How to test a coffee shop location by buying window
Evaluate a coffee location by the periods when suitable customers can and will buy. Count relevant movement in consistent time windows, test a supported purchase assumption and check whether the service bottleneck can deliver the resulting transactions. Footfall by itself is not a sales forecast.
What you will produce: A buying-window transaction test and an evidence record for comparing candidate locations.
Updated September 6, 2026 · Worked examples and editable worksheets
What to have ready
Record the exact frontage, observation dates, window duration, direction of travel, access and trading conditions. Keep customer counts and transactions distinct: one transaction may serve more than one person.
Work through the calculation and decision
What should you observe at each location?
Define the window before counting, such as an hour of morning commuter movement. Count people who can realistically see, access and use the offer. Record weather, school or office patterns, road crossings, queues, parking and nearby alternatives. Distinguish unique people from repeated passes where your method allows.
Repeat comparable windows across relevant weekdays and weekends. A festival day and an ordinary weekday answer different questions. Use the same observation method at candidate sites, retaining the raw counts and circumstances so the comparison can be repeated.
How do observations become a transaction scenario?
Multiply relevant passers-by by a purchase share only after stating its basis. A temporary, permitted paid test can provide more specific evidence than an opinion survey, but differences in signage, offer, price and convenience still limit transfer to the permanent shop.
The editable example uses an assumed 8% share to show the arithmetic. It is not a coffee-industry conversion rate. Run a lower share and report the consequence instead of choosing the percentage needed to justify the rent. Keep preorders, destination visits and other channels separate to avoid counting the same transaction twice.
Can the shop serve that window profitably?
Identify the slowest service resource: ordering, payment, espresso production, food handling or collection. Divide its available lane-minutes by time per transaction. Multiple lanes count only if they operate independently at the bottleneck. Test arrival bunching and queue tolerance because an average hourly ceiling can hide lost peak sales.
Calculate contribution after variable product and payment costs, then combine distinct trading windows into a complete day. Deduct the full paid roster, rent and other fixed commitments in the broader business plan. One attractive hour cannot establish monthly profit or justify a lease by itself.
300 relevant passers-by imply 24 transactions only under an 8% assumption
This authored example uses one 60-minute window, one independent service lane, two lane-minutes per transaction, a $6 ticket and $2 variable cost.
| Question | Calculation | Result |
|---|---|---|
| Assumed demand | 300 × 8% | 24 transactions |
| Service ceiling | 60 ÷ 2 | 30 whole transactions |
| Capacity-limited scenario | Lower of 24 and 30 | 24 transactions |
| Window contribution | 24 × ($6 − $2) | $96 |
| At a 4% purchase share | 12 × $4 | $48 contribution |
| At a 12% purchase share | 36 desired, capped at 30 | $120 contribution before fixed costs |
What this changes: Doubling the assumed purchase share from 8% to 16% cannot double sales through this service lane. Investigate both paid demand and service capacity before using a stronger window to support a more expensive site.
Test transactions in one buying window
Start with the illustrative example, then replace its inputs with your own assumptions. All money amounts are in USD. The result updates in this tab.
Illustrative result · assumptions apply
- Transactions implied by your capture assumption
- 24 transactions/window
- Whole-transaction service ceiling
- 30 transactions/window
- Capacity-limited transactions for this scenario
- 24 transactions/window
- Contribution in this buying window
- $96.00
Capture is an assumption until supported by a comparable paid test. This window excludes other trading periods, fixed labor, occupancy and owner income.
Complete your decision record
A buying-window transaction test and an evidence record for comparing candidate locations. Enter the finding or number, the source and the next action for each row. “Supported” records your assessment of that item; it does not approve the business or certify completed research.
| Item and what to record | Your finding and evidence | Status and next action |
|---|---|---|
| Observation windowAddress, date, start, end, weather and unusual conditions | ||
| Relevant movementCount, direction, access and counting method | ||
| Purchase evidenceOffer, actual paid test, price, sample and transfer limits | ||
| Service testBottleneck, lanes, timing, queue and lost orders | ||
| Unit economicsTicket, variable cost and contribution per window | ||
| Site comparisonComparable windows, total occupancy cost and unresolved facts |
6 items have no evidence recorded yet.
Entries are temporary and are not sent to us or saved automatically. Download your completed work before leaving or refreshing this page.
Choose your next action
| If your finding is… | Your next action |
|---|---|
| The case depends on an untested purchase percentage | Collect a comparable paid test before treating the forecast as supported demand. |
| Peak demand exceeds measured service capacity | Test a feasible service change and its costs before assuming extra transactions. |
| A site works only in an exceptional observation window | Repeat ordinary trading periods and rebuild the daily case. |
Errors that can change the result
- Treating every passer-by as an addressable buyer.
- Multiplying a peak hour by all opening hours.
- Counting baristas as independent service lanes when they share one bottleneck.
Apply this to your business
These operating formats match the decisions in this guide.
Coffee Shop
Independent coffee shop without a drive-through
Open the operating guide and state profiles →Combine tested windows with the paid roster and occupancy commitments in the operating plan. Values entered here are not automatically transferred to another calculator.
Continue with the next part of your plan
- How to test business demand before forecasting revenue
A measured customer funnel, contribution after test acquisition costs and a decision about the next test.
- What to check before committing to business premises
A comparable premises cost summary and a list of conditions to resolve before the next commitment.
- How to build a staffing roster before estimating payroll
A roster with complete task coverage and an annualized monthly staffing budget.
Sources and limits
The sources below provide the stated background. The worked examples, calculator defaults and decision exercises are authored teaching material. They do not establish market prices, local demand, legal applicability or completed state research.
- SBA: planning a business
Background on costs, demand and break-even.
Source pages checked September 6, 2026. Research and review standards · Report an issue
When you need a longer financial plan
Use a financial model to organize a broader forecast after defining your own operating assumptions. The site’s research, your worksheet entries and the purchased workbook are separate; entries are not transferred automatically.