For funds
Average ticket and basket size: how simulated consumer surveys estimate it
TL;DR
- Average ticket (check size in restaurants, basket size in retail) is price times mix. Price is usually disclosed. Mix is the part your model guesses, and it is the part a consumer survey can see.
- A Neon Blue study asks the chain's own customers what they bought last time, what they plan to buy next time, and what they would drop, add or swap at current prices.
- The output is a mix read by segment: who is adding items, who is trading to cheaper ones, and who is splitting the trip. It goes into your ticket line as a labeled input, never a stock call.
- Results come as a report. An ongoing data stream and an API are in build. The first study is free.
Neon Blue builds simulated audiences of a company's customers, AI models that answer the way a defined group of people would, grounded in Neon Blue's own paid studies and AI-moderated interviews with real people. For funds, Neon Blue turns that into answers on the consumer variables in a model, and average ticket is one of the most common.
What is average ticket, and why is it hard to model?
Average ticket is revenue divided by transactions. Restaurants call it check size or check average. Retailers call it basket size or average transaction value. It is the second half of same-store sales, with traffic as the first.
Most models build ticket as price taken plus a mix assumption. The price part is easy because management tells you. The mix part is hard, because it moves with things nobody discloses in time:
- Guests dropping an appetizer, a drink or a dessert.
- Shoppers moving from branded to private label inside the same basket, a quieter form of brand switching.
- Larger households consolidating trips into fewer, bigger baskets.
- Value menus and bundles pulling the average down while traffic holds.
Card data shows you the ticket after the fact. It can't tell you why it moved or what happens next quarter. That gap is covered in how simulated consumer data differs from card panel data.
How does a Neon Blue study estimate ticket and basket size?
- Define the audience as the comp base's customers. Recent customers of the chain, cut by visit frequency, household size and income. Heavy customers carry most of the ticket, so they get their own cell.
- Rebuild the last basket. Ask what they bought on their most recent visit, item by item or by category. This anchors the answer in a real occasion instead of a general attitude.
- Ask about the next basket at today's prices. Same items, fewer, more, or swapped for cheaper versions.
- Ask what would change it. A bundle, a value tier, a new item, a price step. Each lever is a separate question, so you can see which one moves mix.
- Weight by spend, not by head count. A heavy customer trimming one item matters more than a light customer adding one.
The read is the change from the rebuilt last basket. People overstate restraint when you ask about spending in general, so anchoring on a real recent trip keeps the answer closer to behavior. How simulated audiences work covers how the audience is grounded.
What questions go into an average ticket study?
| Question | What it informs | Why it matters |
|---|---|---|
| On your last visit, what did you buy? | Baseline basket | Anchors the read in a real trip |
| Next time, will you buy the same items, more, fewer, or cheaper versions? | Mix direction | The core ticket read |
| Which item would you drop first if you cut back? | Mix risk | Shows the most fragile part of the ticket |
| Have you started buying the store brand for anything you used to buy branded? | Private label shift | Basket size can hold while margin moves |
| Would a bundle at this price make you add an item? | Upside lever | Tests whether a deal lifts the ticket or only discounts it (see promotion sensitivity) |
| Are you making fewer, larger trips or more, smaller ones? | Trip consolidation | Separates basket size from frequency |
What does an average ticket study look like?
Hypothetical: a fast casual chain has guided to roughly flat ticket for the quarter after a small menu price increase. An analyst suspects guests are dropping drinks and sides, which would push ticket below guidance.
- Study question: "On your next visit at today's prices, will you order the same as last time, add something, drop something, or switch to a cheaper item?"
- Audience: adults who ate at the chain in the last 60 days, split into weekly, monthly and occasional guests, and by household income.
- Follow-ups: which item goes first, whether a meal bundle would bring it back, and whether they now order drinks elsewhere.
Example output layout (illustrative, not a real result)
The shape of a Neon Blue ticket read. Cells are blank on purpose; a real study fills them.
| Segment | Same order | Add an item | Drop an item | Cheaper swap | First item dropped (guests' words) |
|---|---|---|---|---|---|
| Weekly guests | __% | __% | __% | __% | Coded themes with sample verbatims |
| Monthly guests | __% | __% | __% | __% | Coded themes with sample verbatims |
| Occasional guests | __% | __% | __% | __% | Coded themes with sample verbatims |
How the analyst reads it: if weekly guests describe dropping drinks while occasional guests hold, mix is leaking from the customers who matter most to the ticket. The analyst trims the mix assumption for the heavy cell, labels the line as simulated, and checks it against the reported ticket at the print. Pair this with a price increase tolerance read if the question is whether guests leave rather than order less.
Want this run on your own name? Neon Blue's first study is free: start a free first study.
When should a ticket study run against the reporting calendar?
Run it inside the quarter and log the date, so the read is a forecast and not a story told after the numbers. The best triggers are the moments when your mix assumption is weakest: a menu or shelf price change, a new value tier or bundle, a new item (pair it with a unit demand for a launch read), a private label push, or management commentary that guests are "managing the check". Running the same questions each quarter gives a trend you can line up against reported ticket. Pre-earnings consumer checks lays out the timing.
Who uses an average ticket read?
- Consumer analysts and PMs use it to set the ticket line before the company reports. See consumer research for consumer analysts and PMs.
- PE deal teams use it to test whether a target's ticket growth is price that sticks or mix that is about to slip.
- Operators at the chain itself use it to decide which lever, bundle or price step, protects the check.
What are the limits of a survey-based ticket estimate?
- It sees what customers plan to buy, not what the kitchen or the shelf actually sold. Out-of-stocks, delivery fees and menu engineering sit outside it.
- Stated restraint runs ahead of real restraint. Read the gap between segments and the direction of change, not the absolute level.
- It works best on clear, recent occasions. Questions far from what the grounding interviews covered get flagged, and Neon Blue labels every simulated result.
- Neon Blue tests simulated answers against real responses on attitudinal and intent questions. Treat the read as one directional input, not a settled number. How to judge whether synthetic respondents are good enough to use explains what to check.
Questions about estimating average ticket and basket size
Is average ticket the same as basket size?
Mostly. Restaurants say ticket or check size, retailers say basket size or average transaction value. All of them mean revenue per transaction, so the same study design works with the wording changed.
Why not just use card data for ticket?
Card data shows what the ticket was. A Neon Blue study shows why it moved and which customers plan to change next quarter, which is the part a forecast needs. Many funds use both.
Can the study separate price from mix?
Yes. Hold the price constant in the question ("at today's prices") and ask only about what goes in the basket. Any change you see is mix.
Does this predict a company's reported ticket?
No. It gives a direction and relative size by segment for one line of your model. Neon Blue doesn't make stock calls or give price targets.
How fast can a ticket read come back?
Well before the next earnings call. A simulated audience doesn't need fieldwork, so the study can run inside the quarter rather than after it.
How do I start a free first study?
Bring the chain or retailer, the quarter, and the ticket assumption you trust least. Neon Blue's first study is free. Start a free first study