into top regional universities at rates well above the area average. The school was real and the differentiators were real. They were also nearly invisible from the outside — a familiar problem.
The competitive landscape had three pieces. First, an entrenched local competitor with a longer history in the area, weaker pedagogy in the eyes of those who knew both schools, but a parent network that had been compounding for thirty years. Second, a handful of smaller bilingual schools with strong individual programs but no real positioning. Third, and most importantly: a new sucursal of a well-known international school brand was opening in the same catchment area for the following year. The market was about to get more crowded, not less.
The goal we wrote down in January was unambiguous: 100 new enrollments for the 2026-2027 school year. Not 100 leads. Not 100 inquiries. One hundred families across the K-12 grade spread, fee-paying, signed, deposit received. The board wanted the number defended on a single page with the funnel math behind it. We wrote that page in the first week of January.
2. Phase 1 — January: the audit and the empty page
The first month was almost entirely diagnostic. No public-facing content. Nothing on the feed except the holiday post that was already scheduled. The temptation to "just start posting" was intense, especially from board members who had read somewhere that velocity was the new strategy. We held the line.
The differentiators audit produced a surprise in the first week. The leadership team's intuition about what made the school distinct turned out to be only partially correct. The phrase "31 nationalities" had appeared in the school's printed prospectus for years and nobody on the leadership team thought of it as remarkable — it had become wallpaper. When we ran the claim past three prospective parents in informal conversations, all three reacted with visible surprise. That detail, internally taken for granted, was externally a headline.
The opposite happened with "academic excellence," which leadership had defaulted to as the primary positioning. The audit dismantled it within an afternoon. Every competitor said the same thing. Parents had stopped hearing it. We struck it from the messaging architecture and replaced it with three claims that could survive scrutiny: the 31 nationalities, the integrated arts-and-robotics curriculum, and the regional university placement rate. We added a fourth — graduates who transferred internationally without losing a grade level — because the extranjero segment needed to hear it explicitly.
Segment work followed. Three segments accounted for the realistic recruitment pool: established Dominican professionals, international families resident in the area, and families in transition (typically arriving mid-cycle for work-related reasons). For each, we wrote out the motivation, the concern, and the unspoken question. The unspoken questions turned out to be different from what leadership had assumed for two of the three segments. The audit was earning its keep.
The funnel math fit on one page. 100 enrollments would require approximately 400 qualified leads at a 25% lead-to-enrollment conversion rate, which would require roughly 2,000 engaged prospects at a 1-in-5 ratio, which would require reach across 80,000 to 100,000 unique households in the catchment. Each number was a forecast, not a guarantee, but it gave us a single page to defend at the board meeting.
And then the campaign nearly died before it launched. Two board members read the strategy memo and pushed back hard. The proposal felt aggressive — too direct, too willing to spend on testing, too willing to walk away from tactics that had felt safe for years. There was a real meeting where the decision to proceed was 4-3. I remember it because I had to defend the funnel math in person and the experience was clarifying. If you cannot defend the numbers when a board is uncertain, the numbers were not honest to begin with.
We proceeded. The empty page at the end of January was not empty by accident. It was a refusal to ship anything before the strategy could hold its own weight.
3. Phase 2 — February: launch, and the first surprise
February was the awareness push. The 60/30/10 architecture went live: 60% aspirational content (student stories, classroom texture, faculty in action, the quiet observable reality of a multilingual hallway), 30% direct content (accreditation, outcomes, curriculum specifics, university placement data), 10% conversion content (open house registration, virtual tour signups). The goal was simple — become familiar to every household in the catchment within thirty days.
The team shipped on a weekly cadence. Two pieces of aspirational content per week, one piece of direct content, one conversion call-to-action every ten days. Production was lean: a single videographer on retainer, faculty and students used as on-camera talent with proper consent, AI used heavily in the back-office work the second article in this series described (transcription, multilingual adaptation, repurposing across formats, first drafts for the marketing pod to edit).
The first surprise hit in week three.
Aspirational content was outperforming direct content on every engagement metric by a factor of four. Reach, completion rate, saves, comments — the aspirational pieces were doing the work the playbook predicted they would do. But the lead generation number was nearly zero. The conversion-content pieces, which we expected to do the lead capture, were being scrolled past. Twelve qualified leads in the first month. Against a target of 400 over the cycle, that pace would not get us there.
The instinct in the room — a strong one, voiced by more than one person — was to pivot. Lean harder into conversion content. Push the 10% to 25%. Run more aggressive calls to action. The argument was reasonable: we have a number to hit, the conversion content is what produces leads, so make more of it.
I argued against the pivot, and the head of school backed me. The reasoning was this: the aspirational content was building familiarity, and familiarity is what makes conversion content work later. Cold conversion campaigns burn audiences. The funnel had not failed; it had not yet had time to compound. We were watching the front of the funnel fill while the middle and back were still empty. That is what the first month is supposed to look like.
We held the mix. February closed with 12 qualified leads, 3,400 newly engaged accounts, and a leadership team that was nervous but not panicking. In retrospect, the decision to hold the line was the single most important strategic call of the entire cycle. The temptation to pivot to conversion under early pressure is, I think, the most common failure mode I have watched in K-12 campaigns. The funnel logic earns the patience or it does not.
4. Phase 3 — March through May: engagement, and the funnel starts behaving
The middle months were where the framework began to feel like it was running on its own.
Open houses launched in mid-March. The first was attended by 28 families — a low number, anxiety-inducing in the moment, but the open houses were being treated as conversion events at the wrong stage of the funnel. We restructured. The second open house was reframed as a curriculum showcase, with faculty demonstrating the integrated arts-and-robotics work to families who were still in the considering phase. Attendance for the second event was 67 families. The third was 94. The format shift was the unlock: parents who were not yet ready to apply were ready to visit, and the visit moved them down the funnel faster than any digital touch could.
The parent-fit tool went live in early April. We had built it through March: a brief web experience that asked a family seven questions about their educational priorities, their child's strengths and interests, and their concerns about transition. The output was a one-page report mapping the family's priorities to the school's actual pedagogical approach, with honest indications of where the fit was strong and where it might not be. We were careful about that last part — the tool sometimes told families the school was not the right fit. That honesty did more for the brand than any aspirational video.
The tool became the highest-converting asset in the entire campaign. Forty-three percent of families who completed it went on to schedule a campus visit. Of those, sixty-one percent applied. It was simultaneously a marketing instrument, a lead generator, and — for the admissions team — a pre-qualification mechanism that made first conversations dramatically more productive.
Lead velocity climbed week over week. Twelve qualified leads in February. Twenty-three in March. Thirty-four in April. Forty-seven by the last week of May. The funnel was behaving. The direct content from earlier months was now landing on warmed audiences. Faculty introduction videos, which had performed modestly in February, were now being saved and shared. A clip of the robotics teacher explaining a project to a parent at an open house was repurposed into a thirty-second social piece that ran on its own legs for six weeks.
Two faculty members became unintentional stars. One — a primary teacher with a quiet, undramatic on-camera presence — appeared in three videos by April and was being recognized in supermarkets by the end of May. The other was the robotics lead, whose enthusiasm translated unedited. We learned, mid-cycle, that faculty introduction videos were the single highest-trust content the school could produce. Parents trusted teachers more than they trusted marketing. We doubled the cadence in May.
The middle of the funnel was no longer empty. It had taken three months of patient front-of-funnel work to fill it, but it was now compounding visibly. Some weeks in May, the lead generation pace was higher than what the funnel math had projected. We let ourselves get cautiously optimistic.
5. Phase 4 — June through mid-July: intensification, and the second surprise
June and the first half of July were the application crunch. The 60/30/10 mix shifted. We pulled conversion content up to roughly 25% of output, kept direct content steady at 30%, and let aspirational content recede to 45%. Application deadline reminders became regular. Scholarship windows were communicated explicitly. Retargeting started on warm leads — families who had attended an open house, completed the parent-fit tool, or downloaded the application packet but had not yet submitted.
This is where the second surprise arrived, and it was larger than the first.
The retargeting budget had been split between paid Meta retargeting and WhatsApp outreach by the admissions team. WhatsApp — sector-appropriate for the Dominican market in a way that is hard to overstate to readers outside the region — was running on a personal, conversational cadence. An admissions counselor would send a message to a parent who had visited two weeks earlier, ask if there were any remaining questions, and offer to schedule a call. We had treated WhatsApp as supplemental. The paid retargeting was the "real" channel.
We were wrong. WhatsApp outreach was converting at roughly three times the rate of paid retargeting on warm leads. The mechanism was obvious in retrospect: a personal message from a named human being is not the same as a programmatic ad. In a market where personal trust is the substrate of nearly every meaningful transaction, WhatsApp was the channel that matched the medium of decision. The paid retargeting was reminding people we existed. WhatsApp was finishing the sale.
The tactical pivot was immediate. We pulled budget from paid retargeting and reallocated it to expanding the admissions team's WhatsApp capacity — two additional part-time counselors, brought on for the final six weeks, with AI-assisted scoring helping them prioritize which families to reach out to first. The AI scoring was straightforward: it ranked warm leads by a composite of recency, engagement depth, parent-fit tool completion, and application progress. The counselors spent their first hour each day on the top of the list. The conversion rate on that prioritized outreach was nearly forty percent.
The parent-fit tool, which had launched as a static asset in April, was being quietly refined through the cycle as we accumulated data on which questions actually predicted family-school fit. By mid-July the tool was on its fourth revision. None of the changes were visible from the outside. They were the kind of small calibrations that you only see if you are reading the back-end completion data weekly.
6. Phase 5 — mid-July through August: the final close
The last six weeks were the close. The funnel was full. The question was how cleanly we could move families from application to enrolled.
The lead-to-applicant conversion rate had been the bottleneck we had not anticipated. Marketing was filling the funnel. Admissions was processing it. But the moment-of-truth conversations between admissions counselors and hesitating families were where the campaign would be won or lost. The marketing team's job for the final stretch was not to generate more leads — it was to make the admissions team's job easier.
We produced a series of short, narrowly targeted videos addressing the specific objections that had surfaced repeatedly: transition concerns for incoming families, scholarship clarity for the established Dominican segment, accreditation specifics for families worried about future international moves. These were not broadcast pieces. They were sent, one to one, by admissions counselors to specific families at specific moments in the conversation. AI helped us produce variants quickly — multilingual, length-adapted, sometimes personalized at the family level. The volume of content rose. The targeting precision rose faster.
Last-minute admits made up a meaningful fraction of the final number. Twelve families enrolled in the final two weeks of August, all of them previously hesitating. Several were the result of single WhatsApp conversations that turned on the right piece of information arriving at the right moment. The admissions counselors deserve the credit. The marketing infrastructure made it possible for them to be that responsive.
On the last working day of August, the admissions director sent me the final number.
7. The final numbers — what worked, what didn't
The campaign closed at 104 enrollments against a target of 100. The other funnel numbers, rounded:
- Total unique reach in catchment: approximately 87,000 households.
- Engaged audience: roughly 1,900 accounts.
- Qualified leads: 423.
- Applications submitted: 178.
- Enrollments: 104.
The implied conversion rates: lead-to-applicant 42%, applicant-to-enrolled 58%, lead-to-enrolled 25% — almost exactly on the funnel math forecast from January. Cost per enrollment came in at a figure roughly fifteen percent below the regional benchmark for comparable private schools, with the caveat that we are not adjusting for the value of unpaid faculty time that went into on-camera work.
What overperformed:
- WhatsApp outreach. Three times the conversion rate of paid retargeting on warm leads. The single largest reallocation of the cycle.
- The parent-fit tool. Forty-three percent of completions led to a campus visit. The highest-leverage asset in the entire campaign.
- Faculty introduction videos. Higher trust than any other content type. Doubled in cadence mid-cycle.
- YouTube long-form school tours. Lower velocity than social, but compounding through search. The video uploaded in March was still generating qualified visits in August.
What underperformed:
- TikTok. Significant investment in the first two months, modest engagement, near-zero attributable leads. The audience was real but not in the decision-making cohort. Worth experimenting with again only if a clearer student-led format can be built with proper safeguarding. As a primary channel, it did not pay back.
- Print campaign. A regional family magazine placement and two billboards in residential areas. Combined cost was meaningful. Combined attributable conversions were three. We will not repeat the print spend at that level next cycle.
- Cold conversion content in February. The "Apply now" pieces shipped in the first month were essentially wasted. They needed the warm audience that arrived in April and May. The pieces that ran later, against a warm audience, converted at expected rates.
| Channel | Reach | Leads | Conversions | |
|---|---|---|---|---|
| YouTube | ||||
| TikTok | underperformed | |||
| Google Search | ||||
| overperformed | ||||
| barely moved | ||||
| Parent-fit tool | overperformed |
8. Three lessons we wouldn't have learned without doing it
Frameworks give you the shape of the answer. Running the framework gives you the texture. These are the three takeaways I would not have learned from reading the playbook.
Lesson one: funnel order matters more than funnel volume. The decision to hold the 60/30/10 mix through February, when conversion was lagging and pressure to pivot was high, was the difference between the campaign hitting its number and missing it. Most campaigns I have seen fail in K-12 fail at this exact moment — they panic in month two, shift to conversion content too early, and burn the audience before the funnel has had time to compound. Cultivating engagement before harvesting it is not a slogan. It is the operational discipline that makes the math work.
Lesson two: decision-window discipline beats year-round shouting. The calendar phases earned the budget. We spent almost nothing in January. We spent heavily in May, June, and July. We tapered in August. The total spend was roughly the same as the previous cycle. The result was a forty-percent increase in new enrollments. The lever was not budget. The lever was timing the budget to when families were actually making decisions. Schools that spend evenly across twelve months are competing against themselves.
Lesson three: admissions experience is half the campaign. The bottleneck was never lead volume. It was lead-to-applicant conversion, which is an admissions number, not a marketing number. The single largest improvement we made in the second half of the cycle was the AI-prioritized outreach by admissions counselors using WhatsApp. That was not a marketing tactic. It was a marketing-and-admissions tactic, run by a team that finally stopped treating itself as two departments. If I had to redesign the cycle from scratch, I would start the marketing-admissions alignment work in November, not in February.
9. What the framework looked like when it met the market
The framework held. The audit produced sharper differentiators than the leadership team had on its own. The segment model predicted behavior accurately. The 60/30/10 architecture worked once we trusted it through the early discomfort. The decision-window calendar earned its budget. The funnel math landed within five percent of forecast.
And the framework bent. Aspirational content overperformed at the front of the funnel by more than expected. Conversion content needed three months of audience warming before it could do its job. WhatsApp turned out to be the highest-converting channel in the entire campaign, displacing paid retargeting in a way no playbook would have predicted in advance. The parent-fit tool — which I had originally classified as a nice-to-have lead magnet — became the highest-leverage asset in the entire campaign and is now, in next year's planning, the centerpiece around which the rest of the architecture is being designed.
What I take from this season is what I have come to take from every season of this kind of work. The framework is leverage. The execution is proof. The judgment is what mediates between them. A team that holds the framework rigidly will miss what the market is trying to tell them. A team that abandons the framework at the first surprise will burn the audience and lose the cycle. The discipline is to know which calls require holding the line and which require bending. That discipline is not in any playbook. It is built, slowly, by people who have run the cycle enough times to recognize the difference. The school I have described in this article ran a season well because the team had that discipline. The framework helped. The team did the work.
And one more thing. The new competitor opened its sucursal in the same catchment area for the 2026-2027 year. We watched their launch carefully. They were aggressive and well-resourced. The campaign we ran did not stop them from opening; nothing was going to. What it did do — what the patient front-of-funnel work in February and March made possible — was establish, in the market's perception, what the school we worked with stood for, before the new entrant had a chance to define it for us. Positioning is a perishable asset. The cycles you run before a competitor arrives are worth more than the cycles you run after. The board members who had voted against the strategy in January told me in September that they understood why we had moved when we did. That conversation was, in its own way, the best outcome of the year.
The four perspectives
One case study is not evidence. It is a single data point that may or may not generalize. What to look for in a piece like this: are the funnel ratios disclosed, are the failures reported alongside the wins, and are the surprises specific enough to be falsifiable? This walkthrough mostly clears those bars. What to dismiss as anecdote: any reader who treats one season at one school as a template to copy. Replication is what turns a story into evidence.
Who was served by this campaign, and who wasn't? The school described here is fee-paying, accredited, and positioned for families with means. The 104 enrolled families are a real outcome. The families in the same catchment who could not afford the school, or who were not reached because the WhatsApp networks they belonged to were different from the ones the admissions team had access to — they are absent from this story. They should not be. A campaign serves the families it converts and shapes the perception of the institution for everyone else. Both matter.
The velocity lesson is the one I keep coming back to. A small team — fewer than ten people across marketing and admissions — shipped weekly for eight months and hit a target the board was uncertain they could even define. They iterated mid-cycle. They killed channels that were not working. They expanded channels that were. The mythology of marketing as a massive-budget exercise is mostly mythology. Small teams that ship and adjust beat big teams that plan and stall.
What makes this story generalizable beyond one school is not the tactics. It is the relationship between framework, execution, and judgment. The framework gave the team a defensible starting position. The execution earned the team the right to learn from the market. The judgment — held the line in February, pivoted on WhatsApp in June, doubled the faculty videos in May — was what turned a plan into a result. A school somewhere else with different differentiators in a different market will not copy this cycle. But the discipline of running their own version of it, honestly, with the patience the funnel requires and the humility the surprises demand, is what shared-prosperity work in education looks like in practice. Numbers serve communities. The numbers in this cycle, in the end, served a community of 104 new families and the institution that will educate their children for the next decade.