TOMOGO! Booking-Data Analysis: Findings & Test Plan

Sources: all Admin bookings, non-CXL N=676 (sales/booking SSOT) / Amplitude proj643614 (access, funnel, country=IP-geo) | Built 2026-09-14

About this analysis

Scope: all 676 non-CXL Admin bookings (sales/booking SSOT) + Amplitude (access/funnel). Cuts: party-size segment / booking timing (lead) / value & repeat / tour type / country. Goal: understand who / when / how much / what / from where people book, to find upside and a test plan.

Conclusion: what we found

82% of bookings are ≤2 people; solo is the largest and highest in per-person value and repeat.
The growth engine is solo travelers' in-trip same-day add-ons (2.1× lifetime value for repeaters; 64% within 3 days).
The more people, the earlier they book; solo decides in-trip.
Consultation tours are ‘neither/nor’ with low conversion; only flexible solo self-select.
US = pre-trip groups, high value; Japan = in-trip impulse + guide-recruitment noise.

The through-line: a key cause of low conversion / missed bookings is ambiguity about who and which situation a tour is for. Clarifying by situation (filtered search / feature pages / how it's shown) lifts conversion, recovers missed demand, and simultaneously measures party-size / purpose intent.

Today: ① who is being missed (groups / ambiguous situations) ② which test to run first

Bookings ≤2ppl

82%

Repeater LTV

×2.1

Add-on in-trip

64%

Solo per-person

¥12,854

top

Consult CVR

0.27%

low

A. Who books

Bookings by segment

Explanation

82% are ≤2 people (solo 43% + couple 39%); family+group(3+) only 18%.
Solo is highest in per-person value and repeat (below).
Even ‘group-looking’ consultation tours are 52% solo.

In short：TOMOGO! today has a booking structure optimized for small parties (especially solo).

Hypotheses & problems

Family/group bookings are under 20% of the total.
Family/group demand may exist but is being missed.
Even when families/groups view a tour's detail page, they may lack the information that makes deciding/booking easy.

Proposed test

Build filtered search and feature pages for each situation

PurposeGuide ‘solo / couple / family / group’ visitors to tours that fit, to win family/group bookings we may have been missing — and, at the same time, capture ‘how many people and what they want’ while they browse.

What to look forHow much families/groups use the filters and go on to book, and whether their bookings actually rise — answering ‘missed demand’ vs ‘naturally few’.

B. Value × repeat (the growth source)

AOV / per-person / repeat by segment

Segment	AOV	Per-person	Repeat
Solo	¥12,854	¥12,854	32%
Couple	¥22,430	¥11,215	21%
Family	¥26,038	¥7,400	25%
Group3+	¥34,614	¥9,161	21%

Repeater lifetime value ¥44,291 = 2.1× single ¥20,749 / 64% of add-ons are in-trip same-trip (return-trip 5%)

Explanation

Totals rise with size, but per-person is highest for solo ¥12,854 (family lowest ¥7,400) = ‘group=high value’ is just more people.
Repeat is what grows revenue: repeaters spend ~2× single-timers; repeat rate highest for solo 32%.
That repeat is adding tours in-trip within a few days (64% within 3 days); very few return later (5%).

In short：The revenue source = solo travelers adding one more tour during their stay — today it just happens naturally; we don't drive it.

Hypotheses & problems

Solo repeats easily because a single person can decide instantly, even last-minute or mid-trip.
Families/groups need to coordinate several people, so on-the-spot add-ons are harder = there's friction (hypothesis) — likely why they repeat less.
So growing family/group repeat needs product/UX that makes deciding easier. Solo is valuable too, but situations differ, so one-size-fits-all messaging causes today's skew.
→ The key is splitting how tours are shown by situation (search/UX) — this connects to D.

Proposed test

Split how repeat/add-ons are prompted, by situation

PurposeRecommend ‘one more during your stay’ to solo, and give families/groups an ‘easy to decide on the spot’ presentation (easy-to-coordinate options, same-day availability). Don't use one presentation — split by situation (→ connects to D).

What to look forWhether solo's ‘second tour’ rate rises, and whether family/group repeat bookings increase.

C. Booking timing × party size

Tightly linked to B: families/groups face repeat friction, yet they plan before the trip — so the key to their revenue is capturing pre-trip bookings.

Median lead by party size

* Lead time = days from booking date to tour date (not sign-up-to-booking).

Explanation

Lead grows with size: solo 5d → couple 10 → family 16 → group 20d; pre-trip(≥8d) share 44%→69%.
= groups/families plan pre-trip; solo decides in-trip.
Solo books even at the highest per-person price; no discount needed, and it's profitable.
Families/groups have a lower per-person price = easy to show ‘cheaper per person’, and the head-count captures total revenue & gross profit.

In short：Pre-trip (planners=family/group) and in-trip (impulse=solo) need different levers. Don't discount solo; win family/group via ‘cheaper per person’ × head-count for revenue & margin.

Hypotheses & problems

Early-booking perks likely work for family/group who move early, but not for last-minute solo.
A blanket discount would cut prices even for solo who'd book anyway — just eroding profit.
‘In-trip costs less to acquire’ is appealing but an unverified hypothesis (the per-person price gap is real; the cost side isn't checked).

Proposed test

Offer ‘reasons to book early’ to families/groups who plan before the trip

PurposeShow families/groups the ‘low per-person price’ to feel like a deal, and capture total revenue & gross profit from the larger head-count. Perks needn't be discounts — a preferred time slot or a free add-on works. Don't offer them to solo/last-minute (high value, already profitable).

What to look forWhether family/group bookings rise with the perk versus without, and whether the increase pays back (via head-count revenue & margin).

D. Tour clarity (consultation conversion)

What ‘consultation’ tours are：tours whose itinerary isn't fixed — we build a plan from the customer's requests (e.g., Tokyo Omakase Tour, Custom Private Tour).

Conversion by type

Who books consultation

Explanation

Clear Walking converts at 0.72%; consultation (Omakase/Custom) is 0.27%, the lowest — many views, few bookings.
52% of consultation bookers are solo; family+group only 21% — it isn't working as a ‘group’ product.
Reading: consultation is ambiguous about who it's for, so only flexible solo self-select while families/groups bounce.

In short：‘Target clarity’ likely drives conversion (clearer = higher).

Hypotheses & problems

Viewed but not booked = a sign of demand we're failing to capture.
Whether the barrier is ambiguity, price, or the hassle of needing consultation is not yet separated.
If ambiguity is the cause, just changing how it's shown could lift bookings (low-cost). The ‘situation-based presentation’ from B applies directly.

Proposed test

Show consultation tours split into ‘for solo / for family / for group’ and test it

PurposeClearly stating who each tour is for helps families/groups feel it's ‘for us’ and book — while separating the real barrier (ambiguity vs price vs hassle).

What to look forWhether family/group bookings rise when the target is stated. If they do, ambiguity was the cause; if not, price or hassle is.

E. Geography (country)

By country: conversion / avg value / lead / repeat

Country	Conv.	Avg	Lead	Repeat
US	3.33%	¥27,130	42日/d	28%
Japan	0.86%	¥20,633	2.4日/d	18%
Australia	2.52%	¥23,147	19日/d	14%

Explanation

US = conversion 3.33%, avg ¥27,130, lead 42d, repeat 28% = pre-trip, group-leaning, high value; multi-city (Kyoto19+Tokyo18+Osaka9).
Japan = most views yet conversion 0.86% (lowest), lead 2.4d. But it's mainly guide-recruitment-ad resident traffic — ‘Japan = lost opportunity’ may be a misread.
AU = Tokyo 43% × nature/hidden-gems (the ‘young bar-hopping’ assumption doesn't match).

In short：US is by far the highest-quality demand; Japan's ‘many but thin’ is explained by noise.

Hypotheses & problems

US books well in advance (42 days) — reachable early in planning.
Japan's high traffic likely mixes in residents who came via guide-recruitment ads, distorting the real traveler picture.
Assumptions like ‘AU = young bar-hopping’ don't match the data — revisit creative assumptions with numbers.

Proposed test

Prioritize the US, and separate Japan's guide-recruitment traffic when reading the numbers

PurposeGive high-value US customers early-planning messaging and a Kyoto+Tokyo+Osaka multi-city path. For Japan, count guide-recruitment arrivals separately to see the true traveler conversion.

What to look forWhether US bookings rise, and what the true conversion of Japanese travelers is once guide-recruitment traffic is removed.

Facts vs hypotheses：Facts = per-person gap is a party-size effect / repeat gives 2.1× lifetime value & in-trip add-on 64% / groups=pre-trip / consultation skews solo. Hypotheses = whether consultation ‘ambiguity’ mainly causes low conversion; size of missed group demand (browse intent unmeasured); family/group add-on friction; ‘in-trip is cheaper’. Lead time = booking→tour date. Booking source of truth = Admin; Amplitude & country (estimated view location) are directional, ~half captured.