AI assistant · N4 Executable plan
Marketing Budget Builder for Restaurants
Works out how much money your restaurant can put into marketing without touching its profit floor, splits it between presence, acquisition and repeat business, and gives every line a yardstick in money per visit and the date it gets switched off if it misses.
N4Executable plan
A plan with owners, deadlines and the metric you review. Not advice: an agenda someone can execute on Monday.
◆ This assistant 23 of the 164 published assistants are at this levelIs this for you?
What it does and what you get
- The question "how much do I invest in marketing?" almost always gets answered with a borrowed percentage, and that percentage knows neither your margin nor what your delivery channel charges you.
- This assistant builds the sum backwards: first the contribution margin of one visit, which is the most that can be paid for a new guest; then what the income statement can carry once prime cost is paid; and only then the split.
- It also leaves the cut-off rule written before any money moves, the part nobody writes and the only one that switches off a line that does not pay back in time.
What you end up with
Five chained pieces: the contribution margin of one visit per channel with its yardstick, the month's envelope in money and in points of sales, the split by objective with minimum visits, the weekly cut-off board and the half page that governs the envelope.
EXAMPLE This is what you getAn example built with Sazón de Origen, the house's fictional restaurant.Not your data: your result is built with YOUR restaurant's.See the full example →Sample result
The owner asked this: "Between the two locations I sell 214 million a month and I do not know how much of that I can put into marketing. This month I paid for ads, photography and a Tuesday two-for-one, and I do not know which of the three gave me anything back."
A new guest is not worth the same through the three doors. Product cost comes from the menu's average food cost, 34%; what each channel charges comes from your own receipts and is assumed here until you confirm it.
| Channel | Check | Product (34%) | What the channel charges | Contribution margin | Yardstick per new guest |
|---|---|---|---|---|---|
| Dining room | 48,000 COP | 16,320 COP | 1,200 COP (card fee, 2.5%) | 30,480 COP | 15,240 COP |
| Own delivery | 48,000 COP | 16,320 COP | 5,300 COP (packaging 1,800 + delivery 3,500) | 26,380 COP | 13,190 COP |
| Delivery platform | 48,000 COP | 16,320 COP | 13,800 COP (25% commission + packaging 1,800) | 17,880 COP | 8,940 COP |
The yardstick is half the contribution margin, not the whole of it: that way the first visit already leaves something and there is no waiting on a second one nobody has measured. Once the frequency log exists and says new guests come back, the yardstick can rise to the full margin. Not before.
What the table says. The same new guest is worth 30,480 COP through the dining room and 17,880 COP through the platform. Paying the dining-room yardstick —15,240 COP— for a guest who arrives through the platform leaves 2,640 COP on that visit, and that does not even cover washing the dishes. So this month's acquisition money points at the dining room and the platform stays in presence.
Illustrative figures built from the data you handed over. The platform commission, the packaging and the cost of your own delivery are assumptions: swap them for yours in one line and the whole table recalculates.
The marketing budget already existed: it lived inside expenses and nobody had added it up. Against the month's sales for both locations:
| Line | Money | Points of sales |
|---|---|---|
| Sales | 214,000,000 COP | 100% |
| Prime cost (34% product + 28% payroll) | 132,680,000 COP | 62% |
| Occupancy and expenses, marketing excluded | 59,100,000 COP | 27.6% |
| Today's marketing | 5,100,000 COP | 2.4% |
| Today's profit | 17,120,000 COP | 8% |
The 5,100,000 COP came from adding five lines that lived in five different places: paid social 900,000, photography and running the accounts 1,400,000, print and in-store material 280,000, comps for guests 630,000 (42 plates at 15,000 COP of product) and the Tuesday two-for-one 1,890,000 (125 giveaway plates at 15,120 COP of product each). The last two never left the bank and are 49% of the envelope.
That 2.4% was copied from no manual: it is what the business was already spending without anyone calling it a budget, and it works as a starting point, not as a target.
The envelope's ceiling is 5,100,000 COP for as long as profit stays where it is. The goal —from 8% to 14% in 12 months without raising menu prices— asks for 12,840,000 COP more profit every month, and every peso the envelope grows today comes out of exactly that. So the envelope does not grow by decision: it grows by half of whatever a line returns above its yardstick, and that sum is done on the first working day of the month.
The same 5,100,000 COP, split by what they buy and not by the channel where they get spent.
| Objective | Amount | Share of the envelope | What it buys | Yardstick per visit | Minimum visits |
|---|---|---|---|---|---|
| Presence | 1,680,000 COP | 33% | Photography, the platform listing, the printed menu and the accounts kept current | Hard cap: a third of the envelope | Not applicable |
| Repeat business | 1,920,000 COP | 38% | A WhatsApp redemption to whoever already came | 10,160 COP | 189 recovered visits |
| Acquisition | 1,500,000 COP | 29% | Paid social pointing at the dining room | 15,240 COP | 99 new visits |
The repeat-business yardstick is a third of the margin and not half: someone who already came does not need paying to discover the house, only to remember it. Presence promises no visits, so it is not asked for a number it cannot give; in exchange it carries a hard cap, and it fits under its own by 20,000 pesos.
If every line lands right on its yardstick, the envelope returns 3,678,240 COP of margin a month: 1,517,520 from acquisition and 3,840,720 from repeat business, less the 1,680,000 of presence. That is 28.6% of the 12,840,000 COP the goal asks for. The other two thirds do not come out of the envelope —they come out of prime cost and menu price—, and this assistant does not touch them.
One finding from the split. The Bandeja de la casa does not fit under a 32% food cost even at full price: 15,120 COP of product against a 42,000 COP menu price is 36%. It cannot carry the discount. The problem there is not the offer, it is the menu price.
Illustrative example generated with a fictional restaurant. Not a promise of results, not a client case.
This is 3 of 5 parts. Behind the wall sit the two pieces that get executed. The cut-off board, week by week, with the columns Acquisition spend, New visits, Cost per visit, Repeat spend, Recovered visits and Cost per visit, plus the month row, which is the one that rules: it is read on Fridays at 9 in the morning, before next week's payment is approved, and the rule says one week above the yardstick is noise while two consecutive cuts switch the line off. There are the two lines that went off on the first Friday, and why one was called blind rather than a loser. And the half page that governs the envelope: the docket box with its four closed answers, the minimum attributed visits for a reading to count and the three figures of the first working day.
See the full example — free accountWorks with these AIs
How to use it
Which of your data it uses
This assistant works with 10 of the 23 fields in «My Restaurant», among them Restaurant name, Business type and Number of locations. It does not ask out of curiosity: these are what make the answer speak about YOUR scale instead of an industry average — the same calculation on a business of another size returns a number you cannot decide with. You fill them once and they apply across the library, so the second assistant you open already starts with them in place. Whatever you leave blank, the assistant asks for one thing at a time instead of inventing it, and says so before answering, so you know what it is working from.
Always up to date
Every assistant carries a visible version and date. When an AI changes how it works, the assistant's version goes up and you see what changed.
v1.0.12026-08-26The body stops contradicting the answer ladder. It used to say «do not move on until it arrives», the very order the ladder was born to replace: the assistant asked the owner for what the owner does not have and stalled. It now says what the house doctrine says — ask one at a time, and if they do not know, continue with a marked assumption. FORMAT OF THE RESULT was not touched, so the example still corresponds and its seal moves up with the card.v1.0.02026-08-25The assistant is born.More assistants in this category
Builds the marketing plan for the whole year: which campaigns get in, in which window of the business, with ho…
N4 · Executable planTurns «why they choose me and not the place next door» into a 12-to-18-word sentence the server repeats and th…
N5 · Complete systemCounts backwards from the day something opens —a location, a menu, a channel or a new line— with the capacity …
Included in the always-growing library
Access to every published assistant, adapted to your AI and personalised with your restaurant's data, updates included.