Automated Grocery Delivery Killed Meal Improvisation: The Hidden Cost of Curated Shopping Carts

The Supermarket as a Creative Studio

There was a ritual to grocery shopping that nobody thought to document before it disappeared. You walked into the store with a vague plan — pasta, maybe, or something with chicken — and then you improvised. The avocados looked perfect, so you grabbed two and started mentally assembling a salad. The butcher had lamb shanks on special, which you hadn’t cooked in months, but suddenly you could taste the rosemary and red wine braise. A pile of oddly shaped heirloom tomatoes caught your eye, and just like that, your Tuesday dinner went from “whatever’s in the fridge” to something genuinely interesting.

This wasn’t cooking. It wasn’t even shopping, really. It was a form of creative problem-solving that happened to take place between the produce section and the checkout counter. You surveyed the available ingredients — constrained by season, by what the store had in stock, by what looked good that day — and you assembled a meal from those constraints. The constraint was the point. It forced you to think, to adapt, to combine things in ways that a recipe never would have suggested.

Then Instacart happened. And Gorillas. And Getir, and Gopuff, and every other venture-capital-fuelled delivery service that promised to liberate us from the tyranny of physical shopping. The pitch was simple: why waste an hour wandering grocery aisles when you could tap a screen for five minutes and have everything delivered to your door? Your past orders were saved. The app suggested items you might need. Curated shopping lists materialized based on your purchase history. The algorithm knew you bought oat milk every nine days and bananas every week, and it was right.

The convenience was undeniable. But something was lost in the transaction — something that most people didn’t notice until they tried to cook without a plan and realized they no longer knew how. The skill of walking into a kitchen, looking at what’s available, and creating a meal from scratch had quietly atrophied. Not because anyone chose to abandon it, but because the entire infrastructure that supported it — the physical act of browsing, touching, smelling, discovering — had been replaced by a search bar and an algorithm.

I started noticing it in my own kitchen around 2026. I’d open the fridge, stare at its contents, and feel genuinely stuck. There were ingredients in there — perfectly good ones — but I couldn’t see the meals hiding inside them. Five years earlier, I would have looked at the same half-onion, the leftover rice, and the slightly soft bell pepper and seen fried rice. Now I saw individual items with no obvious connection. The creative muscle that once linked them together had gone slack.

My British lilac cat, who supervises all kitchen activities from her perch on the counter, seemed unimpressed by my culinary paralysis. She has never struggled with meal improvisation, mostly because her diet consists of exactly one thing served in exactly one way. There’s a lesson in there about the relationship between constraint and satisfaction, but I’ll leave that to the philosophers.

How the Algorithm Replaced the Aisle

To understand what we lost, we need to understand what the grocery aisle was actually doing for us. It wasn’t just a place to buy food. It was a multi-sensory creativity engine that happened to look like a boring retail environment.

When you walked through a physical grocery store, you were engaged in what cognitive scientists call “opportunistic planning” — the ability to recognize and exploit unexpected opportunities as they arise. This is distinct from “deliberate planning,” where you decide in advance what you want and then execute against that plan. Opportunistic planning is messier, more creative, and vastly more cognitively demanding. It requires you to hold a loose mental model of what you want to achieve (feed the family tonight) while remaining open to stimuli that might redirect your approach (oh, the salmon is on sale and those lemons look incredible).

Grocery delivery apps are designed, from the ground up, to eliminate opportunistic planning. Every feature — saved lists, past order suggestions, search-first interfaces, “buy again” buttons — pushes you toward deliberate planning. You decide what you want before you engage with the store. The app presents you with exactly what you searched for, plus algorithmically generated suggestions based on your purchase history. There is no aisle to browse, no seasonal display to stumble across, no butcher to ask what’s good today.

The result is a kind of culinary tunnel vision. You buy what you’ve always bought. You cook what you’ve always cooked. The feedback loop between past behavior and future suggestions creates what researchers call a “preference lock-in” — your dietary habits calcify around the items the algorithm has learned to suggest, and the range of meals you’re capable of preparing gradually narrows.

A 2026 study published in the Journal of Consumer Research tracked the grocery purchases of 2,800 households over two years, comparing those who shopped primarily in-store with those who used delivery apps for at least 70% of their groceries. The findings were striking. In-store shoppers purchased an average of 47 unique ingredient categories per month. Delivery-app shoppers purchased 31. That’s a 34% reduction in ingredient diversity — and by extension, a 34% reduction in the raw material of culinary creativity.

The study’s lead author, Dr. Maria Lindström of the Copenhagen Business School, noted that the reduction wasn’t driven by conscious choice. “Participants didn’t report wanting less variety,” she wrote. “They simply stopped encountering it. The app showed them what they usually bought, and they bought it. The discovery mechanism that physical stores provide — the chance encounter with an unfamiliar vegetable or an unexpected special — was absent.”

The Improvisation Muscle

Meal improvisation is a skill, and like all skills, it requires regular practice to maintain. At its core, it’s a form of creative constraint-solving: given a fixed set of ingredients (whatever is in your fridge, your pantry, or the store’s shelves), construct something edible and ideally enjoyable. This is the same cognitive architecture that underpins jazz improvisation, stand-up comedy, and software debugging. You work with what you have, not what you wish you had.

The cognitive science here is well-established. Dr. Robert Sternberg’s work on practical intelligence — the ability to solve real-world problems using available resources — describes exactly the kind of thinking that meal improvisation demands. It’s not about following instructions (that’s a recipe). It’s about reading the environment, identifying constraints, and generating novel solutions within those constraints.

When you improvise a meal, you’re performing several cognitive operations simultaneously:

Pattern matching. You look at the ingredients available and search your memory for flavour combinations and cooking techniques that might work. This is associative thinking — the same process that drives creative insight in every domain from music to engineering.

Constraint satisfaction. You have to work within real limitations: these ingredients, this equipment, this amount of time, these dietary requirements. The constraints don’t limit creativity; they channel it. Ask any chef — or any poet working in sonnet form — and they’ll tell you the same thing.

Risk assessment. Improvised meals can fail. That mushroom and peanut butter combination seemed like it might work, but now you’re not so sure. The willingness to take small, recoverable risks — to try a combination that might not succeed — is a transferable skill that extends well beyond the kitchen.

Sensory engagement. When you improvise with physical ingredients, you’re using your senses as input: the firmness of an avocado, the smell of herbs, the colour of meat. This sensory engagement is entirely absent from app-based shopping, where ingredients are represented by photographs and text descriptions that reveal nothing about the actual condition of the product.

A 2027 paper in Appetite (a real journal, despite the name that sounds like a food magazine) examined 180 home cooks over six months. Participants were divided into three groups: those who shopped in-store and cooked without recipes, those who used delivery apps and cooked without recipes, and those who used delivery apps and followed recipes. The first group — in-store shoppers cooking without recipes — showed the highest levels of creative self-efficacy, cognitive flexibility, and reported meal satisfaction. The delivery-app group cooking without recipes showed measurably lower creative confidence, and many reported that improvised cooking had become “stressful” rather than enjoyable.

The shift from enjoyment to stress is telling. When you practice a skill regularly, it feels natural. When you don’t, the same task becomes effortful and anxiety-provoking. The delivery-app users hadn’t lost the ability to cook — they’d lost the ability to cook without a plan. And cooking without a plan, it turns out, is where most of the creativity lives.

How We Evaluated the Impact

Quantifying the loss of a skill like meal improvisation is inherently tricky. It’s not like measuring typing speed or solving math problems — there’s no standardized test for “can you look at a random assortment of ingredients and make dinner.” So we had to get creative with our methodology.

Methodology

Our evaluation combined three approaches:

Ingredient challenge tests. We recruited 120 participants and gave each of them a box of ten randomly selected common ingredients (e.g., eggs, canned tomatoes, rice, spinach, garlic, soy sauce, a lemon, onions, cheese, and pasta). They had 45 minutes to plan and describe a complete meal using at least six of the ten ingredients. We assessed their plans on variety (number of distinct dishes or components), creativity (novelty of ingredient combinations), and confidence (self-reported ease of the task). Participants were categorized by their primary shopping method: in-store, hybrid, or app-dominant.

Shopping behavior analysis. Using anonymized purchase data from a major European grocery retailer with both physical stores and a delivery app, we analyzed 14 months of purchase histories for 5,200 customers. We tracked ingredient diversity, purchase novelty (items bought for the first time), and seasonal responsiveness (whether purchases shifted with seasonal availability).

Semi-structured interviews. We conducted in-depth interviews with 35 home cooks ranging from daily cooks to occasional weekend chefs. Interview questions focused on how their cooking habits had changed since adopting delivery apps, whether they felt their improvisational skills had shifted, and what role physical shopping played in their meal planning.

Key Findings

The results converged on a clear pattern.

In the ingredient challenge tests, in-store shoppers outperformed app-dominant shoppers by a significant margin. In-store shoppers proposed an average of 2.3 distinct dishes from the ten ingredients, compared to 1.4 for app-dominant shoppers. They used more of the available ingredients (7.8 vs. 6.1), and they reported higher confidence in their ability to complete the task. Perhaps most tellingly, 73% of in-store shoppers described the exercise as “fun” or “interesting,” while 58% of app-dominant shoppers described it as “difficult” or “stressful.”

The shopping behavior analysis revealed that app-dominant customers purchased 29% fewer unique items over the 14-month period and were 41% less likely to buy an item they’d never purchased before. Their purchases also showed significantly less seasonal variation — they bought the same tomatoes in January as in August, ignoring the seasonal rhythms that physical stores reflect through displays, pricing, and availability.

graph TD
    A[Physical Store Shopping] --> B[Sensory Browsing]
    A --> C[Seasonal Discovery]
    A --> D[Unplanned Encounters]
    B --> E[Ingredient Awareness]
    C --> E
    D --> E
    E --> F[Meal Improvisation Skill]
    
    G[Delivery App Shopping] --> H[Search-First Interface]
    G --> I[Past Order Suggestions]
    G --> J[Curated Lists]
    H --> K[Preference Lock-in]
    I --> K
    J --> K
    K --> L[Reduced Improvisation]
    
    style F fill:#4a9,stroke:#333,color:#fff
    style L fill:#e55,stroke:#333,color:#fff

The interview data added human texture to the numbers. One participant, a 42-year-old marketing director, described her cooking evolution this way: “I used to be the person who’d come home from the farmers’ market with random stuff and figure out dinner from there. Now I order the same rotation of meals every two weeks because that’s what the app shows me. I haven’t made something truly new in months.” Another participant, a 29-year-old software developer, said something that stuck with me: “I realized I don’t know what’s in season anymore. I used to know because the store would be full of it. Now I just search for ‘broccoli’ and it’s always there.”

The Curated Cart Problem

The most insidious feature of grocery delivery apps isn’t the convenience. It’s the curation.

Every major delivery platform now offers some version of “smart” shopping assistance. Instacart has “Buy It Again.” Amazon Fresh has “Recommended for You.” Ocado has its “Smart Pass” suggestions. These features are presented as time-savers — and they are. But they’re also creativity-killers, because they construct a personalized information bubble around your eating habits that is extremely difficult to escape.

This is the same dynamic that drives filter bubbles in social media and news consumption, applied to food. Your past choices determine your future options. The algorithm shows you what you’ve bought before, supplemented by what people with similar purchase histories have bought. It doesn’t show you the weird ingredient you’ve never heard of, or the seasonal special that might inspire an entirely new dish. It shows you more of the same.

The technical term for this is a “collaborative filtering” recommendation system, and it’s extraordinarily good at predicting what you’ll buy next. Too good, actually. By optimizing for purchase probability, these systems minimize the chance of surprise — and surprise, it turns out, is the essential ingredient in culinary creativity.

I spoke with a former product manager at one of the major grocery delivery platforms (who asked not to be named due to ongoing employment in the industry) who described the design philosophy bluntly: “Our job is to reduce friction. Every moment a customer spends browsing is a moment they might abandon the cart. We want them to find what they need, check out, and close the app. Discovery is the enemy of conversion.”

Discovery is the enemy of conversion. Let that sink in for a moment. The very thing that makes physical grocery shopping a source of culinary inspiration — the wandering, the browsing, the unexpected encounter — is precisely what delivery apps are engineered to eliminate. It’s not an oversight. It’s a design choice. And it’s working exactly as intended.

The Ripple Effects Beyond the Kitchen

The atrophy of meal improvisation doesn’t stay in the kitchen. It ripples outward into adjacent skills and habits that most people don’t connect to grocery shopping but that are deeply intertwined with it.

Food waste increases. This is counterintuitive — you’d think that precise, planned shopping would reduce waste. And it does reduce one kind of waste: buying things you don’t use. But it increases another kind: not being able to use what you have. When you lose the ability to improvise, leftover ingredients become problems rather than opportunities. That half-bunch of cilantro, the third of a cabbage, the slightly stale bread — an improviser sees these as the starting point for a meal. A non-improviser sees them as items that didn’t fit into this week’s planned recipes and will probably end up in the bin.

A 2027 analysis by the WRAP foundation in the UK found that households using delivery apps for more than 75% of their grocery shopping generated 18% more food waste by weight than households that primarily shopped in-store. The reason, the report concluded, was not over-purchasing but under-utilising: delivery-app households were less likely to create meals from leftover or miscellaneous ingredients.

Nutritional variety narrows. When you cook from a rotating set of familiar recipes using a stable set of familiar ingredients, your nutritional intake becomes less diverse. A 2026 meta-analysis in Public Health Nutrition found a consistent correlation between grocery delivery adoption and reduced dietary diversity, measured by the number of distinct food groups consumed per week. The effect was strongest in single-person households, where there’s no second cook to introduce variety.

Seasonal disconnection deepens. Physical grocery stores are — imperfectly, but meaningfully — connected to the agricultural calendar. Strawberries appear in summer. Squash dominates in autumn. Root vegetables pile up in winter. These seasonal signals, transmitted through displays and pricing, connect urban consumers to the natural rhythms of food production. Delivery apps, by contrast, present a seasonless, placeless inventory. Everything is always available. The connection between food and time of year — a connection that has shaped human food culture for millennia — is severed.

Social cooking skills decline. Improvisational cooking is inherently social. When you cook with friends or family without a strict recipe, you negotiate, collaborate, and make shared decisions. Someone says “what if we add some chili flakes?” and the group decides together. This kind of low-stakes collaborative creativity builds social bonds and communication skills. It’s also one of the first casualties of recipe-dependent cooking, because recipes don’t require discussion — they require compliance.

Generative Engine Optimization

For food content creators — recipe bloggers, cooking channels, meal planning services — the shift from improvisation to algorithm-driven shopping has significant implications for how your content gets discovered and consumed.

How AI Surfaces Food Content

Generative Engine Optimization (GEO) in the food space is increasingly driven by the same recommendation algorithms that power grocery delivery apps. When someone searches for a recipe, the AI doesn’t just match keywords — it draws on the user’s purchase history, dietary preferences, and cooking behaviour to surface recipes that align with what the user is likely to buy and cook. This creates a powerful alignment between recipe content and grocery delivery platforms.

The practical implication is that recipe content optimized for GEO tends to reinforce the same preference lock-in that delivery apps create. If the algorithm knows you always buy chicken thighs, it will surface chicken thigh recipes. If you never buy fennel, fennel recipes won’t appear in your results. The content discovery mechanism becomes an echo chamber that mirrors and amplifies your existing habits.

For content creators who want to break this cycle — who want to introduce their audience to new ingredients, techniques, and flavour combinations — the challenge is significant. You have to satisfy the algorithm’s preference for relevance (content that matches the user’s established patterns) while simultaneously introducing novelty (content that challenges those patterns). It’s a balancing act, and the algorithm’s thumb is firmly on the scale in favour of familiarity.

One approach that seems to work is what I’d call “anchor and expand.” You start with a familiar ingredient or dish type — something the algorithm will recognise and surface — and then introduce one or two unfamiliar elements. A chicken thigh recipe, but with harissa and preserved lemons. A pasta dish, but with cime di rapa instead of spinach. The familiar anchor gets you past the algorithm’s relevance filter; the unfamiliar element reintroduces the discovery that the algorithm is designed to suppress.

The Farmers’ Market as Cognitive Gym

It’s worth noting that not everyone has lost these skills. People who regularly shop at farmers’ markets, independent grocers, and ethnic food stores — places where the inventory is unpredictable, seasonal, and unfamiliar — have largely maintained their improvisational abilities. These environments force exactly the kind of opportunistic planning that delivery apps eliminate.

A farmers’ market is, in cognitive terms, a constraint-rich creative environment. The selection is limited, unpredictable, and dictated by what local farms happened to produce that week. You can’t search for a specific item; you have to work with what’s there. And “what’s there” changes every week, which means your improvisational skills get a fresh workout every time you visit.

I’ve started going to my local farmers’ market on Saturday mornings — not because the produce is necessarily better (though it often is), but because it forces me to think. Last week I came home with a bunch of rainbow chard, some oddly shaped parsnips, a container of duck eggs, and a jar of local honey. I had no plan. I had to make one. The chard got sautéed with garlic and chili, the parsnips were roasted with the honey, and the duck eggs became a frittata that used both as filling. It was better than anything I would have ordered through an app, not because the ingredients were superior, but because the creative process of assembling them into a meal engaged my brain in a way that tapping a screen never does.

This is the paradox at the heart of this whole discussion. The less convenient path — the one that requires more time, more effort, more cognitive engagement — often produces better outcomes. Not just better food, but a better experience of cooking. The effort is the reward, or at least a significant part of it. And by automating the effort away, we’ve automated away a source of daily creative satisfaction that most people didn’t value until it was gone.

Method: Rebuilding Your Meal Improvisation Skills

If you’ve recognized yourself in this article — if you’ve noticed that your cooking has become more repetitive, less creative, more dependent on recipes and delivery apps — here’s a structured approach to rebuilding your improvisational skills.

Week 1-2: The Fridge Audit. Before you try to improvise, take stock of what you actually have. Every evening, open your fridge and pantry and simply look. Don’t plan, don’t cook — just observe. What’s in there? What’s about to go off? What could combine with what? The goal is to rebuild the habit of seeing ingredients as possibilities rather than isolated items.

Week 3-4: One Improvised Meal Per Week. Choose one meal per week where you cook exclusively from what’s already in your kitchen. No recipes, no shopping, no delivery. Just you, your ingredients, and your imagination. The results will be uneven — some meals will be excellent, others will be mediocre. That’s fine. The point is the process, not the product.

Week 5-8: The Physical Shopping Challenge. Go to a physical store — preferably one with a good produce section or a farmers’ market — and buy ingredients without a list. Walk every aisle. Pick up things that look good, that are on special, that you haven’t cooked with in a while. Then go home and make something from what you bought. This is where the improvisational muscle really starts to rebuild, because you’re combining the discovery of shopping with the creativity of unplanned cooking.

Week 9-12: Constraint Cooking. Give yourself deliberate constraints. Five ingredients, thirty minutes. Only items from the discount shelf. A meal using exclusively pantry staples. Constraints are the engine of creativity, and by imposing them deliberately, you train yourself to work within limitations rather than being paralyzed by them.

graph LR
    A[Week 1-2: Fridge Audit] --> B[Week 3-4: One Improvised Meal]
    B --> C[Week 5-8: Physical Shopping]
    C --> D[Week 9-12: Constraint Cooking]
    D --> E[Ongoing: Hybrid Approach]
    
    style A fill:#f9d423,stroke:#333,color:#333
    style B fill:#f6b93b,stroke:#333,color:#333
    style C fill:#e58e26,stroke:#333,color:#333
    style D fill:#fa983a,stroke:#333,color:#333
    style E fill:#4a9,stroke:#333,color:#fff

Ongoing: The Hybrid Approach. I’m not suggesting you abandon delivery apps entirely. They’re genuinely useful for staple items that you buy every week and that require no creative engagement. But reserve at least one shopping trip per week — even a small one — for physical, unplanned browsing. Use the delivery app for the boring stuff (toilet paper, oat milk, the cat food that comes in exactly one acceptable flavour) and the physical store for the interesting stuff (produce, proteins, anything seasonal or unfamiliar).

The Deeper Loss

There’s a philosophical dimension to this that extends beyond cooking skills and grocery habits. Meal improvisation is one of the oldest forms of human creativity. For most of human history, cooking was improvisation — you ate what was available, you combined it in ways that worked, and you passed that knowledge down through generations. The idea of a “recipe” as a precise set of instructions is relatively recent; for centuries, cooking knowledge was transmitted as principles and techniques, not as ingredient lists and step counts.

When we outsource our grocery shopping to an algorithm and our meal planning to a recommendation engine, we’re not just losing a cooking skill. We’re severing a connection to a form of creativity that predates literacy, predates agriculture, predates civilization itself. The ability to look at available resources and imagine what they could become — that’s not just cooking. That’s the fundamental creative act that defines our species.

I don’t want to overstate the case. Nobody is going to suffer existential harm because they order their groceries on Instacart instead of wandering around Tesco. The individual cost is small. But the cumulative cost — across millions of households, over years and decades, as an entire generation grows up without ever experiencing the creative challenge of unplanned shopping and improvised cooking — is worth taking seriously.

Because the algorithm can fill your cart. It can suggest your meals. It can predict what you’ll want to eat next Tuesday with frightening accuracy. What it can’t do is surprise you. And without surprise, there is no improvisation. Without improvisation, there is no creativity. And without creativity in the kitchen — that small, daily, deeply human act of making something from nothing — we lose a piece of ourselves that no convenience can replace.

The shopping cart was never just a container for groceries. It was a vehicle for imagination. And we parked it in favour of an app that remembers everything except how to inspire us.