CausalMonitor
LiveAuto-refresh · weekly
Week 29 · Jul 14–20, 2026 ▾
Member Churn — Chicago Portfolio
33 clubs · 181 reviews scored this week · model retrained nightly · effects gated at 95% before they alert
Poor-review rate
7.5%
▬ 0.0pp wk/wk
Billing leaving-intent
50%
▼ 0.2pp wk/wk
Front-desk effect
+36pp
▼ 0.2pp wk/wk
Clubs on watch
8
▲ 0.1 wk/wk
Reviews scored
181
▼ 0.4 wk/wk
This week in reviewWeek 29 · Jul 14–20, 2026 · 181 reviews scored
W29
Jul 14–20
Two signals cleared the significance bar this week; the rest is noise. Billing leaving-intent rose +3.1pp for a sixth straight week — the “cancel anytime” repositioning and the click-to-cancel rule are now visible in the text, though they have not yet converted to one-star reviews at scale. The urban tier's 1–2★ share widened +1.8pp, concentrated in Avondale, Little Village and Cicero.
The front desk remains the largest cause of an unhappy member (+36pp) but was statistically flat this week; broken-equipment mentions fell 9 with no causal movement — pure noise. Named-employee (5★) praise drifted +0.6pp, the retention lever quietly working. Do this week: get ahead of the billing trend before it reaches one-star reviews, and send an ops review to the three widening urban clubs.
Causal drivers · what actually moves churnranked by effect, not by volume · gated at 95%
Causal driver impact what actually moves churn, held against club · period · reviewer · length
Front desk & staffDISSATISFACTION
+36ppodds 35× · 13% leaving-intent
Safety & theftCAUSAL
+26ppodds 11× · 12% leaving-intent
Billing & cancellationDEFECTION
+25ppodds 15× · 50% leaving-intent
CleanlinessCAUSAL
+25ppodds 13× · 12% leaving-intent
CrowdingMINOR
+7ppodds 3× · 9% leaving-intent
Broken equipmentLOUD · LOW EFFECT
+5ppodds 2× · 11% leaving-intent
Alerts this week significance-gated
Billing leaving-intentREAL MOVE
+3.1pp
6 weeks climbing · tracks click-to-cancel exposure
Urban tier · 1–2★ shareREAL MOVE
+1.8pp
Avondale, Little Village, Cicero widening
Front-desk effectSTABLE
+0.4pp
largest cause, statistically flat this week
Broken-equipment mentionsNOISE
−9
still loudest topic, no causal move
Named-hero rate (5★)WATCH
+0.6pp
the growth lever, early positive drift
Signal trends · 12 weeks illustrative weekly series
Billing leaving-intent
1–2★ share
Front-desk effect ÷10
Billing leaving-intent is the only line breaking its band — a real six-week trend tracking the “cancel anytime” repositioning and the click-to-cancel rule, not yet converted to one-star reviews. Every other signal is flat.
The two engines of churn one problem, two answers
Makes a member angry
Front desk
+36pp
largest cause of a 1–2★ review · odds 35× · cheapest to fix
Makes a member leave
Billing
50%
of billing complaints say they will cancel · 36× the fleet · prices the stock
Counterfactual — cut front-desk & billing failures by 40%
6.6%
projected poor-review rate · $1.4M annual dues protected
Root-cause drilldown the sub-topics inside each driver — where the effect actually comes from
Front desk & staffBilling & cancellationCleanlinessSafety & theftBroken equipmentCrowding
Charged after cancelling38 mentions
+34pp +2.4
Cannot cancel / runaround27 mentions
+31pp +1.8
Surprise or hidden fee22 mentions
+26pp +1.1
Annual fee shock19 mentions
+22pp +0.6
Double-charged card14 mentions
+30pp +0.9
The member evidenceevery effect traces to words a member wrote
The meaning layer hover any highlighted phrase, or any field, to see exactly where the signal was extracted
★
Terrible gym. Hot water is always out and they won’t let me cancel my membership. Worst decision I’ve made in years.
Logan Square · Urban · 2026-03-31 · one raw sentence → nine structured, queryable fields
9 of 28 dimensions read from this review
Overall sentimentStrongly negative
Member lifecycleDeparting
Primary frictionBilling / cancellation
Billing signalBlocked cancellation
Facility infrastructureHot water out
Trust signalBroken
Issue severitySevere
RecommendationWill not return
Actionability ownerBilling / corporate
Voice of the member the evidence behind every driver · highlighted phrase = what became a field
AllBillingFront deskCleanlinessSafety
BILLING★Logan Square · Urban · 2026-03-31
Terrible gym. Hot water is always out and they won’t let me cancel my membership. Worst decision I’ve made in years.
Sentiment: Strongly negBilling: Blocked cancelTrust: BrokenRecommend: Will not return
FRONT DESK★Logan Square · Urban · 2025-07-19
The lady at the front desk is rude!!!!
Sentiment: NegativeStaff: Rude / namedFriction: Front desk
BILLING★Logan Square · Urban · 2026-02-25
Demasiada gente siempre está lleno pésimo lugar para hacer ejercicio estoy pensando seriamente en cancelar mi membresía
Sentiment: NegativeBilling: Cannot cancelRecommend: Warns others
CLEANLINESS★Crystal Lake · Outer Suburban · 2025-01-26
The women's bathroom was dirty.
Sentiment: NegativeCleanliness: Persistently dirty
SAFETY★Avondale · Urban · 2025-09-10
This place is gross and unsafe.
Sentiment: Strongly negSafety: UnsafeSeverity: Severe
CLEANLINESS★Waukegan Lewis · Outer Suburban · 2026-06-02
Always stinks in the locker rooms/ bathrooms
Sentiment: NegativeCleanliness: Restrooms
The retention lever the growth engine, in members' words
A member naming a specific employee is the single strongest cause of a five-star review (+42pp) and it is invisible to any survey. This is the cheapest lever the company owns.
Maria · Front desk · Belmont Cragin
“Remembered my name and my goal on day one. First gym that ever felt welcoming.”
Denise · Member services · Oak Lawn
“Fixed my billing question in five minutes with no runaround. Staying because of her.”
Coach Ray · Fitness trainer · Naperville
“Walked me through the machines without a hint of judgment. Exactly why I joined.”
Across the fleet · what's bad, and where33 clubs · 3 tiers
Club × driver matrix one-star rate by club and failure · redder = worse · the operational punch-list
| Club | Front desk | Billing | Cleanliness | Equipment | Safety | Crowding |
|---|---|---|---|---|---|---|
| Avondale | 33 | 30 | 22 | 46 | 17 | 34 |
| Little Village | 48 | 20 | 36 | 38 | 18 | 30 |
| Schaumburg Barrington | 41 | 32 | 42 | 22 | 25 | 21 |
| N. Broadway | 34 | 48 | 37 | 43 | 30 | 42 |
| Oak Lawn | 20 | 27 | 40 | 42 | 31 | 20 |
| Cicero | 26 | 24 | 27 | 45 | 24 | 47 |
| Back of Yards | 47 | 19 | 30 | 34 | 18 | 31 |
| Logan Square | 48 | 21 | 37 | 21 | 37 | 47 |
| Pullman Park | 24 | 17 | 20 | 35 | 18 | 20 |
| Elmwood Park | 34 | 48 | 31 | 31 | 39 | 20 |
| Washington Square | 37 | 23 | 38 | 50 | 19 | 36 |
| Chatham | 15 | 9 | 11 | 21 | 21 | 9 |
Tier scorecard per 100 one-star reviews
| Club type | Rating | 1–2★ | Staff | Bill | Clean | Equip |
|---|---|---|---|---|---|---|
| Urban | 3.46★ | 34% | 4 | 4 | 5 | 8 |
| Inner Suburban | 4.21★ | 17% | 3 | 2 | 2 | 2 |
| Outer Suburban | 4.84★ | 3% | 1 | 1 | 0 | 0 |
The urban tier trails a full star and runs 10–20× the failure rate of the outer suburbs at the same price.
Club movers this week largest week-over-week rating shifts · illustrative
| Club | Recent | WoW | Status |
|---|---|---|---|
| Avondale | 2.8★ | ▼ 0.09 | deteriorating |
| Little Village | 2.9★ | ▼ 0.14 | deteriorating |
| Schaumburg Barrington | 3.4★ | ▼ 0.11 | deteriorating |
| N. Broadway | 3.4★ | ▼ 0.15 | deteriorating |
| Oak Lawn | 3.4★ | ▼ 0.16 | deteriorating |
| Joliet Voyager | 4.9★ | ▲ 0.06 | recovering |
| Addison | 4.9★ | ▲ 0.05 | recovering |
| Waukegan Lewis | 4.8★ | ▲ 0.16 | recovering |
Club risk grid — all 33 gap between displayed rating and last-18-month rating · darker = worse
Avondale
+1.3
2.8★
Little Village
+1.2
2.9★
Schaumburg Bar…
+0.9
3.4★
N. Broadway
+0.9
3.4★
Oak Lawn
+0.9
3.4★
Cicero
+0.8
3.2★
Back of Yards
+0.8
3.3★
Logan Square
+0.8
3.4★
Pullman Park
+0.8
3.7★
Elmwood Park
+0.8
3.5★
Washington Squ…
+0.8
3.5★
Chatham
+0.6
3.8★
Melrose Park
+0.6
3.7★
Forest Park
+0.5
3.8★
Crystal Lake
+0.5
3.7★
W Foster
+0.5
4.1★
Tinley Park
+0.3
4.1★
South Loop
+0.1
4.2★
Joliet Jeffers…
+0.0
4.9★
Deerfield
+-0.1
4.8★
Lockport
+-0.1
4.9★
New Lenox
+-0.1
4.9★
Aurora Lake St
+-0.1
4.8★
Bolingbrook
+-0.1
4.8★
Plainfield
+-0.1
4.9★
West Dundee
+-0.1
4.9★
Mundelein
+-0.1
4.9★
Montgomery
+-0.1
4.8★
Naperville
+-0.1
4.8★
Elgin
+-0.1
4.9★
Waukegan Lewis
+-0.2
4.8★
Addison
+-0.2
4.9★
Joliet Voyager
+-0.2
4.9★
Rating gap ≥ 0.7★ (act now)0.4–0.7★< 0.4★ (stable)
The operational layer · beyond reviewsillustrative — what this looks like once first-party data is wired in
What becomes possible with first-party data
Reviews are the loudest 6% of the member voice. Your systems hold the other 94%.
Fuse the billing ledger, key-card check-ins, the membership record, app and booking data, and CRM tickets, and the same causal engine moves from “what causes a bad review” to “what causes a lapsed member” — aimed directly at net member joins, the number the market prices. Every metric below is illustrative.
Member dues run-rate
$5.03B
▲ +0.4%
annualized, system-wide
Net member adds · wk
+3,410
▼ −36%
−36% vs a year ago
Trailing 90-day churn
4.7%
▼ +0.3pp
of members, monthly
Avg revenue / member
$19.51
▬ flat
monthly dues
Check-in frequency
1.8 / wk
▼ −0.1
median active member
Black Card mix
62%
▲ +0.5pp
of members
Causal drivers of 90-day churn estimated on structured operational data · illustrative
Billing failure event (charge-after-cancel)structured billing ledger
+34pp
First-month check-ins under 2 / weekkey-card access logs
+21pp
Front-desk complaint logged in CRMmember-services tickets
+14pp
No class or trainer booking in 30 daysapp & booking data
+12pp
Tenure under 90 daysmembership record
+9pp
Black Card → Classic downgradebilling record
+8pp
$142Mannualized member dues at risk from billing-driven churn if the trend holds — the number a fix protects
Retention by cohort first billing failure vs a clean first 90 days
● Clean cohort · 87% retained● Billing-failure cohort · 51% retained
A single billing failure in the first 90 days is worth a 36-point swing in one-year retention — the mechanism, on your own data.
Method & confidencewhy these are causes, not correlations
Model health every effect is estimated five ways and stress-tested before it earns a place here
9,186
reviews scored
28
dimensions per review
6
causes tested
5
estimators per effect
3/3
refutations passed
300
bootstrap resamples
✓ Placebo test passed✓ Random-common-cause stable✓ Subset re-estimation stable✓ Club fixed-effects confirmed✓ 95% bootstrap intervals
What to do — ranked by return
1 · Instrument churn, not the rating.
The leading indicator missing in Q1, from data you already own.
The leading indicator missing in Q1, from data you already own.
2 · Fund the front desk first.
Largest cause of an unhappy member (odds 35×), cheapest to fix.
Largest cause of an unhappy member (odds 35×), cheapest to fix.
3 · Repair billing trust.
Strongest leaving signal; don't re-add cancellation friction.
Strongest leaving signal; don't re-add cancellation friction.
4 · Work the named clubs.
Damage concentrates in a nameable urban set at 10–20× the rate.
Damage concentrates in a nameable urban set at 10–20× the rate.
Dimension Labs Causal Intelligence · illustrative monitoring preview. Driver effects, odds ratios and club rating gaps are seeded from the July 2026 causal analysis of 9,186 Google reviews across 33 Chicago clubs (g-computation ATEs with 95% bootstrap intervals, confirmed by a club fixed-effects logit). The weekly time series, alerts and week selector are fabricated to demonstrate what a live weekly causal monitor would show once wired to the review feed. No weekly figure here is a real observation.