For funds
Simulated consumer research for restaurant analysts
TL;DR
- Restaurant comps come down to guest behavior: how often people visit, what they order, and how they respond to price and deals. Most of that is invisible until the company reports.
- Neon Blue runs simulated guest studies on the questions that come up every quarter in the sector: traffic, check, value perception, deals, lower-income pullback, loyalty and daypart.
- Each study returns answers by guest segment (visit frequency, income, age, region) as a labeled input to your model, 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 chain's guests, 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. Restaurant analysts at hedge funds and investment firms use Neon Blue to answer the guest questions behind a comp model, such as visit intent, check size and response to a value deal, while the quarter is still open.
What consumer questions matter most for restaurant stocks?
The same handful of guest questions drives most restaurant models, from quick service to casual dining:
- Will traffic hold? Visits per guest, especially among heavy guests.
- Is the check growing from price or from mix? Price is disclosed. Whether guests drop a drink or a side is not.
- Do guests see the brand as good value? Value perception moves before traffic does.
- Do value deals bring new visits, or discount existing ones?
- Are lower-income guests pulling back, and where do they go? The "whether" is a consumer sentiment by cohort read. The "where" is a brand switching question.
- Is the loyalty app driving visits, or just rewarding visits that would have happened anyway? See loyalty program engagement.
Each of these maps to a variable page with the full method, linked below.
Which Neon Blue studies answer them?
| Analyst question | Study to run | Audience | What comes back |
|---|---|---|---|
| Will traffic hold next quarter? | Visit intent vs last 90 days | Recent guests by visit frequency and income | Expected change in visits by segment, and where lost visits go |
| Is check growth price or mix? | Last-order rebuild and next-order intent at today's prices | Recent guests by frequency | Add, drop or swap shares, and the first item dropped |
| Is the brand seen as good value? | Value perception vs named alternatives | Guests and lapsed guests | Value ranking against the occasion's alternatives, with reasons |
| Do value deals add visits? | Deal response | Guests split by deal use | New visits vs swapped visits, and check effect |
| Are lower-income guests pulling back? | Budget and occasion study | Guests by income band | Cut back, trade down to a cheaper chain, or cook at home |
| Is the app driving visits? | Loyalty engagement | App members vs non-members | Extra visits credited to offers vs habit |
How do you read restaurant traffic before the company reports?
Ask the chain's recent guests how many times they visited in the past four weeks, then whether they expect to visit more, the same or less in the next three months, and why. Weight by visit share, so heavy guests carry the read. The method is on the foot traffic page, and the full comp build, traffic and ticket together, is on same-store sales.
How do you tell price from mix in the check?
Hold the price constant in the question and ask only about the order. "At today's prices, will you order the same as last time, add something, drop something or switch to a cheaper item?" Any change is mix. The average ticket and basket size page has the question set and an output layout.
How do you test a price increase or a deal before it shows in comps?
For a menu price increase, ask guests what they would do at the new price: change nothing, order cheaper, visit less, or go elsewhere. Price increase tolerance covers the method. For a value deal or a bundle, ask whether it would bring an extra visit or replace a full-price one, which is the core of promotion sensitivity. A deal that only swaps visits lowers the check without adding traffic.
What does a restaurant study look like?
Hypothetical: a casual dining chain has launched a value bundle and said its lower-income guests have been visiting less. The analyst wants to know whether the bundle brings them back before the next report.
- Study question: "This bundle is a starter, two entrees and a dessert for a set price. In the next month, would it make you visit this restaurant more than you do now, about the same, or would you still go elsewhere?"
- Audience: adults who visited the chain in the last six months, split by household income and by visit frequency, plus lapsed guests who last visited six to twelve months ago.
- Follow-ups: where they went instead, and whether they would order the bundle in place of their usual order.
Example output layout (illustrative, not a real result)
The shape of a Neon Blue guest read. Cells are blank on purpose; a real study fills them.
| Segment | Visit more | About the same | Still go elsewhere | Bundle replaces usual order | Where they go instead (guests' words) |
|---|---|---|---|---|---|
| Income under $50K, lapsed | __% | __% | __% | __% | Coded themes with sample verbatims |
| Income under $50K, current | __% | __% | __% | __% | Coded themes with sample verbatims |
| Income $100K and up, current | __% | __% | __% | __% | Coded themes with sample verbatims |
How the analyst reads it: if lapsed lower-income guests say the bundle would bring them back, traffic upside is real. If current guests mostly say it would replace their usual order, the bundle lowers the check more than it lifts visits. Both lines go into the model labeled as simulated.
Want this run on your own name? Neon Blue's first study is free: start a free first study.
When should restaurant analysts run these studies?
Inside the quarter, after a menu price change, a new value platform, a limited-time item or a loyalty change, and before the quarter closes. A repeat read with the same questions each quarter gives a trend you can line up against reported comps. Pre-earnings consumer checks lays out the calendar, and consumer channel checks shows how a guest read sits next to the usual franchisee and supplier calls.
What can't a guest study see?
- Operations: staffing, speed of service, weather and calendar shifts.
- Unit growth: new openings, closures and remodels change the comp base, not guest intent.
- Fast news swings: a read reflects how guests say they'll act around a known change, not this week's headlines.
- Neon Blue tests simulated answers against real responses on attitudinal and intent questions. Treat each read as one directional input, not a settled number.
Questions about consumer research for restaurant analysts
What consumer data do restaurant analysts use?
Mostly card and foot-traffic data for what already happened, plus channel checks and expert calls. A Neon Blue study adds why guests are changing and what they plan to do next quarter.
Can a simulated study cover a specific chain's guests?
Yes. The audience is defined as that chain's recent guests, cut by visit frequency, income, age and region. You can name the chain in the study.
Can it compare two chains?
Yes. Ask guests of both, or ask one chain's guests where else they go for the same occasion. The answer shows which rival takes lost visits.
Does Neon Blue predict restaurant comps or stock moves?
No. Neon Blue answers the guest variables in your model, such as visit intent or check mix. It doesn't make stock calls or give price targets.
How do I start a free first study?
Bring the chain, the quarter and the guest question you trust least in your model. Neon Blue's first study is free. Start a free first study