The Go Academy — Data Breakdown

Revenue = gross course fees paid (SGD), attributed by registration (Created) date; four streams (Semester / Holiday / Adult / Camp), with the 10-day holiday intensive (10INT) counted as Holiday. Recomputed from source enrolment records. 2026 is part-year (through Jun, plus advance July signups).

From the “How did you learn of TGA?” field — all CRM contacts, regardless of whether they enrolled (87% of 5,697 filled, n=4,999). Chinese social = WeChat + Xiaohongshu + 狮城辣妈 forum; Friends/Referral = word-of-mouth.

Source mix by CRM sign-up year, semester-enrolled students only; bracketed on the axis = students with a source that year. 2022→2026: Instagram rose 1%→9%; Facebook Ads rose 11%→16%; Friends/Referral fell 32%→26%; Chinese social fell 14%→4%; Online Search ~30% throughout. 2022 inflated by migration backfill; 2026 part-year.

Three stages per channel: Contacts (CRM records with a source) → Did a trial (camp/holiday) → Enrolled in a semester class.

• Labels show the two step rates: →trial = trial ÷ contacts, then →sem = enrolled ÷ trial. Hover a bar for both, plus the overall enrolled ÷ contacts.
• Source: all three stages come from each student’s own Student Database record (channel, trial flag, semester enrolment) — one standardized source per student. Only the channel attribution (“how did you hear of us”) is self-reported.
• →sem can exceed 100% (e.g. Other 2023): “enrolled” also counts students who went straight to a semester class with no trial, so on small channels enrollees can outnumber trial-doers. The Sankeys above separate the trial vs direct paths cleanly.
• Use the buttons above to toggle year.

Adult retention is measured in quarters (8-lesson terms, ~Jan/Apr/Jul/Oct); each step is +1 quarter. ⚠ Note the semi-annual return pattern (+2q, not +1q) — flagged for ops review.

⚠ Punggol 2026 (76) includes 38 off-cycle January camp registrations — Punggol’s opening-period 3-Day Go Camp (13–18 Jan 2026), folded into this Q1 window. 0 of the 38 convert within 150 days (only 1 has enrolled in any semester product at all), but this batch is not really measurable: a January camp’s next natural intake is the July semester (~175 days out, just past the 150-day window), and July-2026 enrolment is still being entered. Treat as WIP.

Lifetime semester revenue and tenure by age at first enrolment (kids only; maturity-adjusted to those who started 2022–2024). Age 5–6: ~$2,100 over ~5.2 semesters; age 9–11: ~$1,500 over ~4.4 semesters. Bars = avg lifetime $ (left axis); line = avg semesters stayed (right).

New (first-ever) kid students per year, per centre. Each bar stacks the year’s two intakes: solid = January (H1), paler = July (H2). 2022 shows July only (its January was pre-Airtable migration, excluded); 2026 shows January only (July semester not in yet). Online excluded.

New (first-ever) adult students per year, per centre. Each bar stacks the year’s two intakes: solid = January (H1), paler = July (H2). Real course enrolments only — the Introductory Go Workshop (adult trial) is excluded, same as kid camps are. 2022 = Jul only, 2026 = Jan only. Online excluded.

Catchment map — each dot is an enrolled student's home (geocoded, colour = centre they attend), clustered into counts as you zoom out. Dashed rings mark 2 km (inner) and 5 km (outer) around each centre. Toggle centres top-right; ⍌ top-left for full screen; scroll to zoom while hovering the map. Homes are jittered ~50 m for privacy; PC ring is ~±1 km (new-mall postal). Needs internet to load the base map.

How far enrolled students travel to each centre — home postal → centre, straight-line km (all-time semester students). Medians: MC 4.0, KC 4.3, PC 3.8 km. Geocoded from OneMap building coordinates; 94% of enrolled students matched (Online/LingoAce students — who have no physical catchment — and ~6% with an unmatched/malformed postal are excluded). 20 recently-built addresses missing from the geocode set were added at neighbourhood-level (approximate). PC ~±1 km — Punggol Coast Mall's postal is too new for the geocode dataset, so PC is anchored on the nearest mapped Punggol Coast point.

Home postal district (Singapore postal sector → district) of enrolled semester students, by centre attended; sorted by total demand. A student attending two centres is counted under each (so columns match the travel chart). Based on postal sector, so it covers a few more students than the exact-geocoded map above.