Dialogue · SoCal Category Study · August 2026

Where the better-for-you aisle grows next

In an era of looksmaxxing, sugar-aversion, and mogging (don't worry if you don't know these terms, half of us at Dialogue still don't...), one would think the conditions are perfectly set for the “better-for-you” branded drinks. We conducted a deep-dive of the category, using our AI interviewing platform to conduct 20+ minute conversations with 100 people who already buy these drinks, start to finish in four days.

We drilled in on five brands: Liquid Death, OLIPOP, Health-Ade, GT's Living Foods, and Ghia (all of which can be found at your local Erewhon!). One thing to be upfront about, because it shapes everything below: to qualify for this study you had to have bought at least one of those five in the past six months. Nobody here is a skeptic we dragged in off the street, which means this is not a read on whether the category can win new people. Rather, it's a read on how well the category retains the buyers it already has.

Across the board, the category has a retention issue. 35 out of the 100 participants have not kept any of the five in their rotation, despite having bought at least one of the drinks in the past six months. The best performing brand retained 42% of its triers. When we asked participants what their last better-for-you drink was bought instead of, that — not the brand or the claim on the can — was what tracked with retention.
100SoCal category buyers
23.7min per conversation
4days, start to finish
2,020screened out to find them
35 of 100retained none of the five
01 · The funnel

34 of 99 Triers Retained Nothing

Read the first two bars below as a description of our sample rather than as a conversion or win: the screener in our study rejected anyone who had not bought one of these five drinks in the past six months. The third bar is the one that carries information. “Retained” here means the participant keeps the brand in their rotation; it's their own account of what they drink now.
Liquid Death
96 aware
78 tried
25 retained
32%of triers retained
OLIPOP
99 aware
92 tried
39 retained
42%of triers retained
Health-Ade
69 aware
57 tried
22 retained
39%of triers retained
GT's Living Foods
53 aware
42 tried
16 retained
38%of triers retained
Ghia
38 aware
13 tried
3 retained
23%of triers retained

Some brands certainly retain customers better than others: OLIPOP holds 42% of its triers against Ghia's 23%, close to double. Even though OLIPOP might be the best in this category, its ceiling is still quite low—the best performer in a five-brand set loses well over half of everyone who tries it. Awareness is where brands diverge most, running from 38 to 99 across the five, a 61-point spread compared to 19 points on retention.

Liquid Death recruits triers about as well as OLIPOP does. The entire gap between them opens after the first can.Aware 96 → tried 78 → retained 25

Ghia is the one brand in the set with a genuine awareness issue, and it is a cheaper job than it looks. It is the least known at 38 aware, and the only brand where most people who have heard of it still never try it, even among category buyers: 34% trial among the aware, against 79-93% everywhere else. Nobody is tasting Ghia and turning it down, they're simply walking past it on their morning Erewhon routines. A brand people walk past needs encounter and legibility, which are distribution and pack decisions, not a preference fix.

02 · Substitution

What the drink replaced decides whether it survives.

We've illustrated churn among the people who tried each brand, split by what their “better-for-you” occasion displaced. For example, if a participant in our study mentioned that they picked up a Liquid Death to replace a Coke, Liquid Death in that instance “displaced” a soda. Conversely, if a participant noted that they selected an OLIPOP in lieu of plain water, we categorize them as “giving nothing up.” Each participant counts once for every brand they tried, 282 in all. Churn is defined as a participant trying the brand and no longer keeping it in their rotation of beverages they drink.

Pick a brand to drill in more!
59%churn when something was given upn=188
73%churn when nothing wasn=92
Gave something up
Sodasmall base
base too small to rate
Coffee or teasmall base
base too small to rate
Alcoholsmall base
base too small to rate
Energy drinksmall base
base too small to rate
Sports drinksmall base
base too small to rate
Seltzersmall base
base too small to rate
Foodsmall base
base too small to rate
Juicesmall base
base too small to rate
Nicotine / vapesmall base
base too small to rate
Prescriptionsmall base
base too small to rate
Gave nothing up
Plain watersmall base
base too small to rate
Nothing at allsmall base
base too small to rate

Churn by what the drink replaced, every brand

All five

Gave something up: 59% churn (n=188) · Gave nothing up: 73% churn (n=92) · +14 points difference, small base

What it replacednChurn
Soda9858%
Coffee or tea4168%
Alcohol2255%
Energy drink1765% *
Sports drink1479% *
Seltzer1267% *
Food1155% *
Juice7
Nicotine / vape3
Prescription3
Plain water6371%
Nothing at all3471%

Liquid Death

Gave something up: 66% churn (n=50) · Gave nothing up: 74% churn (n=27) · +8 points difference, small base

What it replacednChurn
Soda2871%
Coffee or tea1060% *
Alcohol5
Energy drink5
Sports drink3
Seltzer3
Food3
Juice3
Prescription1
Plain water1878% *
Nothing at all1070% *

OLIPOP

Gave something up: 51% churn (n=63) · Gave nothing up: 75% churn (n=28) · +24 points difference, small base

What it replacednChurn
Soda3435%
Coffee or tea1392% *
Alcohol6
Energy drink6
Sports drink4
Seltzer3
Food4
Juice3
Nicotine / vape1
Prescription1
Plain water1861% *
Nothing at all1283% *

Health-Ade

Gave something up: 58% churn (n=38) · Gave nothing up: 68% churn (n=19) · +10 points difference, small base

What it replacednChurn
Soda1662% *
Coffee or tea8
Alcohol5
Energy drink2
Sports drink4
Seltzer3
Food2
Juice1
Nicotine / vape1
Prescription1
Plain water1471% *
Nothing at all6

GT's Living Foods

Gave something up: 61% churn (n=28) · Gave nothing up: 64% churn (n=14) · +3 points difference, small base

What it replacednChurn
Soda1486% *
Coffee or tea9
Alcohol5
Energy drink4
Sports drink2
Seltzer2
Food2
Nicotine / vape1
Plain water1070% *
Nothing at all5

Ghia

Gave something up: n=9, base too small to rate · Gave nothing up: n=4, base too small to rate

What it replacednChurn
Soda6
Coffee or tea1
Alcohol1
Sports drink1
Seltzer1
Plain water3
Nothing at all1

* base of 10–19 relationships. Rows with fewer than 10 relationships show no rate.

Sub-buckets sit inside their group and are multi-label, so someone who displaced both soda and coffee appears in both rows and the rows do not sum to the group. Anything under 10 relationships shows its count without a rate, because a percentage on n=3 is noise. And one honest limit: the displacement was recorded in section 4, before any brand was named, so a row means “people who tried this brand and whose occasion displaced X” — not “people who replaced X with this brand.”

OLIPOP keeps 65% of the people whose occasion displaced soda, and 25% of the people who gave up nothing.Churn 35% against 75% · n=34 and 28

The widest gap in the study belongs to one brand. For OLIPOP the trade looks like everything: people whose occasion displaced soda churn at 35%, while people who gave up nothing churn at 75% — on 34 and 28 relationships. OLIPOP is a healthier soda, it is bought against a named rival, and where that rival is real the habit appears to hold.

Liquid Death is the counter-case, and it shows no meaningful gap. Its soda-displacers churn at 71% and its gave-nothing group at 74% — a difference of only 3 points. This makes sense intuitively — Liquid Death is water. Swapping a soda for canned water is not the same trade as swapping it for something that tastes like soda, and this flows down clearly into retention data.

Read across all five and the pattern is milder but pointing the same way: 59% churn where something was given up against 73% where nothing was. Treat every one of these splits as a direction to test, not a result: a hundred people is a small base, they were screened to buyers of these five brands in one region, and the same person appears in the pooled figure once for every brand they tried. The category-level version of this pattern is mild for a reason: it averages OLIPOP's strong effect with Liquid Death's absent one. One participant named the hazard more cleanly than we could have: “every time I grab a drink, it is to replace water in my life”. When a three-dollar can replaces tap water, “better” has nothing left to be better than, and the purchase has to survive on taste, ritual and identity alone.

Whether the swap felt deliberate does not drive retention. Thirty-seven participants called it a conscious decision and 22 said it simply happened, and retention is flat across them — 65% against 68% retaining at least one, means of 1.08 and 1.09. That distinction is worth noting, because a brand can put a rival in the frame and cannot make a habit feel intentional. Where no rival was in the frame, people said so unprompted: “I was trying to replace soda with it, but I just liked soda too much”, and “I love the taste of a Diet Coke that comes out of a fountain”.

Two occasions at the far edge are worth recording even though neither generalizes. One participant replaced vaping with flavored functional drinks. Another replaced a prescription. Single cases, but they mark out the real competitive set, which is wider and more winnable than the soda aisle: every other way a person changes how they feel at three in the afternoon.

03 · Brand by brand

Five brands, five different next moves, and only one of them is an awareness buy

Each barrier is specific, and each carries a different instruction. Select a brand.

The name is the funnel, in both directions

The uber-masculine brand has near-total awareness (96 of 100) and the strongest venue-discovery engine in the set: participants meet it at concerts and festivals where it is often the only water sold. The job it does is social. It is “water that makes you feel better about not drinking a beer”, or as one participant put it, “like a sober man's water”.

While the brand maintains superb awareness, being aware of a brand is different than actually liking the brand's image or even knowing what the brand actually sells. Twelve participants refused or resisted on the name alone, and eleven had the category wrong entirely, six thinking it an energy drink and four thinking it alcoholic. Liquid Death's retention is the lowest of the four established brands at 32%, because once the novelty is spent the arithmetic surfaces: “I'm not paying like three or four dollars for a can of water”.

The deeper risk is identity fit, not price. “I feel kind of corny holding it” and “I don't fit into the stereotype where I feel like I have to fit in by having a canned beverage at, like, a bar” are refusals no discount reaches.

The move. Keep the name; it is buying trial no media plan could afford, and the 12 refusers are the cost of that, not a problem to solve. The gap to close is the second purchase at home, where the drink competes with the tap and loses on arithmetic. Give it a home occasion with a rival in it — the drink you hold instead of a beer after work, which is the job participants already describe — and a multipack price that survives comparison to a case of water.

Owns the soda occasion, leaks on unit economics

The most known (99) and most tried (92) brand in the set, with the most stable self-description in the whole study: a healthier soda. Where it wins, it wins cleanly, displacing a specific Coke or Diet Coke in a specific moment. One participant's rule is exact: “If I'm eating something that's really heavy or fatty, I'll try to, to drink an Olipop instead of a Coke”.

Its losses are not taste losses. Thirteen participants named price directly, and pack format compounds it: “it only comes in a four-pack, which I can finish easily in one day”. Where the format fails, the habit never forms.

The one callout is there is a bit of a fuzzy brand boundary. Two participants could not reliably tell OLIPOP from Poppi and described buying both interchangeably: “I don't know if that's Olipop or Poppi”.

The move. This brand has the category's best claim position already: a named rival, in a named occasion, verifiable on the label. Do not spend on awareness at 99 aware. Spend on unit economics — single-serve in the cooler door at an impulse price, out of the bulk aisle — and on distinctiveness against Poppi, because trial this high with retention at 42% means the leak is in the repeat, and a repeat you cannot attribute is a repeat you may be buying for a competitor.

Recruits on the shelf, loses on the label

Aware 69, tried 57, retained 22. First trial is driven by flavor range and the glass bottle. The losses cluster on sugar content, price against GT's at comparable quality, and bottle ergonomics.

The brand's own name is doing unintended work. One participant read the product as “it represents like an indulgence while pretending to be a healthy drink”, a judgment made harsher by the health promise in the name. Another described the packaging as “almost medicinal”, which is a compliment about seriousness and a warning about pleasure at the same time, and a third found the pack design itself a barrier: “the packaging is a little bit intimidating”.

Its sharpest problem is not rejection but blankness. Asked what stood out, one participant answered: “It seems as good or as not good, as excellent or as not excellent as the next one”. Against GT's, several treat the choice as no choice at all: “they are pretty interchangeable in my mind”.

The move. Blankness is a cheaper problem than rejection: 57 people have tried it and formed no view, which means it is an asset waiting for a reason. Pick one occasion this brand intends to own and make the pack argue for it, rather than competing with GT's on the shelf where the tiebreaker is promotion. The sugar figure is the credible claim here, and putting a number where the health language currently sits also treats the medicinal read.

Story and tenure, not marketing

Aware 53, tried 42, retained 16. The lowest awareness of the three established players, and the most durable loyalty per person reached. Retention runs on narrative and habit rather than claims: participants cite fifteen-year routines, “knowing the backstory of GT's mom's cancer”, and the fact that “It's an LA company”.

The category-entry role matters commercially. GT's is repeatedly named as the brand that introduced someone to kombucha, which makes its awareness ceiling the category's ceiling too.

One structural artifact worth naming: two participants treated Synergy, GT's own flagship line, as a separate and competing brand: “I think it's, like, Synergy, I think is the kombucha”. The branded house is not reading as one house.

The move. This is the one brand in the set where an awareness buy is justified, because what it retains on — founder story, local origin, tenure — is working on everyone it reaches and only 53 have been reached. The story is the asset, so spend it on the kombucha drinkers who have never met the brand rather than on the category at large. Fixing the Synergy boundary is close to free and recovers awareness the brand has already paid for.

A comprehension problem before a preference problem

The outlier on both gates. Least known at 38 aware, and uniquely, most people who have heard of it still never try it: 34% trial-of-aware against 79 to 93% for the other four.

The failure is legibility. Five participants confidently misdescribed it, most often as a kombucha or a chia-seed drink, including “does Ghia have chia seeds in it”. A name that produces the wrong category expectation costs a trial before price or taste is ever reached. One participant declined on the name's sound alone.

The three who retained it describe it precisely and warmly as a non-alcoholic aperitif. The product is not the problem. Encounter is.

The move. Do not buy preference; buy encounter and legibility. The three keepers describe the product perfectly, which says the proposition converts once it lands. Say the category on the front of the pack, and put the bottle where the occasion already exists — the restaurant and bar moment where someone is not drinking and wants something to hold. That is the one displacement in this study with a real vice on the other side of it and almost no better-for-you competition sitting in it.

04 · So what

Five plays, and how you would know whether they worked

If you run a brand in this category or next to it, this is where we would put the next dollar. Each play carries the evidence it rests on, the test we would run first, and the number we would hold it to — because a recommendation you cannot falsify is an opinion.
1

Put a named rival on the other side of the trade

The drink that replaced something retained 1.16 brands against 0.78 for the drink that replaced water or nothing, on matched trial. For OLIPOP the same split is the difference between 35% churn and 75%.

TestTwo creative cells in one market: one names the incumbent and the occasion (“instead of a Coke with wings”), one runs current category language.
MeasureSecond purchase within eight weeks per first-time buyer, not awareness lift.
2

Buy the occasion, not more reach into your own buyers

Awareness is close to spent among the buyers you already have: the top two brands sit at 96 and 99 aware, and retention still tops out at 42%. Whether reach converts people who don't yet buy the category is a real question this sample cannot answer.

TestPick two cities that look alike on size, demographics and retail mix. In one, point a defined share of media at a single occasion for a quarter — the heavy meal, the venue, the 3pm slump. In the other, change nothing, so any difference in repeat buying is attributable to the occasion rather than to the cities.
MeasureRepeat rate per buyer in the test city against the unchanged one.
3

Fix format before you touch taste

Thirteen participants named OLIPOP's price directly, one finishes a four-pack in a day, and bulk-only shelving recurs across brands. These arrive in survey data looking like preference losses; they are format losses, and cheaper to fix.

TestSingle-serve at an impulse price in cooler doors and checkout, out of the bulk aisle.
MeasureShare of first-time buyers who repeat within 14 days.
4

Defend the boundary you already paid for

Two participants could not separate OLIPOP from Poppi, and two treated Synergy, GT's own flagship line, as a rival. That is awareness already paid for and not banked.

TestUnaided attribution check: show pack, ask the brand, before any prompt.
MeasureCorrect attribution rate, tracked against the confusable competitor.
5

Win a slot in the rotation next door

31 people here already hold two or more of the five at once, and 34 hold exactly one, so a second better-for-you brand in the same rotation is observed behavior rather than a hypothesis. The largest co-held pair is Liquid Death with OLIPOP at 12 people, and those participants describe the two as doing different jobs, not competing for the same slot.

A buyer who already accepts the category is cheaper to reach than a soda drinker, and the ask is smaller: you are competing for one more slot in a rotation, not for a whole allegiance. The exception is the kombucha pair, where GT's and Health-Ade are decided by whichever is on promotion — there you would be buying a switch, and a discounted one.

TestTarget declared buyers of one named adjacent brand with the occasion you own and they do not — not a taste comparison. Hold back a matched slice of the same audience so the lift is readable.
MeasureOf the buyers you acquire, the share still holding both brands twelve weeks on, counted separately from those who switched. Then repeat value per acquired buyer over a year, in your own purchase data: this study can size the behavior, not price it.

What we would not fund

More awareness aimed at people already inside the category, where the top two brands sit at 96 and 99 aware. Converting the 12 people who refuse Liquid Death on the name, who are the price of a name that recruits everyone else. New claim language built from the category's own vocabulary, which is the part shoppers already discount. And arguing with the wellness-fatigue minority in creative, which is a tracking question, not a campaign.

Before you go

Get the next study before anyone else

This was the Southern California better-for-you read. There will be more — self-funded, fielded in days, published in full, just like this one. Leave your email and each new CPG study arrives the day it drops.

Leave your email → Want a read like this on your own category? Tell us what you sell.
05 · Method

How this was made

100 interviews, 11 sections, four days

Dialogue's AI moderator ran every interview — open elicitation, critical-incident anchoring, five unaided brand deep-dives — averaging 23.7 minutes and a median of 49 participant turns. Brands were withheld until section 5, so the occasion findings are unprompted. 100 Southern Californians: 79 LA / Orange / Inland Empire, 17 San Diego, 4 Santa Barbara / Ventura; 55 women, 43 men, 2 other.

Who qualified. The screener rejected anyone who answered “none of the above” when asked which of the five they had bought in the last six months. So this is a study of the category's own buyers: read every trial and awareness figure as a within-buyer rate, higher than a general population's. What the design buys you is the retention read — the people who churned here are people the category had and lost.

  • Every count is coded participant by participant, then recomputed from the code files. No figure here is an unverified model estimate.
  • Every quote is an exact string match against its transcript. 3 are withheld because those participants disclosed a health condition; three unverified allegations about brands appear nowhere.
  • No significance tests are quoted here. At 100 participants, screened to buyers of five brands in one region, every split in this readout is a direction worth testing against real purchase data, not a result.
  • Limits: awareness is unaided, so less-known brands are understated; brand order was fixed, so later brands carry fatigue; retention is self-reported, which is why every play is measured against purchase data; coding is single-coder.

Nobody in this category commissioned this study, and no brand named in it has seen the results before you.

Close