Issue 6 — Sources

Every factual claim in Issue 6, with its primary source. All data from a ChatGPT session run on April 19, 2026. Screenshots on file.

---

Claim: The five prompts run in a fresh ChatGPT session on April 19, 2026: (1) Best flavored sparkling water brands, (2) What do you think of Waterloo sparkling water?, (3) Why did you include (or not include) Waterloo in your top flavored sparkling water picks?, (4) What would Waterloo need to change to be cited as a top flavored sparkling water brand?, (5) What sources did you use to evaluate Waterloo specifically?

Source: ChatGPT session dated 2026-04-19, screenshots on file.

---

Claim: ChatGPT’s top three picks for flavored sparkling water: Spindrift (best overall, real fruit juice), LaCroix (best zero-calorie classic), Waterloo (best bold flavor).

Source: ChatGPT session on “Best flavored sparkling water brands,” April 19, 2026. Screenshot at /issue-6-screenshots/01-top-picks.png.

---

Claim: ChatGPT quote on Waterloo overall: “One of the better zero-calorie sparkling waters — but only if you actually want strong flavor without sweetness. If you expect soda-like taste, you’ll be disappointed.”

Source: Same session, Prompt 2.

---

Claim: ChatGPT’s three reasons Waterloo wasn’t ranked #1: (a) no real juice — less authentic than Spindrift; (b) inconsistent flavor lineup — “some flavors are top-tier, others are widely criticized”; (c) still not a true soda replacement — “brands with juice or minerals feel more complete.”

Source: Same session, Prompt 3. Screenshot at /issue-6-screenshots/02-not-number-one.png.

---

Claim: ChatGPT’s five-point product roadmap for Waterloo: (1) Upgrade from natural flavor to real ingredients — add real juice tier (5–10 calories) or minerals like Topo Chico; (2) Cut weak SKUs — Grape and Watermelon named as underperformers; double down on Lemon-Lime, Summer Berry, Raspberry Nectarine; (3) Improve mouthfeel via slight salinity or fuller carbonation; (4) Pick one positioning lane — currently stuck between LaCroix/Bubly (cheap) and Spindrift/Topo Chico (premium); (5) Tighten carbonation consistency.

Source: Same session, Prompt 4. Screenshot at /issue-6-screenshots/03-roadmap.png.

---

Claim: ChatGPT’s self-disclosed source weighting for evaluating Waterloo: Professional taste tests (Sporked, Tasting Table, Daily Meal) — 50%; Comparative rankings (Good Housekeeping, Paste, Consumer Reports) — 25%; Ingredients and brand data — 15%; Consumer sentiment (Reddit, reviews) — 10%.

Source: Same session, Prompt 5. Screenshot at /issue-6-screenshots/04-source-weighting.png.

---

Correction policy

If any claim in Issue 6 or in this sources post is wrong, it will be corrected here. Corrections are dated and attributed. We do not silently edit. See the commit history for a full record of changes. Email sacha@paidaisearch.com to flag an error.