Grow bookings through seasonal message-fit, not ad volume
LUCENA Inc. / 2026.09.20
01Background & Challenge
Tripling summer ad spend did not grow bookings; the cause is not market conditions but “seasonality × presence of a hook.”
Month
Ad spend
Bookings
CPO
US+AU arrivals
Mar (Spring)
¥3,400,000
104
¥32,000
473,000
Jun
¥11,900,000
34
¥350,000
418,000
Jul
¥11,900,000
52
¥229,000
336,000
Aug
¥8,800,000
51
¥173,000
236,000
Key points: Even at ~3× spend, bookings stayed flat; CPO worsened ¥32k→¥350k. Bookings vs arrivals are uncorrelated (r≈0); the US/AU summer dip is annual seasonality.
1
01Background & Challenge
Every month that grew had a “hook” ready — without exception.
Growth month
Bookings
CPO
Hook in place
Nov
107
¥58,000
Black Friday (social + LP + discount)
Mar
104
¥32,000
Spring / cherry-blossom content & tours
Dec
91
¥55,000
Holiday / year-end
Jun-Aug
34–52
¥173,000–350,000
(spend up while autumn content was not ready)
Months with a hook (seasonal content/campaign) also had the lowest CPO. = Bookings grow via message-fit, not ad volume.
2
02Current Analysis & Past Seasonal Trends
The segment we win flips completely by season.
Past period
Solo
Couple
Family/Group (3+)
Ad spend
Nov Black Friday
58%
37%
6%
¥6,200,000
Dec Winter / Holiday
61%
35%
4%
¥5,000,000
Mar Spring / Blossom
24%
43%
32%
¥3,400,000
Reading: Black Friday & winter are a “solo/couple engine” (≤2 people = 95-96%). Only spring (blossom) wins family/group (32% of bookings, top revenue, lowest CPO).
※ “Family/group (3+) booking share” = share of bookings with 3+ participants.
3
02Current Analysis & Past Seasonal Trends
The family/group uplift is driven by UI changes too, not only seasonality.
Family/group (3+) booking share — monthly trend
Old-UI period (Nov-Dec) 4-6% → settled at 24-28% after the 4/10 & 7/20 booking-flow changes (gray = old UI / red = spring onward). Past data alone cannot separate season from UI.
4
02Current Analysis & Past Seasonal Trends
The larger the party, the earlier they book. So delivery timing and messaging must change accordingly.
Average lead time by party size (booking → tour date)
In-trip = solo (instant) → deliver last-minute to in-Japan travelers / “book even at the last minute.” Pre-trip = family/group, US/AU (planners) → deliver early to home countries / early-booking & fit messaging.
Lead time = days from booking to tour date (median). Larger parties tend to plan earlier.
5
02Current Analysis & Past Seasonal Trends
Segments differ by country. AU = the family/pre-trip core, but we are losing it recently.
Country (IP geo)
Total bookings
Solo
Couple
3+ (family/group)
US
288
47%
35%
18%
AU
68
35%
37%
28%
Japan
152
57%
34%
9%
(unknown)
172
34%
43%
23%
Last 90 days: AU 3+ (family/group) = 0 (vanished) / US = 18 (33%) now drives 3+. = Recovering the lost AU family/pre-trip segment is the spring goal.
Lead: AU ~40 days, US ~20 days, Japan ~2 days (March sample) → the earlier the plan, the earlier & home-country the delivery.
Amplitude Complete Booking (undercounts vs Admin booking SSOT; country = IP geo; unknown ~25%), last 12 months. Treat as directional share.
6
03Validation & PDCA Objectives
Since past data alone cannot separate the drivers, we run each seasonal campaign as a “test” and validate it via PDCA.
P Plan
Prepare content, campaign and journey for each seasonal pin
D Do
Don’t scale spend first; shift budget once the hook is ready
C Check
Track bookings, CVR, segment mix and CPO via weekly flash + monthly
Run the three seasonal pins (Black Friday, Winter, Spring) in sync across every team.
Pin
Timing
Target segment
Budget (as prior year)
Black Friday
Late Nov
Solo/couple (+ family early-bird test)
¥6,200,000
Winter / Holiday
Dec
Solo/couple (+ family early-bird test)
¥5,000,000
Spring / Blossom
Feb-Mar
Family/group (primary)
¥3,400,000
Ads, creative, web and product all work to one deadline back-calculated from tour dates. Whether original tours (geisha, sakebaro, etc.) boost acquisition year-round is validated in a separate structure.
8
04Seasonal-Pin Plan
Black FridayWin the solo/couple last-minute.
Aim
Maximize in-Japan solo/couple + test early-bird for family future spring trips
Target
Main = in-Japan travelers (solo/couple) / Sub = family in home countries
Delivery design
Last-minute to in-Japan geo (Meta / Google Search / retargeting) + family early-bird to home countries
Budget
¥3,000,000
Same period last year: 71 bookings, CPO ¥58k, ≤2 people 95% (= grow the segment we already win)
9
04Seasonal-Pin Plan
Winter / HolidaySustain and strengthen solo/couple.
Aim
Winter-experience messaging for solo/couple / Sub = year-end family early-bird test
Target
Main = in-Japan travelers / Sub = family in home countries
Delivery design
In-Japan last-minute focus (Dec delivery)
Budget
¥5,000,000 (same as last Dec)
Same period last year: 54 bookings, CPO ¥55k, ≤2 people 96%
10
04Seasonal-Pin Plan
Spring / Blossom (primary)Win family/group via staged delivery ordered by lead time.
Aim: lift family/group (3+) 32%→40% (primary = US + AU; especially recover AU family/pre-trip, now at 0)
Segment (lead time)
Delivery timing
Delivered to
AU family (~40 days out)
Early Feb
Home country (Australia)
US (~20 days out, largest volume)
Early Mar
Home country (US)
Family/group (16-20 days out)
Early Mar
Home country × family
Couple/solo (10-5 days out)
Mid-late Mar
In-Japan + home country
Last March: 74 bookings, CPO ¥32k (lowest), family/group 32% (won family on the smallest budget) / budget ¥3.4M
11
05Objective & Goals
The objective is to grow bookings via conversion rate. The goals: net booking growth, CPO improvement, family acquisition.
Sustain solo/couple; newly acquire family/group and pre-trip travelers to improve the P&L.
Bookings (per pin)
+20-35%
vs. prior year
CPO
¥30-60k
message-fit level
CVR (visit→booking)
0.4%+
up from summer 0.1-0.2%
Spring family/group
32→40%
capture pre-trip
Family early-bird
>6%
early-bird test
※ Targets are directional placeholders; refine via PDCA after launch (details finalized internally).
12
06Roadmap / Gantt Chart
Back-calculate from tour dates: prepare → deliver. Layer seasonal pins on the year-round base.