How to increase restaurant sales

1Why four levers, in this order?

A restaurant’s sales break into three numbers: sales = guests × visits per guest × average bill. All three are capped by how many people you can serve at your busiest hour. That gives four levers, and they cost very different amounts to pull.

Raising the average bill costs a menu reprint and some training. Bringing guests back costs a message to people who already like you. Serving more people at peak costs process and, sometimes, a screen in the kitchen. Buying new guests costs money every month, and it leaks through every gap the first three levers left open.

Many owners start at the bottom of this list, with ads and discounts. This guide runs the other way: cheapest rupee first.

The order matters for a plain reason. A returning guest costs you the price of a message. A new guest found through a delivery app arrives after the app’s commission, often after a discount as well, and may never come back.

2Lever 1: how do you raise the average bill?

Five moves, and none of them needs a new customer.

1. Engineer the menu. Sort every dish by how often it sells and how many rupees each plate leaves after ingredients. Promote the dishes that do both, reprice the popular ones that earn thinly, move the profitable ones nobody orders to a better place on the card, and cut what does neither. The method, with a worked 20-dish example, is in the menu engineering guide; the menu price calculator does the repricing arithmetic. Cost: an evening with an item sales report and your recipe costs. First result: the next menu print.

2. Price combos on margin, not as a discount. A combo should earn more per order than the main dish would alone. Take an example main at ₹229 with ₹72 of ingredients, and a drink at ₹79 with ₹14. Sold as a ₹279 combo, the order leaves ₹193 after ingredients, against ₹157 for the main on its own: ₹36 more from a guest who would have ordered only the main. From a guest who would have bought both anyway, the combo earns ₹29 less, which is the discount. So aim combos at the orders that arrive without a drink or a side, and check the arithmetic against your food cost before you print. Cost: close to nothing. First result: immediate, if the price is right.

3. One specific suggestion per order. A starter while the mains cook, a named dessert when the plates clear, the large size for ₹30 more. “Would you like the gulab jamun?” is easier to say yes to than “anything else?”, because it gives the guest something to agree with. Train one line per stage of the meal and count how often it lands. Cost: an hour of training a week. First result: the next day’s close.

4. Price endings and anchors. Prices ending in 9 on value dishes, round numbers on premium ones, and one deliberately high-priced dish that makes the rest of its column look fair. These are habits borrowed from retail, not laws. Test them on your own menu before you trust anyone’s rule, this guide’s included. Cost: nothing. First result: small, and immediate.

5. Add-ons one tap away. Extra cheese, the large size, a second roti: set them up as modifiers on the till, so the cashier adds one with a tap instead of typing a note. CountStand’s till carries variants, modifiers, combos and coupons on every plan, and an add-on stays on the bill line of the dish it belongs to. Cost: setting up the modifiers once. First result: every order where a guest says yes.

What does a small lift add up to? Take an example outlet with 120 bills a day at a ₹450 average. If one table in four adds a ₹120 dessert, that is 900 desserts in a 30-day month: ₹1,08,000 more in sales, and an average bill of ₹480 instead of ₹450, without one new guest through the door.

3Lever 2: how do you bring guests back more often?

A guest who comes every month instead of every quarter is worth several new ones, and you already paid once to win them.

6. Ask for the number at billing, with consent. Every bill is a chance to ask for a phone number, as long as you say what it is for and keep that promise. Ask once, say how you will use it (a birthday offer, for example), and stop the moment a guest asks you to. On every plan, from Single at ₹2,999 a month, CountStand keeps customer records with consent, encrypts the personal data, and erases it two years after a guest’s last order. Cost: one line at the counter. First result: indirect, but every later move in this lever draws on the list.

7. Win back the regulars who stopped coming. A guest who has not visited in 30 or 60 days needs a specific, small reason to return: a dish they ordered before, a new item, a weekday offer. The message goes to someone who already likes your food, which is why it costs so much less than finding a stranger. Promotional SMS in India go out through a telecom operator’s DLT registration, so register your sender and templates before you plan a campaign, and honour every opt-out. Marketing messages from CountStand are not open at launch; use a registered SMS provider. Cost: message fees and the offer. First result: two to six weeks, as regulars cycle back.

8. Ask every guest for a review. Ask for feedback after the bill, and ask every guest, not only the happy ones: Google’s rules for reviews forbid picking who gets asked. Give unhappy guests a direct way to reach you as well, so you hear the problem first. Reviews feed local search at no media cost. Cost: close to nothing. First result: slow, and it lasts.

9. Birthdays and festivals. The calendar is full of reasons to eat out: birthdays, anniversaries, and every festival with a food tradition. A small, genuine offer to a named list is also easier to measure than a hoarding. Gift cards, store credit and coupons are on every plan, so the offer and its redemption live on the same till. Cost: message fees and the offer itself. First result: per campaign, and you can count it.

10. A direct line for repeat orders. Give regulars a way to order that skips the delivery apps: a phone number that someone answers, and a pickup at the counter billed as takeaway. No platform takes a share of that bill. Here is the arithmetic for an example kitchen doing 30 app orders a day at ₹500 each: ₹4,50,000 a month through the apps. Move a fifth of those orders to the phone and ₹90,000 a month stops passing through the apps; at an effective deduction of 30%, about ₹27,000 a month stays with you, before the cost of packing those orders yourself. Run your own numbers in the aggregator commission calculator, and read the commission guide for the rest of the playbook. CountStand takes dine-in and takeaway orders; delivery-app orders are not open at launch.

4Lever 3: how do you serve more guests at peak?

You cannot sell a table you cannot turn. At peak, a full restaurant loses sales quietly: in a kitchen that loses track of ticket order, and on a floor where finished tables wait. This lever matters if you turn guests away or make them wait at your busiest hour. If your problem is empty tables rather than a queue, skim it and spend the effort on levers 2 and 4.

11. A kitchen screen at rush. Paper tickets and shouting hold up until the rush, then fail. A kitchen display that shows every ticket’s age makes the slow dish visible in minutes instead of in a review. CountStand’s kitchen display turns a ticket amber at 4 minutes and puts a red “late” ring on it at 8, with one press for each stage. Cost: a screen, and the software to drive it. First result: weeks, measured as ticket time at peak, before and after.

12. Time the turn, then remove dead minutes. Know your real turn time at peak before you try to improve it. Menus go down as guests sit, starters fire fast, and the bill arrives when the plates clear, not when the guest waves. This is brisk hospitality, not rushing. On CountStand’s floor every table shows its state, from running and in kitchen to bill printed and paid, so a table that has paid but not been cleared is visible from the counter. Cost: measurement and floor training. First result: every slot reclaimed at peak is one more cover.

13. Take the order sooner. Every minute a seated guest waits to order is a minute of table time spent on nothing. A captain who takes the order at the table saves the walk to the counter; CountStand’s captain view puts the same floor and the same bill on a phone. Pre-orders by phone for pickup take the wait out of takeaway as well. Cost: low. First result: weeks, clearest at your busiest hour.

5Lever 4: when should you pay for new guests?

Last, on purpose. Paid demand poured into a leaky operation buys one visit from guests a weak repeat engine never sees again. Pull this lever once levers 1 to 3 work, and it compounds instead of draining away.

14. Delivery-app ads. They can pay back in specific windows: a new outlet building its first ratings, a new neighbourhood, a kitchen with idle afternoon hours. As an always-on habit they are easy to overspend on, because the ad cost sits on top of the commission. Measure each burst as extra orders against what your listing brought in without it; many owners never separate the two, which is how the habit forms. The fee stack is broken down in the commission guide. Cost: high, and on top of the commission.

15. Local food creators. Creators whose followers live inside your delivery radius cost a meal and a modest fee, and their audience can actually reach you, which a celebrity’s mostly cannot. Give each creator their own code or link so every post can be counted, and expect results to vary widely. Cost: barter to a small fee per creator.

16. New hours and events. Demand you make rather than buy: a breakfast menu, a late window, corporate lunches, a weekend live counter. You are spreading the same rent and kitchen over more selling hours. A new daypart either covers its own staff and food cost within a few weeks or it does not; decide on the numbers, not the buzz.

17. ONDC as one more channel. On ONDC, the open network, the commission depends on the buyer and seller apps involved rather than on one platform, and you list through a seller app. Order volumes are far smaller than on the two big apps, so treat it as extra demand, not a replacement. The economics are in the ONDC guide.

6What does not work?

These take most marketing budgets, so they are worth naming:

• Deep discounts for everyone. They teach guests to wait for the offer, attract the least loyal, and reset your price anchor lower. A week at 40% off is a spike in sales and a hole in margin, with no repeat visit attached.
• Ads without a repeat engine. Buying the same strangers every month because nothing captured them on their first visit. Fix moves 6 to 10 first.
• Anything you cannot measure. A campaign without a code, a link or a before-and-after number is spending, not marketing.
• Copying the neighbour. The moves are the same everywhere; the weighting is not. A biryani counter lives on add-ons and speed. A family restaurant lives on repeat visits and turns.
• Growing over leaks. More sales on top of unmanaged food cost and theft grow revenue, not profit. Fix the bucket before you pay for more water.

7A 90-day plan

Days 1–30: the free lever. Engineer your top 20 dishes, price two combos on real margins, train one suggestion line per stage, and start asking for numbers at billing.

Days 31–60: the repeat engine. The first win-back message to guests who have not visited in 30 days, a review request for every guest, the birthday and festival calendar set, and your direct phone number printed on every bill and every box.

Days 61–90: speed, then fuel. Ticket times at peak measured and pushed down, the turn timed, and then the first paid tests: one burst of app ads with the extra orders counted, two or three creators with their own codes, and an ONDC listing.

The weighting changes with the format. A counter-service outlet leans on moves 3, 5, 11 and 13: suggestions, add-ons and speed. A family restaurant leans on 1, 8 and 12: the menu, reviews and turns. A delivery kitchen leans on 10, 14 and 17, where the cost of each channel matters most.

8Where CountStand fits

Every move above needs a number you can read without an export. CountStand’s Owner Cockpit shows net sales, orders, covers and the average bill from the first day, and the reports cover item performance, an hourly heatmap, the payment mix, and discounts and voids. Every plan includes variants, modifiers, combos and coupons, gift cards and store credit, customer records with consent, stock alerts, a menu health report every Monday, and morning and evening briefs by email.

Not open at launch: delivery-app orders, marketing messages and online ordering. Single is ₹2,999 a month per outlet, or ₹2,549 a month paid yearly, and starts with 30 days free, no card, by invitation. See the restaurant POS, compare the plans, or book a demo.

9Questions owners ask