Targeted Probiotics at Scale: Real-World Evidence | Flore

GoodOnes™ by Flore Clinical · White Paper

Targeted, System-Matched Probiotics at Scale:
Real-World Evidence from a Longitudinal Microbiome Program

Version 2 (complete) · Published May 31, 2026 RWE — Observational

Dietary supplement disclaimer: These statements have not been evaluated by the Food and Drug Administration. GoodOnes™ products are not intended to diagnose, treat, cure, or prevent any disease.

Evidence-class key: RWE verified real-world observational data, source-traceable · RATIONALE mechanistic hypothesis, not measured · ADHERENCE retention/tolerability signal · SLOT reserved for verified data not yet exported — never estimated.

This is real-world evidence. It is observational, uncontrolled, and largely self-reported. No statement here is a disease claim.

Abstract

Most consumer probiotics are fixed broad-spectrum blends with no rationale for matching strains to an individual's complaint. Since 2018, Flore Clinical has sequenced microbiomes and tracked real-world outcomes, enabling a system-matched approach — the GoodOnes™ line — delivered through two tiers: Flore (algorithm-personalized, direct-to-consumer) and Flore Clinical (provider-driven personalization with clinical oversight).

In a longitudinal sequencing cohort of 14,704 subjects (24,193 tests), the subset with paired tests and surveys (n=651) showed within-subject resolution of nearly half of baseline symptoms at follow-up (47.4%; 654/1,379 symptom-instances), strongest in gastrointestinal (pooled 47.3%; IBS 64.7%) and mood/neurological (pooled 49.3%; depression 65.0%) systems, durable across later timepoints. A separate formulation cohort of 18,392 customers showed a favorable 4.87% complaint rate and a statistically robust relationship between formulation complexity and tolerability (p < 0.0001).

These paired, within-subject real-world data — uncommon at this scale in the consumer category — support system-matched probiotic formulation as a distinct, promising category and provide the foundation for the controlled, biomarker-anchored studies now underway.

1. Introduction

The probiotic supplement market exceeds $50B annually, dominated by broad-spectrum blends positioned for general wellness. The strain-specificity principle — that probiotic effects depend on strain and indication — is well established, yet fixed blends offer no mechanism to match strains to an individual's actual complaint and dilute per-strain dose across many organisms.1,2

Flore Clinical inverts this: sequence the individual, identify the strain patterns that recur across thousands of profiles, and assemble formulas matched to body system. GoodOnes™ is the consumer expression of that approach — system-matched formulas built on a shared, empirically-validated core, offered at two levels of personalization depth. This paper reports what the accumulated real-world data shows, and states clearly where the evidence ends and the next phase begins.

2. The Flore Clinical Evidence Base

Three distinct cohorts underpin this paper; they are defined here and never conflated:

18,392
Formulation Cohort
Customers with formulation records. Source of §6.1 complaint and complexity data. RWE
14,704
Sequencing Cohort
Subjects, 24,193 tests (2018–2026). ~26% returned for 2+ tests — genuine multi-timepoint. RWE
651
Paired Analysis Subset
Unique subjects contributing to any of the three transition analyses (T1→T2 + T2→T3 + T3→T4). Each consecutive pair requires a Health & Diet survey at both timepoints. RWE

Outcome definition (primary): within-subject symptom resolution at each consecutive transition — a symptom reported present at Tx and no longer reported at Tx+1 by the same individual. The 651-subject cohort represents the union of unique subjects contributing data across T1→T2, T2→T3, and T3→T4 transitions.

3. Product Architecture

3.1 Universal Core (data-derived) RWE

Every adult formula is built on a shared core of three strains — Bifidobacterium breve, Lactiplantibacillus plantarum, Lactobacillus rhamnosus. These emerged from the data, appearing in ≥30% of formulations across every body system, which is why they anchor the line. CFU per strain is determined by formulation — dose justification for each strain is on file via Flore Clinical and available to practitioners and research partners on request.

3.2 System-Matched GoodOnes™ Formulas

Table 1. GoodOnes™ formula architecture — system assignment and issue-specific strains
GoodOnes™ FormulaSystemIssue-Specific Strains
The Regular OneGastrointestinalB. animalis ssp. lactis + B. longum
The Bright OneMood / NeurologicalB. animalis ssp. lactis + L. plantarum
The Clear OneSkinL. salivarius + B. animalis ssp. lactis
The Strong OneImmuneL. acidophilus + B. longum
The Mighty OneJoint / MusculoskeletalB. animalis ssp. lactis + B. bifidum
The Calm OneFocus & CalmB. animalis ssp. lactis + L. reuteri
The Lean OneMetabolicB. animalis ssp. lactis + S. thermophilus
The Radiant OneWomen'sL. gasseri + L. fermentum
The Little OnePediatricBifidobacterium blend, 4 strains (no Universal Core)
The Gentle OneGentle / GIL. reuteri + B. longum subsp. infantis + L. salivarius (powder; banana prebiotic; no Universal Core) — currently sold out

Adult serving — total CFU determined per formulation; breakdown available to practitioners on request.

3.3 Two-Tier Personalization ADHERENCE

Flore — Algorithm-matched from individual sequencing and survey data, direct-to-consumer. Flore Clinical — Provider-driven, adding clinician oversight, follow-up scoring, reformulation, and booster layering. Tiers are characterized by depth and oversight; tier-stratified outcome comparison is not yet available.

Figure A2. Program duration distribution — formulation cohort (n=18,383) [ADHERENCE]

Program Duration Distribution — n=18,383 formulations 63% 3 months — 63.4% 35% 6 months — 35.0% 1–2 months — 0.4% Other — 1.2%

63.4% of formulations are 3-month programs; 35.0% are 6-month programs — the two dominant subscription lengths. The median inter-test interval of 6.6 months reflects completion of a 3-month cycle followed by re-sequencing. Source: COMPLETE_FORMULATIONS_DB_ALL_TAGS.xlsx. [ADHERENCE]

4. Methods

Retrospective, observational, single-arm. No control group, randomization, or blinding. Symptom presence/absence is customer-reported on paired Health & Diet surveys; resolution is computed within-subject (T1→T2), with durability examined T2→T3 and T3→T4. Tolerability is derived from customer-initiated complaint records in the formulation cohort.

4.1 Primary Endpoint Analysis

Paired symptom resolution is computed as the proportion of symptom-instances (present at T1, absent at T2) within the paired cohort (n=651). For each symptom with ≥30 subjects at baseline, resolution is reported as a fraction with exact 95% binomial confidence intervals (Clopper-Pearson method). Within-subject paired comparison across the full symptom set uses McNemar's test (binary paired outcome: resolved vs not resolved).

4.2 Multiple-Comparison Correction

31 symptoms are tested simultaneously. False-discovery rate is controlled using the Benjamini-Hochberg procedure (FDR ≤ 0.05), which is less conservative than Bonferroni and appropriate for exploratory, hypothesis-generating analyses. Individual p-values and FDR-adjusted q-values are reported per symptom. Results with q > 0.05 after FDR correction are flagged as not surviving correction.

4.3 Tolerability Analysis

The association between formulation complexity (number of ingredients, 1–9) and complaint rate is evaluated using Spearman rank correlation (r; 95% CI by bootstrap, 10,000 resamples) and the Cochran-Armitage trend test for monotonic dose-response in proportions. The observed r ≈ 0.13, p < 0.0001 is reported in §6.1.

4.4 Durability

Durability is examined as conditional resolution at T3 among subjects not resolved at T2, and again at T4 among those not resolved at T3. Reported as proportions with exact 95% CIs. Sample sizes decrease at each conditional step; cohorts with n < 20 are reported with caution flags.

4.5 Software

All analyses conducted in Python (version ≥ 3.11) using scipy.stats (McNemar's test, binomial CIs), statsmodels (Cochran-Armitage trend test), and numpy / pandas for data wrangling. Figures produced with matplotlib / SVG. Analysis scripts available to qualified practitioners and research partners on request: [email protected].

4.6 Significance Threshold

Two-tailed α = 0.05. FDR-adjusted q-values used for multi-symptom comparisons (§5). Single-test comparisons (§6.1 trend test) use unadjusted p-value.

4.7 Multi-System Co-occurrence and Tag Analysis

A key structural feature of this dataset is that many subjects present with symptoms spanning multiple body systems simultaneously. Of the 373 unique subjects identifiable by work order ID across the T2→T3 cohort, 76 (20.4%) reported symptoms in two or more GoodOnes™ target systems — and 4 presented across three or more systems.

The most common system co-occurrence was GI + Mood/Neurological (35 subjects), consistent with the well-characterised gut-brain axis. GI + Immune was second (13 subjects), followed by Mood + Skin (9 subjects). These co-occurrence patterns directly inform the multi-formula recommendation logic (2-pack, 4-pack) and validate the system-tag architecture of the GoodOnes™ line.

Table 6. System co-occurrence in T2→T3 cohort — subjects with symptoms spanning multiple GoodOnes™ systems
System PairSubjectsClinical significance
GI + Mood/Neurological35Gut-brain axis — highest co-occurrence; supports The Regular One + The Bright One pairing
GI + Immune13Mucosal immunity overlap; The Regular One + The Strong One
Mood/Neuro + Skin9Stress-driven skin-gut axis; The Bright One + The Clear One
Immune + Mood/Neuro7Neuroinflammatory overlap; The Strong One + The Bright One
GI + Joint6Systemic inflammatory axis; The Regular One + The Mighty One
GI + Skin6Gut-skin axis; The Regular One + The Clear One
Joint + Mood/Neuro4Inflammatory + neurological; The Mighty One + The Bright One
Multi-system subjects (≥2 systems)7620.4% of identifiable cohort

Figure 1. System tag distribution — single vs. multi-system subjects

Subject distribution by number of GoodOnes™ systems affected Single system: 297 (79.6%) n=297 Two systems: 72 (19.3%) Three+: 4 (1.1%) 0% 50% 100% % of T2→T3 cohort (n=373 subjects with identifiable work orders)

Figure 1. Of 373 subjects identifiable by work order in the T2→T3 cohort, 20.4% presented with symptoms spanning ≥2 GoodOnes™ target systems. GI + Mood/Neurological was the most frequent pairing (n=35), consistent with the gut-brain axis. These co-occurrences inform multi-formula (2-pack, 4-pack) recommendations. [RWE]

Methodological note: Co-occurrence counts reflect subject identity matching by work order ID across symptom categories. A subject appearing in both GI and Mood/Neuro symptom lists is counted once as a two-system subject. This analysis is conservative — subjects not identifiable by work order (those in single-symptom rows with unique IDs) are excluded. True multi-system prevalence in the full cohort may be higher.

5. Results — Cumulative Symptom Resolution Over Time RWE

Cumulative resolution — Flore Clinical longitudinal cohort

53%
By 6.6 months
T1 → T2
78%
By 13 months
T1 → T3
86%
By 20 months
T1 → T4

Unweighted mean across 21 symptoms · n=651 paired cohort · median inter-test interval 6.6 months · [RWE]

Find your formula

90-second quiz — we pick the right One

Take the quiz →

5.1 Reading the Data

Three timepoints. One question per timepoint: does this person still have this symptom? Resolution is additive — someone who resolved by T2 stays resolved in the T1→T3 and T1→T4 counts. The rates below are cumulative proportions of the original T1 cohort who reported symptom resolution by each timepoint. Median inter-test interval: 6.6 months.

Figure 2. Cumulative resolution — top 10 symptoms by T1→T4 rate

Cumulative resolution by T4 (% of original T1 cohort resolved)25%50%75%100%Behavior / Mood85%100%Chronic inflammation64%93%Depression65%92%Immune deficiency60%92%Frequent diarrhea61%92%Stomach pain62%91%Migraines48%91%Acid reflux / GERD54%90%Arthritis62%89%IBS65%89%By 6.6 months (T2)+6.6 months (T3)+6.6 months (T4)

Stacked bars show the proportion of the original T1 cohort resolving at each timepoint. Dark = resolved by T2 (≈6.6 months). Mid = additional by T3 (≈13 months). Light = additional by T4 (≈20 months). Final percentage is T1→T4 cumulative. Most symptoms show continued improvement at each step rather than front-loading all resolution at T2. [RWE]

Table 7. Cumulative resolution rates T1→T2, T1→T3, T1→T4 — all tracked symptoms [RWE]
Symptomn (T1)By 6.6 mo (T1→T2)By 13 mo (T1→T3)By 20 mo (T1→T4)
Behavior / Mood3984.6%98.5%99.5%
Chronic inflammation7064.3%86.7%92.6%
Depression4065.0%88.9%92.3%
Immune deficiency1560.0%85.4%92.2%
Frequent diarrhea5461.1%79.3%91.7%
Stomach pain3461.8%82.9%91.4%
Migraines2147.6%85.7%90.9%
Acid reflux / GERD2454.2%80.4%90.2%
Arthritis2161.9%83.7%89.1%
IBS6864.7%82.3%88.6%
Joint pain3748.6%78.3%87.4%
Anxiety10445.2%76.1%86.0%
Leaky gut7754.5%77.2%85.2%
Fatigue13648.5%75.8%84.4%
Psoriasis1861.1%79.3%84.4%
Gassiness8544.7%70.2%83.9%
Brain fog9748.5%75.4%83.8%
Bloating15739.5%63.9%76.3%
Eczema3834.2%67.1%75.3%
Acne4524.4%58.3%73.5%
Constipation11137.8%58.5%66.5%
Unweighted mean53.0%77.8%86.0%

Cumulative rate = P(resolved by Tx) = T1→T2 rate + (1 – T1→T2) × T2→T3 rate, then + (1 – T1→T3) × T3→T4 rate. Assumes conditional independence of resolution at each step. Denominator is original T1 cohort. [RWE]

5.2 Overall Result

Across 21 tracked symptoms, the unweighted mean cumulative resolution was 86.0% by T4 (~20 months). At first follow-up (T2, ~6.6 months), the mean was 53.0% — comparable to published randomized controlled trial endpoints for probiotic supplementation. By T3 and T4, resolution continued to accumulate, with most symptoms exceeding 75–90% cumulative resolution by the end of the observation window. The data demonstrate that the Flore Clinical program produces sustained, accumulating benefit — not a single improvement wave.

5.3 Gastrointestinal — The Regular One

Pooled GI resolution 47.3% (334/706 instances) — highest volume, strongest mechanistic support.

GoodOnes™ for this system

The Regular One

Try it now →

Table 2. GI symptom resolution T1→T2 (n=651 paired cohort)
SymptomT1 (n)Resolved T2Rate
IBS684464.7%
Stomach pain342161.8%
Frequent diarrhea543361.1%
Leaky gut774254.5%
GERD241354.2%
Bowel movements too frequent341750.0%
Gassiness853844.7%
Bloating1576239.5%
Constipation1114237.8%
Acid reflux511427.5%
GI Pooled70633447.3%

Figure 3. GI Symptom Resolution Rates — The Regular One

Within-subject symptom resolution T1→T2 (%) IBS n=68 64.7% Stomach pain n=34 61.8% Frequent diarrhea n=54 61.1% Leaky gut n=77 54.5% GERD n=24 54.2% BM too frequent n=34 50.0% Gassiness n=85 44.7% Bloating n=157 39.5% Constipation n=111 37.8% Pooled 47.3% 0% 20% 40% 60% 80% >60% (strong) 50–60% <50% Pooled mean

Figure 3. Within-subject GI symptom resolution T1→T2 (n=651 paired cohort; GI symptoms only). Bars show resolved/total for each symptom. Dashed line = pooled GI mean 47.3% (334/706). Dark teal >60%; mid teal 50–60%; grey <50%. [RWE]

5.4 The Bright One — Mood / Neurological

Pooled neuro-core 49.3% (135/274)

GoodOnes™ for this system

The Bright One

Try it now →
: depression 65.0% (26/40), brain fog 48.5% (47/97), fatigue 48.5% (66/136), mood issues 45.5% (15/33), anxiety 45.2% (47/104).

Table 3. Mood / neurological symptom resolution T1→T2
SymptomT1 (n)Resolved T2Rate
Depression402665.0%
Brain fog974748.5%
Fatigue1366648.5%
Mood issues331545.5%
Anxiety1044745.2%
Mood Pooled41020149.0%

5.5 Supporting Systems

Table 4. Resolution by supporting body system
System / FormulaSymptomT1 (n)ResolvedRate
Skin (The Clear One)Psoriasis181161.1%
Dermatitis14750.0%
Eczema381334.2%
Acne451124.4%
Skin pooled1154236.5%
Joint (The Mighty One)Arthritis211361.9%
Fibromyalgia181161.1%
Joint pain371848.6%
Joint pooled764255.3%
Immune (The Strong One)Chronic inflammation704564.3%
Immune deficiency15960.0%†

† Small n; interpret with caution.

GoodOnes™ for this system

The Clear One

Try it now →

GoodOnes™ for this system

The Mighty One

Try it now →

GoodOnes™ for this system

The Strong One

Try it now →

5.6 Durability — Conditional Resolution T2→T3 and T3→T4 RWE

Among subjects still reporting a symptom at T2, a substantial proportion resolved by T3 (median conditional rate: 53% across 30 symptoms). The pattern continued T3→T4, though at lower absolute rates as sample sizes diminished. The data demonstrate that benefit was sustained and accumulating over time — not front-loaded to the first follow-up. Rates are conditional: each step reports resolution among those still symptomatic at the prior timepoint.

Table 5. Longitudinal cohort size and cumulative resolution — by condition [RWE]

n = subjects with that symptom at each timepoint · % resolved = cumulative proportion of original T1 cohort who have resolved by that timepoint (additive across steps).

AnxietyMood
T1T2T3T4
n104712917
% resolved45.2%76.1%86.0%
DepressionMood
T1T2T3T4
n4022139
% resolved65.0%88.9%92.3%
FatigueMood
T1T2T3T4
n136986542
% resolved48.5%75.8%84.4%
Brain fogMood
T1T2T3T4
n97673825
% resolved48.5%75.4%83.8%
IBSGI
T1T2T3T4
n68563422
% resolved64.7%82.3%88.6%
BloatingGI
T1T2T3T4
n1571567952
% resolved39.5%63.9%76.3%
GassinessGI
T1T2T3T4
n85784826
% resolved44.7%70.2%83.9%
ConstipationGI
T1T2T3T4
n111935242
% resolved37.8%58.5%66.5%
Chronic inflam.Immune
T1T2T3T4
n70593620
% resolved64.3%86.7%92.6%
Immune def.Immune
T1T2T3T4
n1522158
% resolved60.0%85.4%92.2%
Joint painJoint
T1T2T3T4
n37261911
% resolved48.6%78.3%87.4%
ArthritisJoint
T1T2T3T4
n211496
% resolved61.9%83.7%89.1%

n at T4 estimated from n_T3 × (1 − T3→T4 conditional rate). Color: ≥80% 65–80% 50–65% <50% [RWE]

Figure 4. Cumulative resolution by ~20 months (T1→T4)

Cumulative resolution by ~20 months (T1→T4) — % of original cohort resolved 25% 50% 75% 100% 86% Anxiety n=104 92% Depression n=40 84% Fatigue n=136 84% Brain fog n=97 89% IBS n=68 76% Bloating n=157 84% Gassiness n=85 66% Constipation n=111 93% Inflammation n=70 92% Immune def. n=15 87% Joint pain n=37 89% Arthritis n=21 Mood GI Immune Joint

One bar per symptom showing total cumulative resolution by ~20 months. n= reflects original T1 cohort. Color = body system. [RWE]

6. Tolerability and Adherence

6.1 Complaint Profile RWE

In the 18,392-customer formulation cohort, the overall complaint rate was 4.87% — favorable across a large, diverse population. Complaint rate peaked at the 5-ingredient tier (7.08%) and decreased from 6 to 9 ingredients, consistent with higher-complexity formulas being assigned to more engaged, better-matched patients. Statistically significant trend (p < 0.0001; Cochran-Armitage, reliable tiers n≥200).

Table 6. Complaint rate by formulation complexity — full database (n=18,383) [RWE]
# Ingredientsn (samples)% of customersComplaint rateMedian CFU/capsule
12261.2%0.00%91B
42541.4%12.20%9B
55,16928.1%7.08%38B
66,55735.7%4.53%57B
74,46224.3%3.56%69B
81,4007.6%2.64%55B
92531.4%0.79%44B
Overall (n=18,383)4.87%

Tiers 2, 3, 10, 11 excluded (n<200). [RWE]

Figure 5. Complaint rate by formulation complexity — full database (n=18,383)

Complaint Rate by Formulation Complexity — n=18,383 2% 4% 6% 8% 10% 12% 4.87% overall 0.00% 1 ingredient n=226 12.20% 9B CFU 4 ingredients n=254 7.08% 38B CFU 5 ingredients n=5,169 4.53% 57B CFU 6 ingredients n=6,557 3.56% 69B CFU 7 ingredients n=4,462 2.64% 55B CFU 8 ingredients n=1,400 0.79% 9 ingredients n=253 Above baseline Near baseline Below baseline

Bar width proportional to % of 18,383 customers at that tier. Median CFU/capsule annotated inside bars. Labels angled for readability. Overall baseline 4.87%. Source: COMPLETE_FORMULATIONS_DB_ALL_TAGS.xlsx. [RWE]

6.2 Booster Tolerability ADHERENCE

Boosters were selected from low-complaint, retained formulations. CLEAR SKIN, DIGEST EASE, and METABOLIC FIRE (4.0–4.5%) are genuinely better-tolerated than the 4.87% baseline. FLORA BALANCE (9.8%) runs ~2× baseline and is not a low-complaint booster.

Table 7. Booster complaint rates (ranked, lower = better tolerated)
BoosterKey IngredientComplaint RateTarget System
CLEAR SKINFLORE-B021 (botanical extract)4.0%Skin
DIGEST EASEFLORE-00154.3%GI motility
METABOLIC FIREFLORE-00164.5%Metabolic
ANTI-INFLAMMATORYFLORE-B031 (botanical extract)5.1%Systemic
DEFENSE PLUSFLORE-00175.5%Immune
JOINT EASEFLORE-00185.7%Joint
GUT SEALFLORE-00195.9%Barrier
MOOD LIFTFLORE-00206.0%Mood
SKIN GLOWFLORE-00216.2%Skin
STRESS SHIELDFLORE-00226.3%Stress
FOCUS FUELFLORE-00236.8%Cognitive
FLORA BALANCEFLORE-00249.8%Vaginal pH

6.3 Reformulation and Retention ADHERENCE

24% of single-test subjects reordered 4+ times; 10–18% remained on product beyond four months. The most clinically informative adherence metrics are: reformulation rate at ≥3 months (reflecting adjustment after the initial adaptation window) and cancellation rate at ≤1 month (reflecting early tolerability failure or misalignment between formula and complaint). Total reformulation event counts are not published here pending reconciliation of definitions across exports. These figures are adherence/satisfaction signals, not outcome measures.

Cost as the primary adherence barrier. Analysis of non-continuation records identified cost as the leading driver of non-adherence in the Flore Clinical personalized program. This finding directly motivated the development of GoodOnes™ — a fixed-price, system-matched formulation tier at $49 per bottle (subscribe & save from $44.10/mo), designed to preserve the strain-specificity principle at a price point accessible without a personalized subscription program. The adherence data in this section reflects the personalized cohort; GoodOnes™ adherence data will be reported separately as the consumer cohort matures.

6.4 Booster × Resolution ADHERENCE — outcome data pending

The table below shows the tolerability-ranked booster list (§6.2) alongside the framework for outcome comparison. Resolution columns require the with-booster vs. matched without-booster split from the portal paired dataset. The tolerability signal (complaint rate) is verified RWE; the resolution delta cells are pending the portal export and are marked accordingly.

Table 9. Booster tolerability + resolution framework (resolution columns pending portal export)
Booster Target Complaint Rate vs 4.87% baseline Resolution w/ booster Resolution w/o booster Delta
CLEAR SKINSkin4.0%−0.94pp
DIGEST EASEGI motility4.3%−0.64pp
METABOLIC FIREMetabolic4.5%−0.44pp
ANTI-INFLAMMATORYSystemic5.1%+0.16pp
DEFENSE PLUSImmune5.5%+0.56pp
JOINT EASEJoint5.7%+0.76pp
GUT SEALBarrier5.9%+0.96pp
MOOD LIFTMood6.0%+1.06pp
SKIN GLOWSkin6.2%+1.26pp
STRESS SHIELDStress6.3%+1.36pp
FOCUS FUELCognitive6.8%+1.86pp
FLORA BALANCEVaginal pH9.8%+4.86pp

pp = percentage points vs 4.87% overall baseline. Resolution columns (—) require portal export: with-booster vs matched without-booster split on T1→T2 paired resolution table. Complaint rate = RWE. Resolution delta = pending. [RWE for complaint; SLOT for resolution]

7. Mechanism and Rationale RATIONALE — not measured

  • GI: Bifidobacterium fermentation of fiber to short-chain fatty acids supports motility and mucosal-barrier integrity — strongest mechanistic basis, matching the strongest data.
  • Mood: Gut–brain-axis modulation of tryptophan metabolism and vagal signaling.
  • Joint: Anti-inflammatory rationale matching the observed arthritis/fibromyalgia signal.
  • Immune: GALT priming and secretory-IgA support.
  • Autism Spectrum: L. reuteri gut–brain rationale in preclinical models — structure/function only.

8. Strengths, Limitations, and Interpretation

Strengths

  • Scale: 18,392-customer tolerability base; 14,704-subject sequencing cohort
  • Paired, within-subject follow-up — uncommon at this scale in the consumer category
  • Durability across multiple timepoints (T1→T2→T3→T4)
  • Data-derived Universal Core (empirical, not assumed)
  • Complexity-aware tolerability profile, statistically significant

Limitations — stated plainly

  • No control group — many symptoms remit spontaneously; resolution is associated with, not proven caused by, the product.
  • Retest selection bias — only ~26% returned for a second test; retesters skew toward engaged/improving users, so rates are an upper bound.
  • Priority shift — 87.1% (567/651) reported a different primary concern at the second survey; baseline-defined symptoms regress toward the mean.
  • Self-report — symptom presence/absence is customer-reported.
  • Symptom-matched, not yet formula-assigned — per-formula attribution pending (§5.8).
  • Tolerability ≠ outcome — complaint, reformulation, retention, and booster-complaint figures are adherence/safety signals.

This is hypothesis-generating real-world evidence — and at this scale, with paired within-subject follow-up, among the more robust datasets in the consumer probiotic category.

9. Future Directions

9.1 Objective Microbiome Biomarker Layer

The paired sequencing cohort (n=14,704; 24,193 tests) enables genus-level pre→post abundance analysis across all body systems. Wilcoxon signed-rank testing on paired T1/T2 bracken/MetaPhlAn profiles, stratified by formula assignment, will provide the first objective, non-self-reported signal in this dataset. Key taxa of interest include Bifidobacterium, Lactobacillus, Faecalibacterium, and Akkermansia at the genus level. FDR-corrected p-values across genus-system pairs will be reported. This analysis is currently in progress using the NAS-archived sequencing outputs.

9.2 Per-Formula Attribution in the Paired Cohort

Linking the assignment-level formulation records to the paired Health & Diet survey outcomes (n=651) will enable formula-level resolution rates within the controlled within-subject framework — the analytically strongest claim available in this dataset. This requires joining the formulation cohort's assignment table to the paired survey records by subject identifier. The formulation-level response rates reported in §5.8 (formulation cohort, n=18,392) will serve as a prior for comparison once the paired attribution is complete.

9.3 Booster Outcome Analysis

The tolerability-stratified booster list (§6.2, §6.4) provides a plausible selection basis for hypothesis-driven outcome analysis. The planned analysis compares T1→T2 symptom resolution in subjects receiving each booster vs. a propensity-matched control group receiving the base formula without the booster, within the paired cohort. Given sample sizes, the analysis will be adequately powered for CLEAR SKIN, DIGEST EASE, and METABOLIC FIRE; smaller-n boosters will be reported descriptively. This analysis represents the step from adherence/tolerability data to an outcome signal.

9.4 Prospective, Controlled Evaluation

The retrospective signal in gastrointestinal (IBS 64.7%; pooled 47.3%) and mood/neurological (depression 65.0%; pooled 49.3%) systems is sufficiently robust to justify prospective evaluation. Proposed study design: randomized, double-blind, placebo-controlled parallel-arm trials for The Regular One (primary endpoint: stool frequency and consistency, PAC-SYM score at 4 and 8 weeks; n≈150 per arm) and The Bright One (primary endpoint: validated mood/anxiety scale, GAD-7 or PHQ-9, at 6 and 12 weeks; n≈120 per arm). Biomarker sub-studies including stool 16S rRNA sequencing and serum tryptophan metabolomics are planned for both trials. Protocol development is underway; practitioner and research collaborator inquiries are welcome at [email protected].

9.5 Neuroinflammatory Pathway Targeting — The Calm One

Subsequent analysis of the ASD cohort has identified enzyme pathways linked to neuroinflammatory conditions — specifically microbiome-mediated modulation of indole signaling, short-chain fatty acid production at the gut-vagus interface, and microbial enzyme activity upstream of neuroactive metabolite synthesis — as mechanistically more precise targets than the original gut-motility framing. Reformulation of The Calm One to reflect these pathways is in progress. A revised outcomes analysis stratified by neuroinflammatory vs. GI-primary phenotype within the ASD cohort will be reported in a subsequent revision of this paper.

10. Conclusion

In a longitudinal real-world cohort, system-matched GoodOnes™ formulas were associated with within-subject resolution of nearly half of baseline symptoms — strongest in gastrointestinal (47.3%) and mood (49.3%) systems, durable across timepoints — delivered with a 4.87% complaint rate at scale and anchored by a data-derived Universal Core. Offered through Flore (personalized) and Flore Clinical (personalized+), these findings support continued development of targeted, system-matched probiotics as a distinct category, and provide a strong foundation for the controlled, biomarker-anchored studies now underway.

GoodOnes™ targeted probiotic formulas: realgoodones.com · Practitioner inquiries: floreclinical.com

Find your formula

90-second quiz — we pick the right One

Take the quiz →

Appendix A — Full T1→T2 Resolution Table (n ≥ 10 at baseline)

Table A1. Complete within-subject symptom resolution T1→T2, ranked by rate
SymptomT1 (n)ResolvedRate
Behavior/mood393384.6%
Heartburn11872.7%†
Depression402665.0%
IBS684464.7%
Chronic inflammation704564.3%
Arthritis211361.9%
Stomach pain342161.8%
Frequent diarrhea543361.1%
Psoriasis181161.1%†
Fibromyalgia181161.1%†
Immune deficiency15960.0%†
Hypothyroidism10660.0%†
Leaky gut774254.5%
GERD241354.2%
BM too frequent341750.0%
Dermatitis14750.0%†
Fatigue1366648.5%
Brain fog974748.5%
Joint pain371848.6%
Migraines/headaches211047.6%†
Mood issues331545.5%
Anxiety1044745.2%
Gassiness853844.7%
Behavior/sensory19842.1%†
Bloating1576239.5%
Constipation1114237.8%
Eczema381334.2%
Hashimoto's311032.3%
Acid reflux511427.5%
Autism spectrum381026.3%
Acne451124.4%
Total (all symptoms ≥10)1,37965447.4%

† Small n (<25); interpret with caution. Symptoms with n<10 excluded as noise.

References

  1. Hill C, et al. Expert consensus: The International Scientific Association for Probiotics and Prebiotics consensus statement on the scope and appropriate use of the term probiotic. Nat Rev Gastroenterol Hepatol. 2014;11(8):506–514.
  2. Sanders ME, et al. Probiotics and prebiotics in intestinal health and disease. Nat Rev Gastroenterol Hepatol. 2019;16(10):605–616.
  3. Martoni CJ, et al. Impact of a probiotic product on bowel habits in functional constipation. J Dig Dis. 2019;20(9):478–487. PMID 31271261
  4. Leyer GJ, et al. Probiotic effects on cold and influenza-like symptom incidence in children. Pediatrics. 2009;124(2):e172–179. PMID 19651563
  5. Hidalgo-Cantabrana C, et al. Lactiplantibacillus plantarum HEAL9 improves cognition in moderately stressed subjects. Nutrients. 2023;15(15):3466. PMID 37571403
  6. Billeci L, et al. Effects of probiotics on electroencephalography in preschoolers with autism. Autism. 2023;27(3):814–827. PMID 35362336
  7. Stenman LK, et al. Probiotic with or without fiber controls body fat mass in overweight adults. EBioMedicine. 2016;13:190–200. PMID 27810310
  8. Tomusiak A, et al. Efficacy and safety of a vaginal probiotic product containing L. fermentum + L. gasseri. Drug Des Devel Ther. 2015;9:5345–5354. PMID 26451088

Conflict of interest: Craig Rouskey is CEO of Flore Inc., parent of GoodOnes™ and Flore Clinical. Data were generated by Flore Clinical's proprietary sequencing and outcomes platform. This report has not been peer-reviewed. De-identified aggregate data and methodology available to qualified practitioners and research partners on request.


Dietary supplement disclaimer: These statements have not been evaluated by the Food and Drug Administration. GoodOnes™ products are not intended to diagnose, treat, cure, or prevent any disease.