Small travel agency in Argeș. Six people, 45 packages in portfolio, prices between 800 and 4,500 lei per person. Mountain tours, weekend retreats, rural tourism in the villages under Iezer. 90% of sales came from word of mouth. The rest came from wherever, with no system.
How their manager ended up on our doorstep.
She wrote us in December 2024. Short message: „I've been reading your posts about POAS. I get the logic. I'd love to talk, but I don't have a big budget." That's how most serious conversations begin, honestly.
On the first call she told us about her previous agency. Four months. Three thousand euros. Reach and engagement campaigns everywhere. At the end she had 12,000 new page likes, about 400 enthusiastic comments, lots of „how beautiful!", and one simple question nobody could answer: „How many actual bookings came from those ads?" None trackable. Not because the ads were badly made. Because nobody had asked them to bring bookings. They'd been asked to bring attention. They had.
When we got into the account we found a ghost town. No Pixel. No Conversion API. Not even a custom event on phone-number click. Their booking page was essentially their Facebook page: people sent DMs, the team answered when they got around to it. But the worst thing was this: of their 45 packages, nobody on the team could say with certainty which were profitable after deducting guides, transport, accommodation, meals, and commissions. They had instinct. They had experience. They didn't have numbers.
What we found in the first two weeks.
Our audit wasn't just on ads. We sat with the ops team, two guides, and the admin. From those conversations plus an Excel with net margins per package, a picture emerged that nobody had seen before.
- 01
Of 45 packages, only 22 had net margin over 20%. Of the rest, 15 were borderline and 8 were straight-up losses once you added returns, cancellations, and complementary packages sold below cost as „customer acquisition". Several of the packages they pushed most were exactly the ones losing them money.
- 02
Conversion infrastructure was basically nonexistent. Someone wanting to book a Rucăr tour had to click the ad, land on their Facebook page, see a 2019 album with 23 photos, then send a DM. Estimated click-to-booking rate: probably under 1%.
- 03
Seasonality was treated as an enemy, not a reality. Summer and winter budgets ran the same as spring and fall. Result: in spring CPA quadrupled because nobody was in the market for a mountain retreat when it rained every weekend.
- 04
Old campaigns optimized for engagement. Meta was learning precisely to bring people who liked and left. Which was normal. That's what you'd asked for. But from there to an actual booking is a long road, and nobody had built that road.
What we built, piece by piece.
Two weeks just on the numbers.
Before anything else, we sat with them and did the net margin exercise for each of the 45 packages. Guide cost (fee plus their accommodation), transport owned or rented, traveler accommodation, meals, insurance, partnership commissions. Result: a list of 22 packages worth advertising dollars and the rest living on organic and returning clients. Nothing was „killed", just pulled from ads.
Clean tracking, from Pixel to offline.
Meta Pixel plus Conversion API on every funnel step: package view, details click, checkout initiation, confirmed booking, payment received. Offline event tracking for in-office cash payments (which they received often, clients who wanted to come meet face to face before paying). First month of data was already gold: we could see exactly where people were dropping off.
Dedicated landings, not a full website.
We didn't build a new site. Too expensive, too long. We built three simple landings for three big categories (mountain, retreat, rural), each with clear social proof (the 5-star Google reviews nobody had been using), real trip photos (not stock), transparent pricing, FAQ, and three-step booking. WhatsApp Business connected for fast reply.
Full funnel on Meta, controlled spend.
TOFU with short Reels of 15-30 seconds, filmed by a guide on his phone, with real destinations and authentic commentary. MOFU with carousels: package plus price plus testimonial. BOFU with deadline offers and retargeting on abandoners. Starting budget 800 euros a month, raised gradually as POAS held up.
Seasonal scaling, not blind.
Winter packages (wellness retreats, snow-themed tours) had higher POAS than summer. Counterintuitive, but clear in the data: premium clients seek calm during holidays, they don't chase discounts. So winter budget went ×1.5. Summer ×2 on classic mountain packages. In spring and fall, drastically reduced budget, fine targeting on existing clients, and email on their list of over 800 past buyers.
Google Ads and email, after month eight.
Once Meta had pulled over 1.1 million lei in the first seven months, we added Google Ads on local and brand keywords, plus an email flow with seasonal reminders and personalized recommendations based on past packages. Retention on ads-acquired clients hit 85% over 12 months, confirming what we'd hoped: the ads were attracting the right people, not bargain tourists.
Their SmartBill. Tax numbers, not what we want to see.
These aren't Meta reports, prettied up and color-coded. These are their invoices, in their system, which anyone who wants to verify them can do on a call with their manager. Full fiscal year 2025, plus the first three months of 2026.
SmartBill, full 2025 (993 invoices, 1,700,657 lei)
SmartBill, Q1 2026 (390,349 lei, 4,436 lei/day average)
The numbers that mattered most.
Data from SmartBill and Meta Ads Manager. Full fiscal year 2025 (12 months), compared to the prior year.
What we missed along the way.
It wasn't Swiss-watch precise. Three moments we come back to and think we'd do differently.
- ✕
We underestimated autumn. In September 2025 we kept the budget at summer level, because it was working. Within two weeks CPA doubled, because the cold audience had stayed the same but purchase intent had naturally dropped. We caught it when we saw two days in a row with zero bookings. Cut the cold campaigns, moved spend to retargeting and email. Had we prepared the seasonal plan two weeks earlier, we wouldn't have lost those two weekends.
- ✕
In month three we got excited and built separate landings for seven sub-categories. Of the seven, only three got meaningful traffic. The other four ate development time and sat empty. A lesson in restraint. Next time: three well-thought-out landings, then A/B test, then add another if it justifies itself.
- ✕
In month four, when we were scaling hard, we didn't discuss operational capacity enough with the team. Result: two weeks with WhatsApp response times between 12 and 18 hours. We lost a few bookings to competitors. Next time, the capacity conversation comes before scaling, not after.
What we took with us into the next projects.
On services with strong seasonality, the budget plan gets built one quarter ahead, not on the fly. Travel made it obvious. With any client running cyclical demand (education, fitness, beauty with specific dates), we now build a 12-month budget calendar before launching the first campaign.
Premium audiences don't react to discounts, they react to calm and to their own time. Winter packages pulled higher POAS at higher cost per booking. That's why we now drop „% off" creative on premium segments and run exclusive offers, short lists, personal messaging instead. Big POAS gap inside 30 days.
WhatsApp Business with under-an-hour reply beats any perfect landing page. On this agency, every time response time dropped below 45 minutes, lead-to-booking rate hit 38%. When it climbed over 4 hours, it dropped to 12%. All new projects now carry a response-time SLA in the contract, not a suggestion.
Email to your existing client base is the cheapest revenue reserve. Their list of 800 past buyers, touched monthly with seasonal reminders, pulled around 110,000 lei in a year, with zero ad cost. On any business with strong retention, email is the first lever we build, before scaling cold traffic.
„First time I saw the SmartBill numbers after six months with you, I called to ask whether this was real. I didn't think you could pull this off for a small travel agency in a small country with small budgets. Looking back, the difference was simple: you wanted to see, at the end of the month, that we made money too, not just you. That shows in everything you built."
Client's words: Travel agency manager in Argeș, January 2026