Reports

Survey-augmentation studies wired to the Ario panel.

Pinned
1 report
MW Chocolate USA · Survey Augmentation
Stated-pref survey · N=504 human + 500 synthetic Ario panel · 2,924 buyers · 2021–2026 Updated today
15 / 24 questions fully answerable from transactions
›
CPG · Confectionery

MW Chocolate USA — Survey Augmentation

A 25-question stated-preference survey on online chocolate buying. Ario augments it with revealed behavior from a panel of US Amazon shoppers — 2,924 panelists, 578K orders, 2021–2026.

Source survey:MW_Chocolate_Programming_Draft · Field:Apr 2026 · Audience:US 18+ · online grocery 6mo+ · choc 6mo+
Panel
2,924
unique Amazon panelists
Confectionery buyers
665
23% of panel
Confectionery orders
2,539
0.4% of all orders
Confectionery spend
$31K
5-yr, 2,945 line items
Time window
5 yrs
Jan 2021 → May 2026
Q coverage
92%
22 of 24 full or partial

Survey question coverage

Can Ario answer each of the 25 questions from transactions alone?
ID Question Ario verdict Why

Answers we can give from transactions

Revealed preference, not stated
C3 · First-choice brand
Reese's, Hershey, M&M's lead — top 3 reach two-thirds of buyers
Across 665 confectionery buyers, Reese's reaches 25%, Hershey 23%, M&M's 18%. Top 3 = 65% of buyers.
Inside "Private Label / Other" — confectionery only
The remaining 58% of confectionery buyers fall into the unbranded / specialty bucket — mostly hot cocoa mix, small craft brands, and store-brand chocolate. Decomposed (top contributors):
Brand
Buyers
% reach
Hot cocoa mixes (drinking chocolate, generic)
72
8%
Chocolove (craft)
18
2%
barkTHINS
7
0.8%
Theo Chocolate (craft)
6
0.7%
Tony's Chocolonely (craft)
5
0.6%
365 / Whole Foods (private label)
4
0.4%
Milka (import)
3
0.3%
Amazon Brand chocolate (private label)
2
0.2%
Trader Joe's (private label)
2
0.2%
Ritter Sport, Great Value, Green & Black's, Nestle (misc, ≤1 buyer each)
4
0.4%
Unbranded / store-brand long tail (2,401 items)
—
—
Buyer reach = % of 665 confectionery buyers who bought brand X at least once. Multi-brand buyers counted in each. Decomposition based on regex scan of 2,945 confectionery line items, excluding the named brands shown in the main chart.
Q043 · Trip type
Most chocolate baskets are tiny — fill-in and last-minute dominate
Median basket containing chocolate is just 2 items. 38% of confectionery orders are chocolate-only (1 line item); only 10% are stock-up trips of 16+ items.
What these baskets actually look like
Four real receipts pulled from the panel — chocolate items highlighted.
Last-minute · 1 item User 2428 · 2026-02-04
🍫
$23.02
DOVE PROMISES Milk, Dark & Caramel Assorted Chocolates
Solo chocolate purchase 10 days before Valentine's — clear gift / treat intent.
Fill-in · 2 items User 5107 · 2026-03-31
$8.99
Easter Bunny Cake Topper w/ Eyes & Nose
🍫
$20.99
Ferrero Rocher 13-Count Premium Gourmet Hazelnut
Easter prep run — chocolate is the centerpiece of an occasion bundle.
Main grocery · 9 items User 213 · 2025-02-24
$11.51
PREPARED FOODS Hot Bar
$1.43
Icelandic Provisions Key Lime Skyr Yogurt
$11.49
AURORA Australian Crystallized Ginger
$2.89
365 by Whole Foods Italian Herb Pasta Sauce
🍫
$3.69
CHOCOLOVE Rich Dark Chocolate Bar, 3.2 OZ
$8.99
Rao's Homemade Vodka Sauce, 24 oz
$1.43
Icelandic Provisions Cold Brew Coffee Skyr
+ 2 more items. Real weekly grocery run — one dark chocolate bar as a small treat amid yogurts & pantry staples.
Stock-up · 22 items User 482 · 2024-11-01
$52.90
Whole Foods Pasture-Raised Whole Chicken
$26.45
SILVER HILLS BAKERY Little Big Bread (frozen)
$14.97
Teton Waters Grass-Fed Beef Hot Dogs
🍫
$20.07
Unreal Dark Chocolate Peanut Butter Cups, 4.2 oz
🍫
$6.69
UNREAL Dark Chocolate Almond Butter Cups 6-Pk
🍫
$6.69
UNREAL Dark Chocolate Coconut Bars (Vegan, GF)
$22.47
Boulder Canyon Avocado Oil Sea Salt Chips (6 ct)
+ 15 more items. Full pantry restock — three UNREAL chocolate SKUs at once is a clear "stockpile favorites" pattern.
Trip type inferred from total basket line-item count of the order containing chocolate.
NQ3 · Planned vs impulse — via shopper trait
~1 in 4 shoppers shows an impulsive trait — they're the ones adding chocolate without planning
We can't see whether a specific chocolate was added impulsively. But we can measure each shopper's lifetime return + cancel rate across all categories — a strong marker of impulsive buying behavior. 28% are impulsive (≥15% return rate), 25% are deliberate (<5%), 47% are in between.
Why this works: almost no one returns a chocolate bar — too small to bother. So return data on chocolate alone is useless. But the same shopper's overall return rate on bigger items (apparel, electronics, household) reveals whether they tend to buy impulsively and regret it. Heavy returners are the same people who toss chocolate into a cart on a whim.
Q042 · Occasion — event-window lift
Mother's Day is the only holiday that under-indexes online
Holiday 2.0x, Valentine's 2.0x, Easter 1.9x above baseline. Mother's Day 0.8x under-indexes — that moment goes through physical retail. Halloween near-flat at 1.0x.
How to read this: we look at the week or two before each holiday and ask "did chocolate sales spike?" A score of 1.0x = no spike (sales same as a normal week). 2.0x = sales doubled. 0.8x = sales were below normal — i.e. shoppers are buying chocolate somewhere else for that holiday.
C5 · Repeat purchase
Roughly 1 in 3 confectionery buyers rebuy the exact SKU
28%
of 665 confectionery buyers repurchased the same chocolate product at least once over the 5-year window.
Exact product-name match. Lower bound — variations in title formatting suppress the true rate.
B7 · Importance ranking — inferred
Flavor variety and brand dominate the inferred #1 spot
When we score each shopper on what they appear to optimize for, 35% rank Brand #1, 32% Flavor / variety, 17% Price, 14% Pack size. Past purchases and ingredients trail at <2% each as the inferred top driver.
How we score it: for each shopper we calculate a 0–1 signal on six factors using their actual purchases. Brand = how concentrated their buys are in one name. Price = inverse of their average item price. Flavor / variety = distinct SKUs / purchases. Pack size = share of party-pack / fun-size / single-serve formats. Ingredients = share tagged organic / sugar-free / vegan / dark / fair-trade. Past purchases = repeat-buy rate. Rank top 3. The stated ranking still requires the survey, but this is a strong starting point.
S7 · Order frequency — annualized
85% of confectionery buyers shop online weekly or every other week
62% shop weekly+, 23% every 1–2 weeks. Confectionery online is a habit of high-cadence shoppers — chocolate lands in their cart because the cart is open often.
How to read this: we count all Amazon orders per user (any category), then divide by the user's tenure on the panel to get orders/year. Buckets map to the survey's frequency options.

User stories — what the data tells you about real people

3 panelists, complete confectionery timelines
Holiday Valentine's Easter Mother's Day Halloween Everyday
U149
The Hershey loyalist on autopilot
User 149 · 5-yr history · 76 confectionery items
76conf items
$3545-yr spend
0Reese's bought
Bought the same HERSHEY'S Extra-Large Milk Chocolate with Almonds Bar nine times in Jan-Feb 2021 alone, and keeps coming back every Holiday season. Zero Reese's despite the shared parent company. A textbook single-brand loyalist whose preference is provable from timestamps.
Sample real transactions
U506
The craft explorer who graduated to baking
User 506 · 5-yr history · 33 confectionery items
20items in 2021
7items in 2022–24
6×raw cacao in 2025
2021: sampled 6+ small-batch craft brands — Endangered Species, Theo, Hu Kitchen, Lily's, NuGo, barkTHINS. Almost all dark chocolate. 2022-2024: tapered down to a trickle (1-4 items/year). 2025: stopped buying chocolate bars entirely — six purchases of Navitas raw cacao powder instead. She's still buying ~100 orders/year of other groceries, so she didn't churn off Amazon. She churned out of the category — and is now making her own at home.
From craft sampling to home-baking
U560
The lapsed-then-returned switcher
User 560 · 5-yr history · 49 confectionery items
49conf items
$4265-yr spend
4 yrssilent
Bought Hershey miniatures heavily in 2021, then bought zero chocolate online for almost four years. Came back in late 2025 with a totally different brand mix — Reese's, Kit Kat, 3 Musketeers, Milky Way. Lapsed loyalists who return with different brand preferences are a high-value cohort: their old brand has a re-acquisition opportunity, the new brand has a vulnerable foothold.
2021 Hershey → 4-yr gap → 2025 multi-brand
Method Confectionery SKUs identified via regex on product name — must match a chocolate-bar / candy / hot-cocoa / known-confectionery-brand pattern AND not match a known non-confectionery pattern (cookies, ice cream, protein bars, baking mixes, cocoa-powder for baking, cocoa-butter skincare, nutritional shakes, supplements, cereal/granola, coffee). Filter yields 665 confectionery buyers · 2,539 orders · 2,945 line items · $31K spend over Jan 2021–May 2026. Trip-type, planned/impulse, occasion, and OOS-switch are revealed-preference proxies — see per-card method notes for exact derivation. Source data: Amazon receipt export, May 2026 snapshot.
Digital Twin Simulator

MW Chocolate · Twin Responses

Predict survey responses for 500 chocolate buyers using their 5-year Amazon history. Compare twin distributions vs. what your N=504 human respondents would say.

Audience pool
665
confectionery buyers in Ario panel
Sample size
500
randomly drawn (seed=42)
Questions modeled
7
behavioral (S7, C3, C5, Q041, Q042, Q043, NQ3)
Est. runtime
~6s
behavior-conditioned, no LLM call
Pulls 5-yr basket history → derives 7 survey answers per twin → aggregates distribution.
Simulating 500 twins…
Drawing 500 chocolate buyers from the panel
Computing dominant brand per twin (C3 / C5)
Inferring trip type from basket size (Q043 / NQ3)
Mapping seasonality → occasion (Q042)
Detecting brand-switch history (Q041)
Aggregating responses

Browse individual twins

Click any twin to see their 7-question simulated response, derived from real purchases