# Shopper Insight Lab: virtual shelf for shopper research

Shopper Insight Lab is a [Cassi.ai](https://www.cassiai.com) product for shopper research. It is a virtual shelf purchase simulator: the participant shops a virtual aisle as they would in a real supermarket, with a cart, a budget and three missions, and the platform observes the behavior in place of asking about it. The researcher reads the result in a dashboard with an attention heat map, four attention and choice quadrants, substitutions, and additions versus removals per product.

## Shelf decisions come every month. Observed shopper behavior rarely comes with them.

Planogram, new pack, price change, out-of-stock. Each one changes what happens in front of the shelf, and most are decided with what shoppers declared in a survey.

- Topic: Shelf and pack decisions. Where a product sits in the planogram, how a new pack reads next to its neighbors, what happens when the price moves or the favorite is missing. Trade and category teams decide these all year.
- How it is done today: A declared survey, or a mock store. The survey shows the packs and asks which one the person would buy. The observed route is a physical mock store or an in-store test, which takes weeks, needs a location and works with a small sample.
- The pain: What shoppers say and what they do at the shelf differ. In a survey, people answer what sounds right. At the shelf they grab the usual brand, trade down when money is short and walk past what they never noticed. The method that observes this is too heavy for everyday decisions, so those get made on declared data.
- Where Cassi.ai comes in: A virtual shelf that observes the purchase. Participants shop your category in a browser, with a cart, a budget and three missions. The platform logs attention, additions, removals and the final choice, and the dashboard reads them by attention, choice and trade-offs, the swaps people make when the budget tightens or a product is missing.

## Shelf questions, answered with behavior.

- Attention heat map ("Does the new pack get noticed next to its neighbors?"): A heat map over the shelf, from the top shelf to the bottom, built from mouse movement and the time spent on each product. Attention share by product shows how much of the shelf time each item earns.
- Attention × Choice ("It draws attention. Does it get bought?"): Every product lands in one of four quadrants: Champions, Attracts but does not convert, Brand loyalists and Invisible. A pack that is seen and left on the shelf is a different problem from a pack nobody saw.
- Reduced budget ("What do they give up when money is shorter?"): In the second mission the budget is cut and the participant adjusts the cart. You see what leaves first, what is protected, and how the average ticket and the items per cart move from one mission to the next.
- Substitutions ("Where does my buyer go when I am out of stock?"): In the third mission the favorite product is unavailable. The dashboard lists the most frequent substitutions, product by product, so you know which neighbor inherits the purchase.
- Additions vs removals ("Which products go into the cart and come back out?"): Every addition and every removal is logged. A product that is often added and often removed is being considered and losing at the last step, something a table of final choices would hide.
- Post-purchase questionnaire ("Why did they choose that one?"): Only after shopping does the participant answer: reason for the choice, how hard it was to decide, importance of price, alternatives considered, satisfaction and influence of the budget. The declared answer is read next to the observed behavior.

## Six steps, from the shelf setup to the dashboard.

1. Shelf setup. The Cassi.ai team configures the study with you: category, products, prices, planogram and missions. Your shelf is built for your study and is never reused for another client.
2. Consent and avatar. The participant opens a link, accepts the consent terms and picks an avatar. Sessions run in Portuguese, English or Spanish, and a desktop is recommended. Sessions are anonymous. Nobody is identified.
3. Mission brief. Each mission states the task and the budget before the shelf opens: free choice, reduced budget, out of stock. Every participant goes through the three missions in the same sequence.
4. Shopping at the shelf. The participant browses, adds and removes products, watches the budget and checks out. The platform logs mouse movement, time on each product and decision time. Attention here means mouse and time. It is never called eye tracking.
5. Post-purchase questionnaire. Reason for the choice, difficulty, importance of price, alternatives considered, satisfaction, budget influence, recommendation and an open comment. The questions come after the behavior, so they do not steer it.
6. Insights dashboard. KPIs per mission and three tabs, Attention, Choice and Trade-offs, with the heat map, the four quadrants and the substitutions. Every reading shows the number of sessions behind it.

## Rules of the simulator, stated with their scope.

- 3: missions in sequence: free choice, reduced budget and out of stock
- 5 to 8 min: estimated per session, from consent to the last question
- 4: quadrants crossing attention and choice
- 3: native languages: Portuguese, English and Spanish

## What it never does

- Never calls mouse attention eye tracking. Attention is measured by mouse movement and time on each product.
- Never identifies the participant. Sessions are anonymous.
- Never presents behavior at a simulated shelf as a sales forecast.
- Never reports a quadrant without the number of sessions behind it.
- Never reuses one client's shelf for another client.

## Who it is for

Shopper marketing and trade teams, Category management, Insights teams, Research firms, Packaging and brand teams. For shelf decisions that come too often for a mock store and matter too much to settle with a declared survey.

## Questions and answers

### What is Shopper Insight Lab?

Shopper Insight Lab is a Cassi.ai product for shopper research. It is a virtual shelf purchase simulator: the participant shops a virtual aisle as they would in a real supermarket, with a cart and a budget, and the platform observes the behavior in place of asking about it. The researcher reads the result in an insights dashboard.

### How does a session work for the participant?

The participant accepts the consent terms, chooses an avatar, reads the mission brief, shops at the shelf with a cart and a budget, answers a post-purchase questionnaire and finishes. A session is estimated at 5 to 8 minutes, a desktop is recommended, and the flow is native in Portuguese, English and Spanish.

### What are the three missions?

Free choice asks the participant to buy snacks for the week within the budget. Reduced budget cuts the budget and asks them to adjust their choices. Out of stock makes their favorite product unavailable and asks them to find an alternative. The three missions run in sequence, so the same person is observed under three conditions.

### How is attention measured? Is it eye tracking?

Attention is measured from mouse movement over the shelf and from the time spent on each product, together with decision time. It is a behavioral indicator recorded in the browser. The platform never calls it eye tracking, and the dashboard labels it as mouse and time based attention.

### Does the simulator predict sales?

The simulator shows comparative behavior at a controlled shelf: which product earns attention, which one is chosen, what is swapped when the budget tightens or a product is missing. It supports shelf and pack decisions, and it never presents simulated behavior as a sales forecast.

### Who configures the shelf for a study?

The Cassi.ai team configures each study together with the client: the category, the products, the prices, the planogram and the missions. A shelf belongs to the study it was built for and is never reused for another client.

### What does the researcher dashboard show?

KPIs for participants, average ticket per mission, items per cart, average time at the shelf and the share of participants who went over budget. Three tabs, Attention, Choice and Trade-offs, bring the heat map over the shelf, attention share by product, the four attention and choice quadrants, budget influence, the most frequent substitutions and additions versus removals per product.

### Are participants identified?

Sessions are anonymous. The platform records behavior at the shelf, such as mouse movement, time on each product, products added and removed and the final choice, plus the answers to the post-purchase questionnaire. It never identifies the participant.

## Bring the shelf decision you have on the table.

Tell us the category and the decision: a new pack, a planogram, a price move, an out-of-stock. We show how the study would run in the simulator, from the mission brief to the dashboard. https://shopperlab.cassiai.com/#acesso