Two different things under one heading
Meal kits and food delivery are often discussed together because both change where a meal comes from. They are different services, though. A meal kit typically sends a set of ingredients, often pre-portioned, with a recipe to be cooked at home, usually arriving on a recurring schedule. Food delivery usually means prepared food from a restaurant or shop, brought to the door through an app or by phone. Grocery delivery is a third category again, and some services blur the lines by offering prepared meals, ingredient boxes and groceries together.
When someone says that ordering food at home is growing, the first question is which of these is meant. A figure for restaurant takeaway, one for subscription boxes and one for online supermarket orders describe separate behaviours, even when a headline joins them.
Where the records come from
Changes in how people get meals at home are tracked through several kinds of evidence. Each shows part of the picture.
Company reports. Firms that run delivery apps or meal kit services sometimes publish figures on orders, customers or revenue. These come from complete internal records for that company, which is useful, but they cover only that firm, and companies choose what to publish and how to define terms such as “active customer”. Definitions may change between reports.
Payment and transaction records. Aggregated card or bank data can show spending with certain merchants or categories. They can reveal trends across many firms, but merchants have to be categorised, and a purchase through an app may appear under the platform rather than the restaurant. The strengths and limits of this kind of source are described in more detail in the general discussion of sales data.
App and download data. Counts of downloads or sessions are sometimes used as a proxy for interest. Downloads are not orders, and many apps are installed and rarely used. Platform data of this kind has well-known blind spots, including what the platform chooses to share.
Surveys. Asking people how often they order meals or use a meal kit service gives information about households and habits that business records cannot, such as who is using the service and why. But answers depend on recall and wording, and people may misremember how often they order. This is a clear case where self-reported behaviour and observed behaviour can diverge.
Official statistics. In some countries, national statistical offices track household spending, retail sales or food service output. These are typically carefully designed but may update slowly and use broad categories that do not isolate delivery.
Search trends. Search charts for terms like “meal kit” or “food delivery” are sometimes used as signals. They show attention rather than use, and they can spike for reasons unrelated to ordering. What search interest charts do and do not measure is worth understanding before relying on them.
What the records leave out
Even when several sources agree in direction, there are gaps that are easy to overlook.
- Substitution. If someone orders a delivered meal, what did they otherwise do? They may have cooked, eaten out or skipped the meal. A rise in delivery orders does not by itself say what it replaced.
- Who is ordering. Totals can hide whether many people order occasionally or a smaller group orders often. A change in the average could come from either.
- Churn. Meal kit subscriptions are often described as having people who sign up, pause and cancel. A count of sign-ups does not reveal how many remain, and a count of current subscribers does not show how many tried and left.
- Promotions and prices. Orders can rise when discounts, free trials or delivery fee waivers are offered and fall when they end. Growth driven by promotions says something different from growth at stable prices.
- Fees and mark-ups. Delivered food can cost more than the same food collected in person, once fees and menu mark-ups are included. Spending figures may mix food cost with service charges.
- Informal and local channels. Orders taken by phone, by messaging apps or through small local services may not appear in app or payment data at all.
- Unmeasured places. Data tends to be richest in cities with dense delivery coverage. Rural areas, where delivery may be limited or unavailable, are often underrepresented.
- Waste and packaging. Records of orders usually say nothing about the food or packaging that is thrown away afterwards. Meal kits are sometimes described as reducing food waste because of portioning, and sometimes criticised for packaging. Without measurement of the same households before and after, neither claim is easy to support from order data alone.
Why timing and context matter
Delivery grew for many people during periods when going out was restricted or less appealing, and patterns afterwards may differ from those during the peak. A single year of data can mislead if it falls on an unusual moment. Seasonal habits, weather, local events and changes in work patterns such as more time at home with hybrid work can also affect how often people order. Assigning a change to one cause is risky, since several plausible explanations may fit the same trend.
A related caution concerns the “trend” itself. Services that attract many early users may later find their customer base looks different. People who adopt something first tend to be unrepresentative, and early rapid growth often slows as the pool of likely customers shrinks. Growth rates from a small starting point can look dramatic, while a mature category may grow slowly even when it is large.
Reading a claim about home ordering
A few checks tend to help:
- Is the claim about meal kits, prepared food, groceries or all of them?
- Is the figure orders, customers, spending or app activity?
- Does it come from one company or from a wider source?
- Is the comparison between consistent definitions and periods?
- Does it say anything about what the ordering replaced or who is ordering?
None of this is to say that the shift is imaginary. Ordering food at home has plainly become easier through apps and subscriptions, and many sources point the same way. The caution is about precision: the records are better at showing that a channel exists and moves than at explaining how much eating habits changed or how lasting the change is. As always, the strongest conclusions come from combining sources, noting what each leaves out, and avoiding the temptation to read a single metric as the whole story.