A quarterly business review gives franchisors and franchisees a dedicated chance to look at how a business is performing. Sales, orders and average check are a natural starting point, but those numbers don't necessarily show what's changed in the business or how a store compares with the rest of the system.
A few of the brands we work with at Bikky have shared how they've rebuilt their QBR programs around better guest, transaction, and feedback data, and how franchisees responded to having a more complete, detailed view of their business performance. We were curious to take a closer look at what those QBRs included, from the metrics they use to how they compare performance across the system and the data needed to calculate them.
This playbook starts with what can be measured using POS data alone, then looks at what a data platform like Bikky can add. At the end, we’ve included an AI prompt that creates a location-level scorecard from POS exports, giving you a starting point for building more structured, data-driven location reviews.
A QBR usually starts with the topline. That might include sales for the quarter, order count and average check, each compared with the same quarter last year. From there, franchisees may see a store-by-store breakdown, monthly sales trends or a look at how a recent promotion performed.
These numbers give franchisees a clear view of what happened during the quarter. They come straight from the POS, so they're also metrics most brands already have available for every store.
The limitation is that the topline tells you what happened, but not necessarily why. A store might be up 4% in sales, for example, but that number alone doesn’t show whether the growth came from new guests, regulars visiting more often or lapsed guests coming back. These metrics go a step further, helping franchisees understand if there’s an operational issue that leads to lower guest frequency, or highlighting how a specific LTO or promotion impacted lapsed guests. =
Metrics to include in a data-driven QBR
The most productive QBRs we’ve heard about from customers go beyond the topline to give franchisees a clearer picture of what’s driving performance. They start with sales and orders, then look at channels and the menu, and get more specific about guests, including whether they come back, how often they visit and where they are in their lifecycle.
The metrics below follow the same path, grouped by the data they require. For each one, the table shows what it tells you about the business. The first seven can be calculated from POS data alone. The remaining four require customer data that ties orders back to individual guests across channels, such as the data available through a CDP.
With POS data
Metric
What it shows
Same-store sales
Whether the business is growing or shrinking compared with the same period last year, using only stores open in both periods
Same-store transactions
Traffic on its own, helping separate a traffic problem from a check-size problem
Average order value, indexed to the system
How check size compares with the system average, with 100 representing the system average
Annualized AUV
Provides size context so stores or franchisee groups of different sizes can be read side by side
Channel mix and growth
Share of sales and orders from in-restaurant, owned digital and third-party, along with year-over-year change
Add-on attach rates by channel
How often orders include a beverage or dessert, split between owned and third-party
Menu mix and price vs. the system
How key items or categories are selling and how prices compare with the system
With guest-level data
Metric
What it shows
Guest retention at 30, 60 and 90 days
Whether guests return after their first visit, and how many are still coming back up to three months later
Lifecycle mix
The share of guests who are new, retained or lapsed, along with their share of revenue and orders
Product and LTO penetration, new vs. returning guests
Who the core menu or an LTO reaches and whether it is bringing in new guests or engaging existing ones
Guest frequency
How often guests visit and whether that frequency is rising or falling
Spend and visit frequency by demographic
Who a store’s guests are by age and income band, and how that mix compares with the system’s
Age and household income mix
Which age or income groups bring in the most revenue, and whether they spend more per visit or simply visit more often
Benchmarking against the system
Every metric above means more next to a benchmark. It gives the brand and the franchisee a shared starting point to plan from. A few rules keep the comparison fair:
Compare against the rest of the system without the franchisee's own stores. Otherwise a large group is partly compared with itself.
Use per-store numbers when comparing franchisees with different numbers of locations
Use same-store comparisons and exclude stores that opened or closed during the period
Add a peer group when it provides useful context, such as region, store age or format
Show where the franchisee is in line with or ahead of the system, not just where there is room to improve
When possible, identify a top-performing store or franchisee as a concrete example to learn from
The guest-level layer
Guest-level data helps explain what's happening behind the POS numbers. Instead of just seeing that sales or orders changed, brands can see what happened with the guests behind those transactions. That's what turns "sales were flat" into a clear picture of what a franchisee can work on next quarter.
A few examples:
Separate a new-guest problem from a retention problem. Two stores with flat sales can be in very different spots: one keeping its regulars but bringing in fewer new guests, the other drawing first-timers who don’t come back or seeing regulars visit less often.
See what's behind a channel shift. Delivery growth can mean new guests, or existing guests moving over from dine-in and the brand's own app. The sales look the same either way; the margin doesn't.
Know who an offer reached. An LTO that sold well may have brought in new guests, or mostly sold to regulars who would have come in anyway.
See who a store’s guests are. A store’s age and income mix can look quite different from the system’s, which helps decide which offers to test and how to position the menu locally.
Understand where growth is coming from. A sales gain driven by new guests, regulars spending more, or lapsed guests coming back each says something different about how long it will last.
Know how many guests are first-timers. When that share is high, how a store welcomes guests and asks them back matters more than almost anything else.
Why data matters, in both directions
For franchisors, looking across stores can reveal patterns that are harder to see at the individual location level. A brand might see that a pricing change is working in some markets but not others, that certain menu items are performing better across the system, or that a marketing initiative is driving different results by location. Those patterns can help inform what the brand tests, recommends or shares more broadly.
For franchisees, the QBR becomes more than a recap of the quarter. It gives them a clearer view of what is happening in their stores, where the change is coming from and how their performance compares with the rest of the system. That makes it easier to leave the conversation with specific areas to work on, whether that's a particular store, channel, menu item or guest cohort.
What a data-driven QBR looks like in practice
Blaze Pizza used to run its QBRs on sales, order volume and average check. "These would be listening sessions," said Chad Sims, Sr. Director, Omnichannel & Loyalty Marketing. "We'd say, 'Hey, what do you think is hurting in your marketplace?'" Blaze rebuilt those reviews into standardized, benchmarked decks for each franchisee. Franchisees who had never seen lifecycle or demographic data started tracking targets and asking sharper questions.
One finding that stood out: nearly half of Blaze's guests were first-time visitors. "A lot of franchisees thought that can't be right," Chad said. "But knowing that so many guests are first-time, we agreed that we have to introduce them to the brand on that first visit, make a strong first impression and earn the next visit back."
At The Great Greek Mediterranean Grill, a system comparison helped the brand and a franchisee get aligned on pricing. The franchisee had priced the gyro above the recommended price, betting that the item could carry a higher price. Instead, gyro sales at that store fell year over year while the rest of the system grew.
Marc Manopla, Director of Operations, used Bikky data to show where the store’s pricing and performance sat against the system, including items priced in line with the system that were growing. Once the price came back in line, gyro sales went up. "Everyone wants the business to succeed, but sometimes there are different opinions on how to get there," Marc said. "Using data to tie performance to our recommendations is the best way to get everyone on the same page."
Build your own location-level scorecard
A system-wide benchmark requires detailed exports from every store, which can be difficult and time-consuming to pull from a POS, so we built a prompt that turns your POS data into a one-page scorecard for a single location, comparing a time period you choose with the same period last year. It’s a simpler starting point and offers a structured look at a location’s performance.
An important limitation to keep in mind when using a general-purpose AI tool with a raw POS export is that it cannot account for your restaurant’s unique transaction and menu structure. A modifier like “add jalapeño” may show up as a separate item, for example, even though it’s really an add-on to another menu item. The same can happen with combos, modifiers and other restaurant-specific details. Closely review any output against your own reports to make sure the AI is interpreting those details correctly.
For Bikky users: The full version of this scorecard, benchmarked against your system, is already available in Bikky. It requires no exports or uploads and stays up to date as new data comes in. From the Dashboards tab, select New Dashboard → Location Level Scorecard. Reach out to your CSM with any questions.
For non-Bikky users: Paste the prompt into the AI tool you use and follow the steps. You’ll upload two POS exports for the location, each covering the period you’re reviewing and the same period last year: sales and order count by channel, and item sales. Be sure to check the results against your own reports.
The scorecard covers sales, orders, average order value, channel mix, add-ons and menu mix and price, each compared with last year, so you can see at a glance what’s growing, what’s steady and where there’s room to grow. It’s an easy way to start looking at your data in a more structured way ahead of your next franchisee conversation.
The prompt also only needs store-level totals, so there’s no need to upload guest names, emails, or payment details. If your company has guidelines on using AI tools with business data, consult them first before continuing.
Disclaimer: This is an AI prompt, and AI can make mistakes. Review numbers and assumptions against your own data before sharing.
Get the location-level scorecard AI prompt
Upload two standard POS exports and get a one-page scorecard for any location: sales, orders, channel mix, add-ons and menu pricing, each compared with the same period last year.
You are helping me build a one-page scorecard for one restaurant location, using only POS data…The scorecard compares a period I choose with the same period last year, so the location’s team…Step 1: Ask me three questions in one message, then wait for my answers…
Prompt unlocked
Copy this into Claude, ChatGPT or a similar AI assistant. It will ask you a few questions and for two POS exports from one location, then build the scorecard.
You are helping me build a one-page scorecard for one restaurant location, using only POS data. The scorecard compares a period I choose with the same period last year, so the location's team can see at a glance what's growing and where there's room to grow.
The tone is supportive and strategic. The location's team is a partner, not someone being graded. Use status labels like "Growing," "Steady" and "Room to grow." Never use language like "underperforming," "failing," "compliance" or "debate."
STEP 1: ASK ME THREE QUESTIONS
Ask these in one message, then wait for my answers. Don't ask anything else yet. You'll work out the rest from the data and confirm it with me in Step 3.
1. What's the brand, and which location is this for?
2. What period should it cover (a week, a month, a quarter or any date range), and do you use a calendar or a fiscal year? I'll compare it with the same period last year.
3. Do online ordering and third-party delivery orders flow into your POS, or do they live in separate systems?
If I don't know an answer, suggest a sensible default and tell me what it is.
STEP 2: ASK ME TO UPLOAD MY POS EXPORTS
After I answer, ask me to upload these two exports for this location as CSV or Excel files, covering both the period I chose and the same period last year:
- Sales and order count by order channel or source (dine-in, drive-thru, app, DoorDash and so on). Daily is best, so you can spot days the location was closed.
- Item sales: item name, menu category, quantity sold, sales and average price
Tell me in one line that if either export has a different name in my POS, I can send the closest match. If an export is missing, tell me which metrics it affects before you continue.
Also tell me that the exports can be order-level or summary reports. For example, in Toast, I can export the Order Details and Item Selection Details reports. Some POS reports leave out the order channel (dine-in, takeout, online, delivery and so on) by default, so remind me to check that it's included before I export.
Let me know that detailed exports, like item-level data, can take a few minutes to download, especially for longer periods. If a file is too large to upload, I can zip it or choose a shorter period.
If a connector or data warehouse with my POS data is already set up in this chat, tell me you'll pull the data from there instead, and skip the upload.
STEP 3: CHECK THE DATA AND CONFIRM YOUR ASSUMPTIONS
Before you exclude anything (catering, gift cards, employee meals), check whether it's actually included in the sales totals. Compare the item sales total with the total sales for the same period. Only subtract something from sales if it's part of the sales total. If it only appears in the item report, leave it out of the menu section and don't subtract it from sales. If the data doesn't let you identify something I asked you to exclude (employee meals, for example), include it and say so in the method notes.
Once you have the data, look through it and send me one message with your assumptions below. Work out each one from the data, then ask me to confirm or correct them. Don't start calculating until I reply.
- Files and dates: each file you received, what it contains and the exact date range it covers. Check the dates in the data itself, not the file name. If a file doesn't cover both periods, say so here.
- Comparing the two periods: whether the location was closed, remodeled or relocated for any part of either period, and whether it opened less than 12 months before last year's period started. If any of these apply, say how it affects the comparison. Also say whether the two periods have a different number of each weekday, or a holiday that falls in one period but not the other, since that can move a short period's numbers by a point or two.
- Channel groups: how you've mapped each order source label into in-restaurant (dine-in, drive-thru, counter, to-go, phone, kiosk), owned digital (the brand's app and website) and third-party (DoorDash, Uber Eats, Grubhub and so on). Call out any label you weren't sure about.
- Exclusions: catering, gift cards and employee meals left out, and sales net of discounts. Say which of these you found in the data and which you couldn't identify.
- Add-ons: the menu categories that work as add-ons at this restaurant, meaning items guests usually buy alongside a main item rather than on their own. Suggest one or two add-on groups that fit this menu, based on the item data, and ask me to confirm them. If some add-ons are sold as modifiers on another item rather than as items of their own, they won't appear in the item export, so say so. If a category is only sold in some channels, ask whether to include it.
- Main items: which menu categories you'll treat as main items, and the top six by units sold this period. Ask whether there's any other item I want to watch.
STEP 4: CALCULATE THE SIX METRICS
Calculate each metric for this period and the same period last year. Apply the same exclusions to every metric and both periods.
1. Sales, year over year: % change in sales vs. the same period last year
2. Orders, year over year: % change in order count. Shows traffic on its own
3. Average order value: sales ÷ orders, with the dollar and % change vs. last year
4. Channel mix and growth: each channel group's share of sales and of orders, and its year-over-year sales change
5. Add-ons per 100 orders: for each add-on group confirmed in Step 3, units sold ÷ orders × 100, for both periods
6. Menu mix and price: for each menu item shown, its share of sales, its change in units sold, and its average price this year and last year, with the % price change
Status labels: Give each of the first three metrics a label based on the change vs. last year:
- Growing: up more than 2%
- Steady: within 2% either way
- Room to grow: down more than 2%
In the menu section, flag any item whose price rose 5% or more while its units sold fell more than 5%, by showing the words "Price up, units down" next to it.
AOV note: A higher AOV isn't always good news. If AOV is up vs. last year, work out how much of the increase comes from price:
Price-only AOV change = last year's AOV × (this year's item sales ÷ the same items' units this year valued at last year's average prices − 1)
AOV growth is "mainly price" if the price-only change is more than half of the total AOV increase.
Add a short note to the AOV tile only if AOV growth is mainly price AND orders are down. For example: "AOV growth is mostly from price; orders are down 3.1%." Otherwise, show no AOV note. Keep the note in the same supportive tone as the rest of the scorecard.
Accuracy checks: Do these before building the scorecard.
- Channel sales add up to total sales, and item sales add up to total sales, each within 1% after exclusions. If they don't, find out why before moving on.
- Both periods cover the same number of days.
- AOV × orders ≈ sales.
- If a number looks surprising (sales or orders up or down more than 20%, an item's price change far from the others), check it against the raw data and ask me whether something explains it.
Never estimate or invent a number. If a metric can't be calculated from the data, leave it out and say why in the footer.
Use only POS data. Don't add anything about guest behavior, such as return visits, frequency, new vs. returning guests or demographics, since POS data can't show it.
STEP 5: BUILD THE SCORECARD
One page, laid out like an operations dashboard: summary first, detail below. Every number appears next to last year's number.
Header: brand, location name, the period and the period it's compared with.
Row 1: three summary tiles
- Sales: the period's sales, last year's sales, the % change and a status label
- Orders: the order count, last year's count, the % change and a status label
- Average order value: the AOV, last year's AOV, the dollar change and a status label, plus the AOV note only where the rule above calls for it
Row 2: two panels side by side
- Channel mix and growth: a table with each channel group's sales mix and order mix this year and last year, and its sales growth
- Add-ons: a bar for each add-on group per 100 orders, named after the group, each with a marker showing last year's number
Row 3: menu mix and price (full width)
A table of the menu items, with columns for share of sales, change in units, average price this year, average price last year and % price change
Footer
Method notes: the two periods, any weekday or holiday differences, channel groupings, add-on categories, exclusions and anything that couldn't be calculated
End the footer with this exact sentence, with no label in front of it: Generated with AI from POS data. Review for accuracy before sharing.
Design: Build the scorecard as a single web page (HTML) that prints cleanly on one page, so it can be saved as a PDF from the browser. Keep it clean and easy to scan. Use green for "Growing," a neutral gray for "Steady," and a soft amber (not red) for "Room to grow." The page should work on a phone as well as a laptop.
STEP 6: HAND IT BACK
After the scorecard, always end with both of the lists below, even if one is empty. If there's nothing to report in a list, write "None." Take every number and status label in them from the finished scorecard, and check them against the scorecard before you send.
- Worth double-checking: anything that looked unusual, any number you checked against the raw data, and any assumption you made
- Couldn't calculate: anything you left out, and what data would let you add it next time
No items found.
A playbook for data-driven franchisee QBRs
Posted
November 2, 2026
Riley Knowles
Introduction
A quarterly business review gives franchisors and franchisees a dedicated chance to look at how a business is performing. Sales, orders and average check are a natural starting point, but those numbers don't necessarily show what's changed in the business or how a store compares with the rest of the system.
A few of the brands we work with at Bikky have shared how they've rebuilt their QBR programs around better guest, transaction, and feedback data, and how franchisees responded to having a more complete, detailed view of their business performance. We were curious to take a closer look at what those QBRs included, from the metrics they use to how they compare performance across the system and the data needed to calculate them.
This playbook starts with what can be measured using POS data alone, then looks at what a data platform like Bikky can add. At the end, we’ve included an AI prompt that creates a location-level scorecard from POS exports, giving you a starting point for building more structured, data-driven location reviews.
A QBR usually starts with the topline. That might include sales for the quarter, order count and average check, each compared with the same quarter last year. From there, franchisees may see a store-by-store breakdown, monthly sales trends or a look at how a recent promotion performed.
These numbers give franchisees a clear view of what happened during the quarter. They come straight from the POS, so they're also metrics most brands already have available for every store.
The limitation is that the topline tells you what happened, but not necessarily why. A store might be up 4% in sales, for example, but that number alone doesn’t show whether the growth came from new guests, regulars visiting more often or lapsed guests coming back. These metrics go a step further, helping franchisees understand if there’s an operational issue that leads to lower guest frequency, or highlighting how a specific LTO or promotion impacted lapsed guests. =
Metrics to include in a data-driven QBR
The most productive QBRs we’ve heard about from customers go beyond the topline to give franchisees a clearer picture of what’s driving performance. They start with sales and orders, then look at channels and the menu, and get more specific about guests, including whether they come back, how often they visit and where they are in their lifecycle.
The metrics below follow the same path, grouped by the data they require. For each one, the table shows what it tells you about the business. The first seven can be calculated from POS data alone. The remaining four require customer data that ties orders back to individual guests across channels, such as the data available through a CDP.
With POS data
Metric
What it shows
Same-store sales
Whether the business is growing or shrinking compared with the same period last year, using only stores open in both periods
Same-store transactions
Traffic on its own, helping separate a traffic problem from a check-size problem
Average order value, indexed to the system
How check size compares with the system average, with 100 representing the system average
Annualized AUV
Provides size context so stores or franchisee groups of different sizes can be read side by side
Channel mix and growth
Share of sales and orders from in-restaurant, owned digital and third-party, along with year-over-year change
Add-on attach rates by channel
How often orders include a beverage or dessert, split between owned and third-party
Menu mix and price vs. the system
How key items or categories are selling and how prices compare with the system
With guest-level data
Metric
What it shows
Guest retention at 30, 60 and 90 days
Whether guests return after their first visit, and how many are still coming back up to three months later
Lifecycle mix
The share of guests who are new, retained or lapsed, along with their share of revenue and orders
Product and LTO penetration, new vs. returning guests
Who the core menu or an LTO reaches and whether it is bringing in new guests or engaging existing ones
Guest frequency
How often guests visit and whether that frequency is rising or falling
Spend and visit frequency by demographic
Who a store’s guests are by age and income band, and how that mix compares with the system’s
Age and household income mix
Which age or income groups bring in the most revenue, and whether they spend more per visit or simply visit more often
Benchmarking against the system
Every metric above means more next to a benchmark. It gives the brand and the franchisee a shared starting point to plan from. A few rules keep the comparison fair:
Compare against the rest of the system without the franchisee's own stores. Otherwise a large group is partly compared with itself.
Use per-store numbers when comparing franchisees with different numbers of locations
Use same-store comparisons and exclude stores that opened or closed during the period
Add a peer group when it provides useful context, such as region, store age or format
Show where the franchisee is in line with or ahead of the system, not just where there is room to improve
When possible, identify a top-performing store or franchisee as a concrete example to learn from
The guest-level layer
Guest-level data helps explain what's happening behind the POS numbers. Instead of just seeing that sales or orders changed, brands can see what happened with the guests behind those transactions. That's what turns "sales were flat" into a clear picture of what a franchisee can work on next quarter.
A few examples:
Separate a new-guest problem from a retention problem. Two stores with flat sales can be in very different spots: one keeping its regulars but bringing in fewer new guests, the other drawing first-timers who don’t come back or seeing regulars visit less often.
See what's behind a channel shift. Delivery growth can mean new guests, or existing guests moving over from dine-in and the brand's own app. The sales look the same either way; the margin doesn't.
Know who an offer reached. An LTO that sold well may have brought in new guests, or mostly sold to regulars who would have come in anyway.
See who a store’s guests are. A store’s age and income mix can look quite different from the system’s, which helps decide which offers to test and how to position the menu locally.
Understand where growth is coming from. A sales gain driven by new guests, regulars spending more, or lapsed guests coming back each says something different about how long it will last.
Know how many guests are first-timers. When that share is high, how a store welcomes guests and asks them back matters more than almost anything else.
Why data matters, in both directions
For franchisors, looking across stores can reveal patterns that are harder to see at the individual location level. A brand might see that a pricing change is working in some markets but not others, that certain menu items are performing better across the system, or that a marketing initiative is driving different results by location. Those patterns can help inform what the brand tests, recommends or shares more broadly.
For franchisees, the QBR becomes more than a recap of the quarter. It gives them a clearer view of what is happening in their stores, where the change is coming from and how their performance compares with the rest of the system. That makes it easier to leave the conversation with specific areas to work on, whether that's a particular store, channel, menu item or guest cohort.
What a data-driven QBR looks like in practice
Blaze Pizza used to run its QBRs on sales, order volume and average check. "These would be listening sessions," said Chad Sims, Sr. Director, Omnichannel & Loyalty Marketing. "We'd say, 'Hey, what do you think is hurting in your marketplace?'" Blaze rebuilt those reviews into standardized, benchmarked decks for each franchisee. Franchisees who had never seen lifecycle or demographic data started tracking targets and asking sharper questions.
One finding that stood out: nearly half of Blaze's guests were first-time visitors. "A lot of franchisees thought that can't be right," Chad said. "But knowing that so many guests are first-time, we agreed that we have to introduce them to the brand on that first visit, make a strong first impression and earn the next visit back."
At The Great Greek Mediterranean Grill, a system comparison helped the brand and a franchisee get aligned on pricing. The franchisee had priced the gyro above the recommended price, betting that the item could carry a higher price. Instead, gyro sales at that store fell year over year while the rest of the system grew.
Marc Manopla, Director of Operations, used Bikky data to show where the store’s pricing and performance sat against the system, including items priced in line with the system that were growing. Once the price came back in line, gyro sales went up. "Everyone wants the business to succeed, but sometimes there are different opinions on how to get there," Marc said. "Using data to tie performance to our recommendations is the best way to get everyone on the same page."
Build your own location-level scorecard
A system-wide benchmark requires detailed exports from every store, which can be difficult and time-consuming to pull from a POS, so we built a prompt that turns your POS data into a one-page scorecard for a single location, comparing a time period you choose with the same period last year. It’s a simpler starting point and offers a structured look at a location’s performance.
An important limitation to keep in mind when using a general-purpose AI tool with a raw POS export is that it cannot account for your restaurant’s unique transaction and menu structure. A modifier like “add jalapeño” may show up as a separate item, for example, even though it’s really an add-on to another menu item. The same can happen with combos, modifiers and other restaurant-specific details. Closely review any output against your own reports to make sure the AI is interpreting those details correctly.
For Bikky users: The full version of this scorecard, benchmarked against your system, is already available in Bikky. It requires no exports or uploads and stays up to date as new data comes in. From the Dashboards tab, select New Dashboard → Location Level Scorecard. Reach out to your CSM with any questions.
For non-Bikky users: Paste the prompt into the AI tool you use and follow the steps. You’ll upload two POS exports for the location, each covering the period you’re reviewing and the same period last year: sales and order count by channel, and item sales. Be sure to check the results against your own reports.
The scorecard covers sales, orders, average order value, channel mix, add-ons and menu mix and price, each compared with last year, so you can see at a glance what’s growing, what’s steady and where there’s room to grow. It’s an easy way to start looking at your data in a more structured way ahead of your next franchisee conversation.
The prompt also only needs store-level totals, so there’s no need to upload guest names, emails, or payment details. If your company has guidelines on using AI tools with business data, consult them first before continuing.
Disclaimer: This is an AI prompt, and AI can make mistakes. Review numbers and assumptions against your own data before sharing.
Get the location-level scorecard AI prompt
Upload two standard POS exports and get a one-page scorecard for any location: sales, orders, channel mix, add-ons and menu pricing, each compared with the same period last year.
You are helping me build a one-page scorecard for one restaurant location, using only POS data…The scorecard compares a period I choose with the same period last year, so the location’s team…Step 1: Ask me three questions in one message, then wait for my answers…
Prompt unlocked
Copy this into Claude, ChatGPT or a similar AI assistant. It will ask you a few questions and for two POS exports from one location, then build the scorecard.
You are helping me build a one-page scorecard for one restaurant location, using only POS data. The scorecard compares a period I choose with the same period last year, so the location's team can see at a glance what's growing and where there's room to grow.
The tone is supportive and strategic. The location's team is a partner, not someone being graded. Use status labels like "Growing," "Steady" and "Room to grow." Never use language like "underperforming," "failing," "compliance" or "debate."
STEP 1: ASK ME THREE QUESTIONS
Ask these in one message, then wait for my answers. Don't ask anything else yet. You'll work out the rest from the data and confirm it with me in Step 3.
1. What's the brand, and which location is this for?
2. What period should it cover (a week, a month, a quarter or any date range), and do you use a calendar or a fiscal year? I'll compare it with the same period last year.
3. Do online ordering and third-party delivery orders flow into your POS, or do they live in separate systems?
If I don't know an answer, suggest a sensible default and tell me what it is.
STEP 2: ASK ME TO UPLOAD MY POS EXPORTS
After I answer, ask me to upload these two exports for this location as CSV or Excel files, covering both the period I chose and the same period last year:
- Sales and order count by order channel or source (dine-in, drive-thru, app, DoorDash and so on). Daily is best, so you can spot days the location was closed.
- Item sales: item name, menu category, quantity sold, sales and average price
Tell me in one line that if either export has a different name in my POS, I can send the closest match. If an export is missing, tell me which metrics it affects before you continue.
Also tell me that the exports can be order-level or summary reports. For example, in Toast, I can export the Order Details and Item Selection Details reports. Some POS reports leave out the order channel (dine-in, takeout, online, delivery and so on) by default, so remind me to check that it's included before I export.
Let me know that detailed exports, like item-level data, can take a few minutes to download, especially for longer periods. If a file is too large to upload, I can zip it or choose a shorter period.
If a connector or data warehouse with my POS data is already set up in this chat, tell me you'll pull the data from there instead, and skip the upload.
STEP 3: CHECK THE DATA AND CONFIRM YOUR ASSUMPTIONS
Before you exclude anything (catering, gift cards, employee meals), check whether it's actually included in the sales totals. Compare the item sales total with the total sales for the same period. Only subtract something from sales if it's part of the sales total. If it only appears in the item report, leave it out of the menu section and don't subtract it from sales. If the data doesn't let you identify something I asked you to exclude (employee meals, for example), include it and say so in the method notes.
Once you have the data, look through it and send me one message with your assumptions below. Work out each one from the data, then ask me to confirm or correct them. Don't start calculating until I reply.
- Files and dates: each file you received, what it contains and the exact date range it covers. Check the dates in the data itself, not the file name. If a file doesn't cover both periods, say so here.
- Comparing the two periods: whether the location was closed, remodeled or relocated for any part of either period, and whether it opened less than 12 months before last year's period started. If any of these apply, say how it affects the comparison. Also say whether the two periods have a different number of each weekday, or a holiday that falls in one period but not the other, since that can move a short period's numbers by a point or two.
- Channel groups: how you've mapped each order source label into in-restaurant (dine-in, drive-thru, counter, to-go, phone, kiosk), owned digital (the brand's app and website) and third-party (DoorDash, Uber Eats, Grubhub and so on). Call out any label you weren't sure about.
- Exclusions: catering, gift cards and employee meals left out, and sales net of discounts. Say which of these you found in the data and which you couldn't identify.
- Add-ons: the menu categories that work as add-ons at this restaurant, meaning items guests usually buy alongside a main item rather than on their own. Suggest one or two add-on groups that fit this menu, based on the item data, and ask me to confirm them. If some add-ons are sold as modifiers on another item rather than as items of their own, they won't appear in the item export, so say so. If a category is only sold in some channels, ask whether to include it.
- Main items: which menu categories you'll treat as main items, and the top six by units sold this period. Ask whether there's any other item I want to watch.
STEP 4: CALCULATE THE SIX METRICS
Calculate each metric for this period and the same period last year. Apply the same exclusions to every metric and both periods.
1. Sales, year over year: % change in sales vs. the same period last year
2. Orders, year over year: % change in order count. Shows traffic on its own
3. Average order value: sales ÷ orders, with the dollar and % change vs. last year
4. Channel mix and growth: each channel group's share of sales and of orders, and its year-over-year sales change
5. Add-ons per 100 orders: for each add-on group confirmed in Step 3, units sold ÷ orders × 100, for both periods
6. Menu mix and price: for each menu item shown, its share of sales, its change in units sold, and its average price this year and last year, with the % price change
Status labels: Give each of the first three metrics a label based on the change vs. last year:
- Growing: up more than 2%
- Steady: within 2% either way
- Room to grow: down more than 2%
In the menu section, flag any item whose price rose 5% or more while its units sold fell more than 5%, by showing the words "Price up, units down" next to it.
AOV note: A higher AOV isn't always good news. If AOV is up vs. last year, work out how much of the increase comes from price:
Price-only AOV change = last year's AOV × (this year's item sales ÷ the same items' units this year valued at last year's average prices − 1)
AOV growth is "mainly price" if the price-only change is more than half of the total AOV increase.
Add a short note to the AOV tile only if AOV growth is mainly price AND orders are down. For example: "AOV growth is mostly from price; orders are down 3.1%." Otherwise, show no AOV note. Keep the note in the same supportive tone as the rest of the scorecard.
Accuracy checks: Do these before building the scorecard.
- Channel sales add up to total sales, and item sales add up to total sales, each within 1% after exclusions. If they don't, find out why before moving on.
- Both periods cover the same number of days.
- AOV × orders ≈ sales.
- If a number looks surprising (sales or orders up or down more than 20%, an item's price change far from the others), check it against the raw data and ask me whether something explains it.
Never estimate or invent a number. If a metric can't be calculated from the data, leave it out and say why in the footer.
Use only POS data. Don't add anything about guest behavior, such as return visits, frequency, new vs. returning guests or demographics, since POS data can't show it.
STEP 5: BUILD THE SCORECARD
One page, laid out like an operations dashboard: summary first, detail below. Every number appears next to last year's number.
Header: brand, location name, the period and the period it's compared with.
Row 1: three summary tiles
- Sales: the period's sales, last year's sales, the % change and a status label
- Orders: the order count, last year's count, the % change and a status label
- Average order value: the AOV, last year's AOV, the dollar change and a status label, plus the AOV note only where the rule above calls for it
Row 2: two panels side by side
- Channel mix and growth: a table with each channel group's sales mix and order mix this year and last year, and its sales growth
- Add-ons: a bar for each add-on group per 100 orders, named after the group, each with a marker showing last year's number
Row 3: menu mix and price (full width)
A table of the menu items, with columns for share of sales, change in units, average price this year, average price last year and % price change
Footer
Method notes: the two periods, any weekday or holiday differences, channel groupings, add-on categories, exclusions and anything that couldn't be calculated
End the footer with this exact sentence, with no label in front of it: Generated with AI from POS data. Review for accuracy before sharing.
Design: Build the scorecard as a single web page (HTML) that prints cleanly on one page, so it can be saved as a PDF from the browser. Keep it clean and easy to scan. Use green for "Growing," a neutral gray for "Steady," and a soft amber (not red) for "Room to grow." The page should work on a phone as well as a laptop.
STEP 6: HAND IT BACK
After the scorecard, always end with both of the lists below, even if one is empty. If there's nothing to report in a list, write "None." Take every number and status label in them from the finished scorecard, and check them against the scorecard before you send.
- Worth double-checking: anything that looked unusual, any number you checked against the raw data, and any assumption you made
- Couldn't calculate: anything you left out, and what data would let you add it next time
