How Much Footfall Makes a Food Stall Viable? MOMA's Waterloo Station Calculation
Short answer: Count eligible passers-by in the hours you will trade, then multiply that number by a tested purchase rate and contribution per sale. A pitch is viable only when expected daily contribution covers pitch fees, paid labour, waste and a realistic wage for you, with, as a working guide, a 25% traffic or conversion downside still survivable. Do not sign from headline station footfall, because most people counted by the landlord may never pass your position, want your product or have time to stop.
Footfall looks like demand because it is visible. It is only opportunity.
The stall still has to interrupt the right person, complete the transaction before their train and leave enough money after food and labour. A busy corridor can therefore produce a weaker business than a quieter pitch used by the same customers every weekday.
Use the Footfall-to-Wage Viability Equation
The Footfall-to-Wage Viability Equation converts movement into owner income through six consecutive measures. If one measure is assumed rather than observed, the final profit figure inherits the guess.
| Measure | Definition | How to obtain it | |---|---|---| | Passing footfall | People crossing the stall's sight line | Count in 15-minute blocks | | Eligible share | People matching the buying occasion | Observe direction, timing and behaviour | | Stop rate | Eligible people who approach | Trial the offer and position | | Purchase rate | Stoppers who pay | Count completed transactions | | Contribution per sale | Price less variable food, pack and payment cost | Cost the actual serving | | Daily operating burden | Pitch, labour, travel, preparation and waste | Record cash and time by day |
The core calculation is:
daily operating surplus = passing footfall x eligible share x stop rate x purchase rate x contribution per sale - daily operating burden
My position is that you should not take a permanent food pitch until, as a working rule, at least six representative trading days cover a proper owner wage under the downside case. One successful Saturday or a landlord's annual traffic number is not enough.
How MOMA turned commuters into a calculation
Tom Mercer wanted to sell breakfast to commuters rather than wait for them to discover a new grocery product. In a detailed founder account of MOMA's first stall, he describes counting the people passing possible station sites and using a simple estimate: about 10,000 passers-by, 2% buying and roughly £2 per purchase, producing £400 of daily revenue.
Mercer eventually secured permission to trade at Waterloo East. He says he prepared products from around 2.15am, loaded the van by 6am and sold between about 6.30am and 10.30am. MOMA's company history records the early railway-arch kitchen and converted filing-cabinet stall. Contemporary local reporting later described the business using small station trolleys rather than immediately taking permanent retail units.
The calculation did not prove profit. It established an order-of-magnitude test worth running. The more useful decision came next: could real transactions validate the assumed 2% before Mercer committed to a larger operation?
Treat the figures as Mercer's retrospective account, not a universal station benchmark. Your conversion can be far lower when passengers are rushing, the stall is poorly visible or an established competitor sits before you in the flow.
Count the traffic you can actually serve
Stand where the counter will be, not at the station entrance. Count people who cross the stall's visible frontage in each 15-minute interval. Record their direction because inbound morning commuters and outbound visitors can represent different occasions.
Repeat the count on at least three ordinary weekdays and one apparently weak day. Exclude events, strikes and holidays unless they are part of normal trade. Obtain permission before conducting activity on private property and follow any site rules.
Capacity sets a ceiling. If you can complete one order every 35 seconds, a single serving position can theoretically process about 103 orders an hour. In reality, payment delays, questions and replenishment reduce that number. Observe the longest queue customers will tolerate and record abandoned approaches.
Separate eligible traffic from raw traffic
Define the occasion narrowly. A breakfast stall needs people passing before they have eaten, with enough time to stop and a way to carry the item. Someone walking past at 3pm may count as footfall but not breakfast demand.
Run a low-cost intercept test only where the property owner permits it. Offer the actual product at its intended price, not free samples alone. Free tasting can diagnose flavour, but it inflates stopping and says little about willingness to pay.
Track four numbers separately: passers-by, people who look or slow, people who approach and completed sales. If many look but few approach, visibility, queue risk or the offer is weak. If people approach but do not buy, investigate price, choice and product explanation. If they buy once but weekday repeat remains low, the location may depend on constant new traffic.
Cost every opening hour
A four-hour selling window can require nine hours of work. Include purchasing, preparation, loading, travel, site access, cleaning, cash reconciliation and unsold-food handling.
Give your own time an hourly value even if you do not draw that cash immediately. Otherwise a pitch can appear profitable only because the founder works before dawn for nothing. Separate variable serving costs from daily costs and monthly overhead.
| Cost type | Examples | Decision use | |---|---|---| | Per sale | Ingredients, cup, spoon, card fee | Contribution per transaction | | Per day | Pitch charge, assistant shift, travel, ice | Daily break-even | | Per month | Insurance, storage, accounting, permits | Required trading-day contribution | | Step cost | Second worker, larger kitchen, another trolley | Capacity expansion threshold |
Food registration, licences, allergen duties, labelling and site permissions vary by country, council, product and premises. In the UK, check current national and local requirements and obtain qualified advice for your circumstances.
Worked example: Early Line Porridge
Early Line Porridge is considering a four-hour weekday pitch beside a Birmingham commuter route. The owner counts an average of 3,200 passers-by during the proposed window. Observation suggests 55% are travelling towards offices and fit the breakfast occasion.
A six-day paid test produces these results:
| Measure | Calculation | Result | |---|---:|---:| | Eligible passers-by | 3,200 x 55% | 1,760 | | People who approach | 1,760 x 7% | 123 | | Completed purchases | 123 x 62% | 76 | | Average selling price | Observed average | £4.80 | | Variable cost per sale | Food £1.18 + pack £0.32 + fee £0.12 | £1.62 | | Contribution per sale | £4.80 - £1.62 | £3.18 | | Daily sales contribution | 76 x £3.18 | £241.68 |
Daily pitch, travel and waste cost £74. Preparation and selling require nine owner hours valued at £18, or £162. Daily operating burden is therefore £236. Expected surplus is only £241.68 - £236 = £5.68.
Revenue looks respectable at 76 x £4.80 = £364.80, but the owner has not built a safe business. Under a 25% fall in eligible footfall, purchases decline to 57 if conversion holds. Contribution becomes 57 x £3.18 = £181.26, creating a daily loss of £54.74.
The stall needs £236 divided by £3.18 = 74.2, rounded up to 75 sales, merely to cover the operating burden. For a £50 daily buffer, it needs £286 divided by £3.18 = 90 sales. The owner should improve average contribution, reduce preparation time or find a cheaper pitch before signing a permanent agreement.
These figures are illustrative. Use your actual tax treatment, waste, labour and site costs.
Compare the pitch with the next-best location
Do not evaluate a site in isolation. Run the same count and paid test at two alternatives where possible. A site with 30% less traffic may win because customers have more time, rent is lower and repeat purchases are stronger.
Compare contribution per trading hour, not just daily revenue. Also record weather exposure, storage access, water, electricity, delivery restrictions, security and the earliest permitted setup time. A pitch that requires a two-hour round trip consumes capacity that a closer site could use for preparation or a second service period.
Landlords may quote total site footfall using a different counting zone or period. Ask how it was measured, obtain the exact location and count your own sight line. Include rent reviews, service charges, required opening days and break clauses in the downside case before signing.
Make the decision within 21 days
In the first three days, define the occasion and count the exact sight line in 15-minute blocks. During the next week, repeat counts across ordinary strong and weak days and cost one serving accurately.
Use the second week for a permitted paid test. Record approaches, sales, queues, waste and every hour worked. In week three, compare at least one alternative, calculate base and 25% downside cases, and read the proposed site agreement with qualified property and legal help. Proceed only if the downside covers all cash costs and the base case pays your target wage.
Related guides
Frequently asked questions
What is a good conversion rate for a food stall?
There is no reliable universal rate. Conversion changes with the counting point, product, occasion, price, queue, weather and nearby competition. MOMA's founder describes modelling 2% of roughly 10,000 passers-by, but that was one early calculation, not a benchmark. Establish your rate through paid trading at the actual position. Count passers-by and purchases using the same time blocks, then separate people who approach from those who pay. A lower conversion can still work when contribution per sale is high and rent is low. Borrowed percentages are useful only for a first sensitivity model.
How many days should I test a stall before signing?
Test at least six representative trading days, including the weekday you expect to be weakest. More days are needed when weather, tourism, school terms or events materially affect traffic. A short test should cover complete preparation, travel, selling and cleaning rather than a two-hour promotional appearance. Pay the normal pitch cost and charge the intended price where possible. The purpose is to observe an operating range, not create a launch crowd. A seasonal location may require tests in two periods or a contract with a break clause because one week's evidence cannot represent the year.
Should I count people myself or use the landlord's figures?
Use both, but make your own decision from a count at the stall's exact sight line. Landlord figures may measure an entrance, whole centre or annual average that includes hours you will not trade. Ask for the counting method, dates and direction of flow. Your manual count can use 15-minute blocks and a simple tally, with a second person checking busy periods. Follow property rules and do not record personal information unnecessarily. The number you need is eligible traffic passing your visible frontage while you can serve, not everyone entering the wider site.
How should I include my own labour?
Value every hour required to operate, including shopping, preparation, loading, travel, setup, selling, cleaning and administration. Use at least the wage you would need the business to pay you sustainably, plus employer costs when comparing a future employee. Record cash drawings separately from economic labour cost. You may choose to take less cash during a short test, but hiding the value makes a labour-intensive stall look scalable when it is not. If the business covers ingredients and rent only because you work nine unpaid hours, you have bought yourself a demanding job without a wage.
What if the pitch works only in good weather?
Price that dependence explicitly. Split test days by weather and calculate contribution for each group. Use local historical weather only to estimate frequency, not to replace trading evidence. Then model the number of weak days the cash reserve can support. Covered sites, pre-orders, delivery to nearby offices or a seasonal contract may reduce exposure, but each adds cost and operational work. Do not average one exceptional sunny day with one wet day and assume the result is stable. If bad weather turns contribution negative, require a break clause or treat the stall as seasonal.
When should I hire a second person?
Hire when observed lost contribution from queues consistently exceeds the fully loaded shift cost and a second serving position can remove the bottleneck. Count abandoned approaches during each 15-minute period. If 18 customers leave at peak time and each would contribute £3, the maximum lost contribution is £54, not automatically 18 guaranteed sales. Test a second person for several shifts and compare incremental completed orders with wages, payroll costs and training. Employment, working-time, pension and food-handling requirements vary, so obtain qualified local advice before engaging staff.
Can free samples prove the location works?
No. Samples can test visibility, stopping behaviour and flavour, but free demand exaggerates conversion and removes the price decision. Use them to diagnose the product, then run a paid test at the intended price and portion. Record how many tasters buy immediately and return on another day. If site rules permit only sampling, collect no forecast from the crowd size alone. Secure a short paid-trading licence or pop-up period before taking a long commitment. Compliance, allergen communication and sampling permissions still apply even when no money changes hands.
Comments (0)