--- title: "From Idea to Execution · Irina Kostina" url: https://kostina.online/speaking/from-idea-to-execution/ language: en description: "Material by Irina Kostina: how to turn a gastronomic idea into a product that the team can execute, the guest can appreciate and the business can repeat. A case, a concept, control, one operation with a neural network and a worksheet." updated: 2026-10-16 author: Irina Kostina (Ирина Костина) contact: https://me.kostina.online alternates: - https://kostina.online/speaking/from-idea-to-execution/ - https://kostina.online/speaking/from-idea-to-execution/index.md --- > Agent version of the page: the readable text without navigation, scripts and decoration. The HTML page is the reference; quote with a link to it. Irina Kostina / Talk material # From Idea to Execution How to create gastronomic projects: creativity, technology and decisions that hold up at a real event. [Stage mode · 12 screens](https://kostina.online/speaking/from-idea-to-execution/#stage-1) [Worksheet](https://kostina.online/speaking/from-idea-to-execution/#questions) How do you turn a gastronomic idea into a product that the team can execute, the guest can appreciate and the business can repeat? - Author · Irina Kostina · Managing partner - ≈ 9 min read - October 2026 Material on the topic of the talk. It contains one management case, one concept, one training analysis of execution and one training operation with a neural network. The status of each example is given next to it: practice, concept or training example. Practice · in Irina’s words, published ## A decision that changed the model A gastronomic idea rarely starts with a dish. More often it starts with the question of how the business that will execute it is organised. Irina Kostina has worked in the catering division of a restaurant group since 2010. Until 2020 the division had its own production — a dark kitchen. That is the name for a kitchen without a dining room that works for delivery and off-site events. During the pandemic it was closed and the financial model was rebuilt: off-site projects began to be assembled on the basis of the group’s restaurants. > “…because of the pandemic we completely closed the dark kitchen, our own production, and completely rebuilt the financial model, in which we have now worked successfully with the group’s restaurants for 6 years.”Irina Kostina · article on the catering market, 4 May 2026; Field Notes No. 003 In her description, the model works like this. A request turns not into a menu but into an event project. Managers put together a single menu to suit the client’s wishes and select the restaurants that will execute it. Between two and six restaurants can go to one venue. Irina names the limitation of the model herself: with early set-ups and large volumes, bringing several restaurants together at one venue can be quite difficult, but, in her words, everything is achievable when the team works well together. Details of internal coordination are not disclosed publicly. **What follows from this** Before looking for a new idea, check which resources you already have that stand idle: kitchen, team, purchasing, equipment, trust in the name. Sometimes something new is a different way of organising the business, not a different dish. The limit of transfer: a model with several contractors requires strong coordination; without it, it multiplies risks. Concept · catalogue v1, no event is claimed ## How a concept is built Let us take one author’s concept — Noodle Lab. Its task is to show how an instant product — noodles, chips, mashed potato, snacks — becomes material for restaurant presentation. 1. **Task**what the guest should see in a new way 2. **Material**one product class, eight bases 3. **Classification**a key of four axes 4. **Connections**the distinctness rule 5. **Menu**50 objects 6. **Space**four stations and the guest route The main move here is classification. Each object in the catalogue is described by a key of four axes. Base: eight options, from ramen and soba to chips and mash. Processing method: ten techniques, such as slow braising, emulsion, dehydration. Role of the product in the dish: ten options, from base to crunchy element. And register: cold, warm, hot, sweet. Three catalogue objects and their keys | Object | Key: base · processing · role · register | | Ceviche on a chip | chips :: dehydration :: chip :: cold | | Lagman with slow-braised lamb | lagman :: slow_braising :: base :: warm | | Carbonara on ramen | ramen :: emulsion :: sauce :: warm | The key works in two directions at once. For the kitchen, it sets the technique and the role of the product. For the space, the fourth axis becomes the layout: cold, warm, hot and sweet are the four stations in the hall, with 12, 13, 17 and 8 objects. The guest receives a finished presentation with a label. The key also has a checking function. The catalogue has a rule: two objects must not match on three of the four axes, otherwise it is a repetition. The rule immediately shows weak spots. In the published version v1, 35 pairs of objects match in this way — this is noted on the project page. The working notes for the catalogue record that one of the techniques is hardly used: it should either be developed in the menu or removed from the vocabulary. **What follows from this** Describe your product through several axes: material, processing method, role, place or time of consumption. An axis that does not distinguish between options is superfluous. An axis that governs space or logistics is the most useful. The approach transfers to a coffee shop, a bakery, a shop or a service: the axes change, the principle remains. Practice · in Irina’s words and from the market article ## How an idea holds up in execution An off-site project is tested not on paper but on the day of the event. In the article on the catering market, Irina puts it this way: “catering is not about food, it is about managing chaos so that guests see only the celebration.” > “Catering is an extremely complex mechanism where the cost of a mistake off-site is too high.”Irina Kostina · on the catering course for practitioners, 2026 The same market article (article on the catering market, 2026; Field Notes No. 003) names four conditions without which an off-site line of business does not become sustainable: separate processes rather than a “tail” of the main kitchen; costed logistics — routes, transport, temperature control, set-up and dismantling; one’s own equipment base; and an operational owner — a person who is responsible for the line of business all the time, not between other tasks. ### Training analysis: an early-morning set-up, three kitchens, one venue Training example · not a description of a real event Imagine an event where three kitchens prepare one menu for one venue, and the set-up is early in the morning. Where can a mistake still be corrected before it reaches the guest? 1. Before the eventOne technical card per item and one person responsible per item, not “per kitchen”. It is decided in advance which items move to another kitchen if one is late. 2. The day beforeChecking routes and arrival windows, backup transport, temperature control in transit. 3. ArrivalOne receiving point: one person receives the products of all kitchens against a common list — quantity, temperature, allergen labelling. 4. ServiceA control first plate of each item before it goes out to guests. 5. AfterwardsA short debrief: what worked and what to change next time. **What follows from this** Assign responsibility for critical decisions in advance and by item. Write down three to five checkpoints where a mistake can still be corrected before it reaches the guest, and one backup solution for the most likely failure. Training example · no implementation is claimed ## One operation with a neural network In the study “Gastro Engineering”, Irina describes where data and neural networks can help in working with the menu and where their limits lie. Below is one operation within this framework: generating dish options under given constraints. It is a training scenario, not a description of a working system. 1. InputConstraints: one base from the catalogue, the cold station, role — crunchy element; no nuts or raw fish; a maximum cost; preparation off-site without an open flame. 2. OperationThe model proposes twelve sketches: a combination, a technique and a short description for the label. 3. Human choiceThe chef selects two or three sketches that make sense for the idea and discards the rest. 4. VerificationTasting, a technical card, an allergen and safety check by a food technologist, a trial batch in off-site conditions. 5. OutputOne or two items ready for production — or a reasoned rejection. > “The chef receives not a recipe but a field of sketches… The chef decides whether it makes sense.”“Gastro Engineering”, v3, 2026 The limits of the operation matter more than its speed. The model is not responsible for allergens, product safety, recipes or feasibility off-site — specialists check these. Its output is a hypothesis that has to be tested in practice. Any claim of economic effect requires real data, a period and a method of comparison. **What follows from this** Give the tool the breadth of options, and give people the choice and the verification. Before launch, decide who checks the result and against which criteria. The framework of this material ## Five questions before launch These questions gather the material into a worksheet. It is a framework for checking your own project, not a standalone method. 1. **For whom?**Who the guest or buyer is and in what situation they choose. 2. **What is the value?**What they get — in one sentence, without the words “best” and “unusual”. 3. **Who executes?**Who is responsible for each critical step and what happens in case of failure. 4. **What are the economics?**Cost, price, use of resources: where the project is in profit. 5. **What confirms the result?**What you will measure and when, to know that the idea works. If you apply these questions to the 2020 case, the third and fourth come to the fore: who will execute, and what the economics rest on. ### A completed example Training example A bakery in a regional city wants to launch weekend breakfasts. For whom? Families and couples from nearby neighbourhoods, Saturday and Sunday mornings, 9:00–12:00; they choose a place they can come to without a booking. What is the value? Hot bread from their own oven as the centre of breakfast: three sets where bread is the main dish. Who executes? The baker — baking ready by 8:30; a second employee — assembling the sets; the manager — dining room and till. Backup: if the oven is not ready, sets with the previous day’s bread at a reduced price. What are the economics? The cost of a set is calculated from the technical card; the price covers it with a margin; the oven, which used to stand idle on weekend mornings, is now in use. What confirms the result? After six weekends: the number of sets per morning, the share of returning guests, leftover bread. The decision to continue is based on these figures. ## Summary An idea gains value when taste, space, service, organisation of work and economics support one another. If you look at the 2020 case through these five questions, you can see that a new model can come not from a new dish but from a different way of organising execution. The next idea is worth choosing with the same question: not “what will surprise”, but what the team can execute, the guest can appreciate and the business can repeat. [Stage mode](https://kostina.online/speaking/from-idea-to-execution/#stage-1) ### Related materials - [Field Notes No. 003 · Managing Chaos](https://kostina.online/projects/field-notes/003/) - [Noodle Lab · catalogue](https://kostina.online/projects/noodle-lab/) - [Gastro Engineering](https://kostina.online/projects/gastro-engineering/) - [All projects](https://kostina.online/projects/) ### Sources 1. Catering in 2026 — a market review (with comments by Irina Kostina), industry article, 4 May 2026. The link to the partner’s publication was removed by decision of the site owner (14.10.2026). 2. Field Notes No. 003 “Managing Chaos”: [kostina.online](https://kostina.online/projects/field-notes/003/) 3. Noodle Lab, catalogue v1: [kostina.online](https://kostina.online/projects/noodle-lab/) 4. “Gastro Engineering”, v3, 2026; overview: [kostina.online](https://kostina.online/projects/gastro-engineering/)