PURE/RAPID / INSIGHTS · RESTAURANTS & FRANCHISES
Your busiest restaurant may be the worst place to put the next £1,000.
Multi-location restaurant marketing often begins with a group budget and ends with a group dashboard. Neither tells you which kitchen, postcode or trading period can turn more advertising into more money left in the business. The better rule is to invest where additional profitable orders can be won and actually served, not where last month’s sales looked strongest.
A full Friday restaurant cannot sell the same table twice. A quieter Tuesday can be full of opportunity. But a spare table has no value if nobody wants it, and an extra order has little value if the cost of winning it consumes the margin.
One restaurant group. Several completely different businesses.
Picture a five-site group. The restaurants share a name and menu, but one relies on lunchtime office workers, another sells family takeaway orders, and a third is already turning away Saturday bookings. Give them the same campaign and you are buying different outcomes with the same money.
That matters more when consumers are selective. A UKHospitality report drawing on Meaningful Vision’s tracking of more than 60,000 UK foodservice outlets reported a 2.3% year-on-year fall in customer traffic in Q1 2026. It is sector-level evidence, not a forecast for your restaurants. Still, it makes the case for understanding where demand exists rather than treating every site as an average.
Equal budget splits feel fair. Allocating by existing revenue feels rational. Neither answers the question that matters: what can the next £1,000 buy?
The location with the best historical return is not necessarily the location with the best next return.
Existing sales also reflect location, local loyalty, opening hours, pricing and demand you may never have paid to create. The next marketing decision needs to start with a specific service period and a specific additional order, not an impressive group total.
Three locations. The same spend. Only one passes the test.
Imagine a UK restaurant group considering one extra £1,000 advertising test over four weeks. It can put that money into Harbour, Market or Station. Each option also requires £150 in genuinely additional creative and campaign costs. This is an incremental test above the network’s existing budget, not a recommended monthly account minimum. The comparison is not about which site looks popular. It is about how much contribution an extra batch of orders could leave.
Illustrative example, not a PureRapid client result. Sales exclude output VAT; variable costs exclude recoverable input VAT and include non-recoverable VAT where applicable. Contribution per additional completed order is the net sale after any discount, food, packaging, payment, marketplace commission, variable labour, royalties and other costs that order creates. No cost is counted twice. Fixed rent, existing salaried payroll, group overhead and corporation tax are not included.
| Four-week forecast | Harbour / busiest | Market / underused | Station / high volume |
|---|---|---|---|
| Possible additional orders, before capacity limit | 90 | 115 | 180 |
| Genuine spare capacity in targeted slots | 40 | 130 | 110 |
| Contribution per extra completed order | £18 | £16 | £10 |
| Orders the site could fulfil | 40 | 115 | 110 |
| Contribution before added marketing | £720 | £1,840 | £1,100 |
| Media plus incremental execution | −£1,150 | −£1,150 | −£1,150 |
| Contribution after test costs | −£430 | +£690 | −£50 |
| Break-even additional orders | 64 | 72 | 115 |
FOUR WEEKS / THREE OPTIONS / £1,150 ALL-IN INCREMENTAL COST PER TEST
Market wins this hypothetical decision with £690 left before fixed overhead and tax. Harbour cannot break even in those slots even if the demand forecast proves accurate. Station has plenty of potential orders, but too little contribution per order and not enough spare capacity to reach its break-even count.
The uncomfortable bit: the market-demand forecasts are still assumptions. A forecast is a reason to run a measured test, not a promise to spend the full budget. Nor does this prove that shifting orders to another branch creates group-wide growth. If Market steals orders from Harbour, the network gain is smaller.
VAT and cost basis: Use your actual taxable sales mix. HMRC explains why catering and hot takeaway supplies commonly carry standard-rate VAT, while treatment of some other food differs. Recoverable input VAT on marketing and supplies is not a permanent cost in this example. Confirm the treatment with your accountant before using it in a live P&L.
Can the kitchen physically earn back the marketing cost?
There is a useful threshold most restaurant dashboards never show. Divide the total additional marketing investment by the contribution from one extra completed order. Round up to the next whole order. That is the number of genuinely additional orders required before the test covers its own cost.
Then compare that count with the orders the site can actually fulfil in the targeted periods. If its spare capacity is lower, this particular spend-and-shift plan cannot break even under the stated assumptions.
For Station, £1,150 ÷ £10 = 115 orders. It has room for only 110. No better advert can fix that constraint in the same shifts. A revised offer, menu mix, operating hours or fulfilment plan might. Or the smarter use of the money could be somewhere else.
Be precise about spare capacity. A dining room with empty tables at 4pm may have no more room for diners at 7pm. A delivery kitchen with room for 60 orders on Monday might already be at safe dispatch capacity on Friday. More orders that cause late deliveries, refunds and poor reviews can reduce, rather than improve, contribution.
There is an exception. Some businesses can flex kitchen shifts or paid staff at little additional cost. Others can improve contribution with collection, group bookings or higher-value menu mixes. Recalculate the variable costs and capacity after the operational change; do not assume yesterday’s margin survives a busier service.
An advert can claim an order without creating one.
A customer sees a restaurant advert, searches for the restaurant by name, opens its Google profile and orders through a delivery app. The campaign may earn reporting credit, the Google profile may record an interaction and the delivery app records the sale. Yet the customer might have ordered anyway.
That does not make advertising pointless. It makes the measurement question more specific: what changed in completed orders and contribution because we spent the extra money? Compare point-of-sale totals, first-party ordering, bookings and delivery-platform settlements by site and service period. Do not add every platform’s credited conversions together and call it extra business.
For mixed networks, the extra order could be a transfer from the restaurant two streets away. Compare the affected catchment as well as the promoted site. Nearby branches competing for the same Friday customer are not two separate growth opportunities merely because they have separate ad accounts.
Online bookings and direct orders are easier to reconcile than many in-person visits, but they are still not a perfect experiment. Google says location assets can show shop information and that store-visit conversions depend on account eligibility. Do not promise store-visit attribution if the account does not support it.
A campaign report answers who received credit. A commercial test asks whether the network has more contribution to show for the spend.
Give the next £1,000 a job and a stopping rule.
Do not divide the test equally between all sites just to keep everyone happy. Choose one credible opportunity or a sufficiently powered, deliberately designed comparison. The right setup depends on how much data you have and whether sites share customers.
- Choose the service slot, not only the branch. Specify the restaurant, relevant postcode or catchment, weekdays and trading hours. Remove slots that cannot safely take orders or covers.
- Rebuild the real order economics. For the meals likely to sell, record revenue excluding VAT, discounts, food, consumables, payment, delivery/platform costs, incremental staff and applicable royalties. Base the break-even count on the resulting marginal contribution.
- Record the baseline before spending. Use several comparable weeks where available: completed orders or covers, average net spend, contribution, cancellations and delivery times. Note weather, holidays, promotions, local events and pricing changes.
- Run a controlled increase with a comparison. A suitably matched untreated time period or location can help, but only if spillover and seasonality are considered. For larger networks, a properly designed geo or holdout test is stronger. Small samples may only justify directional learning.
- Decide in advance what would change your mind. Stop or change the offer if additional completed orders, service standards or contribution fall below the threshold. Extend only when results are positive, repeatable and the site still has headroom.
Use a four-week window for the illustration, not as a universal test duration. Some restaurants need more data to separate marketing from weather and seasonality. For a new opening with no established baseline, treat the first campaigns as demand and operations learning, with tighter spending limits. Pretending every opening has a reliable incremental forecast is worse than admitting uncertainty.
Sometimes the right place for £1,000 is not an ad account.
Imagine putting paid demand behind the right location, only to send people to a generic home page. They must find their branch, work out whether it delivers to them, locate the menu and choose an ordering provider. Some will simply go elsewhere.
Before buying more attention, check the path from nearby intent to completed sale. Google’s restaurant Business Profile guidance covers details such as opening hours, booking options and food ordering. Its food-ordering documentation also explains custom links and preferred providers, where supported. Each restaurant needs the correct address, service hours and action. One wrong link can waste good local demand.
Then look at regular customers. A credible reason to order again can be cheaper than repeatedly reacquiring familiar diners, but only if the incentive preserves contribution and your message is permitted. The ICO’s PECR guidance explains when email or SMS marketing to existing customers may qualify under the soft opt-in and why opt-out opportunities matter.
Correct menu, branch, ordering or booking links before paying for more visitors.
Protect kitchen capacity, availability and delivery reliability before adding orders.
Set a contribution threshold and assess truly additional business.
If the broader acquisition economics need work, PureRapid’s advertising profitability guide explains the distinction between reported ROAS, contribution and what the business actually keeps.
Put your multi-location restaurant marketing budget to the test.
Change the forecast orders, viable spare capacity and contribution per additional completed order. Use the same four-week time window and a like-for-like definition of costs across the sites. The result is a planning comparison, not an estimate of advertising’s proven causal effect.
Where can the next budget produce contribution?
Market has the strongest modelled contribution after the additional £1,150 test cost. This is not a causal forecast or a substitute for a controlled trial.
The working example above is readable without JavaScript. To change figures and recalculate inside WordPress, the separately supplied PureRapid Location Budget Calculator plugin must be installed and active.
Formula: forecast completed extra orders = the lower of expected incremental orders and viable capacity. Estimated remaining contribution = completed extra orders × contribution per order − media − other additional marketing costs. Break-even orders = round up(total additional marketing costs ÷ contribution per order). Zero contribution per order cannot fund positive marketing cost.
The calculator does not account for uncertain demand, customer transfers between branches, future repeat visits, the effect of changing prices, fixed overhead, tax or cash-payment timing. If the assumptions differ by location, make them explicit rather than using the most flattering set for each site. Inputs are calculated locally in the browser when the plugin is active; no customer data is needed.
Put the money where the next sale still matters.
The decision is not always to spend more. Sometimes it is to correct a restaurant’s ordering route, protect service quality, adjust a low-margin offer or collect better baseline data. A group that is full at the wrong times does not have a media problem that extra volume will solve.
For a credible location-level budget decision, I would put three things on one page before the meeting: break-even additional orders, viable headroom in the targeted slots, and the evidence that demand would really be additional to the group. Then choose a test small enough to learn from and meaningful enough to change the next allocation.
That is the difference between distributing a marketing budget and managing an investment.
Several locations. One question worth answering properly.
If your restaurants share a marketing budget but not the same demand, margin or capacity, PureRapid can help connect local acquisition, ordering, retention and performance reporting into a coherent network decision. Start with the sites and the figures, not a generic campaign proposal.
Discuss your restaurant network ↗Explore the franchise and multi-location approach if you want to understand how the wider work fits together.
Sources, method and limitations
Research checked 16 October 2026. External operational guidance: UKHospitality, Q1 2026 foodservice traffic; HMRC, VAT rates and food service; Google Business Profile, restaurant guidance; Google Ads, location assets; ICO, electronic marketing. Examples, budget choices, forecasts and diagnostic rules are PureRapid editorial analysis, not empirical claims about client outcomes or the whole sector.