She Wants to Lure You Back Into the Kitchen With "Meal Kits" — and She Knows It's a Hard Business | Founders Speak

The semi-prepared ingredient delivery business is in bad shape. So what opportunity did Pan Xiaoyue see when she founded 321 Cooking?

Braised oxtail with chestnuts and taro; pork-and-cabbage rolls in tomato sauce; beef stewed in fresh chicken broth with celtuce; a creamy soup of shrimp paste and mushrooms. This Wednesday evening, it took me about 20 minutes to make all four dishes.

I'm no great cook, and my work hours mean I never reach the market before it closes. These ingredient kits — already partly prepped, with the cooking steps written out — saved me time.

I ordered the kit from Pan Xiaoyue last Saturday.

The package held four neatly arranged meal boxes. A paper band around each box gave the dish's name and the order to cook in; inside were the ingredients, washed and cut, along with the seasonings. Four color printouts of the steps came with it. Following them, each dish took 5 to 10 minutes.

321 Cooking, Pan Xiaoyue's subscription service for semi-prepared ingredients, went live this September. She had spent five years as content editor-in-chief at the recipe site Xiachufang. In April, with 1 million yuan in startup money from her former employer, she and several senior colleagues moved into an apartment of just over 100 square meters near Wangjing West and began.

The business delivers ingredients for a fixed set of four dishes each week, with a new menu every week. 321 Cooking also pushes monthly and annual subscriptions, though you can skip any week's menu you don't like and roll it forward.

Even in the busy early months, Pan kept up her habit of posting regularly on her Xiachufang page. Beside a photo of one breakfast she told a story from Cooking Fairy 《料理仙姬》 and wrote, "A sense of ritual can save a gray life."

Every so often she shares her fine-grained, carefully tended life online. Even for a glass of spinach juice, she'll post a well-angled photo and set out the steps clearly.

How Is a Meal Kit Made?

With a long braid and a linen dress, Pan Xiaoyue at first looks like the food photos she posts on Xiachufang — arty, bohemian.

That impression goes the moment she starts talking about work. She speaks fast, thinks in order, and hands out assignments to her staff as 1, 2, 3.

321 Cooking's office has nothing arty about it either. It is a somewhat shabby mixed-use apartment that serves as both the team's office and its dish development center.

The dish development center is simply the kitchen in that apartment. Pan insists on developing dishes in an ordinary kitchen. "Our users cook in their own kitchens at home, so the setting has to be the same."

From May to August this year, the team's main work was developing new dishes. Much as you'd picture a restaurant doing it, Pan jokes that her job back then was to eat, eat and eat. "But it's a complicated process. To start with, it isn't easy to find a chef who fits what we need."

"Plenty of chefs can cook remarkable dishes, but every one of our dishes has to have a traceable process, and a chef's habit is often not to remember how much salt went in, or what the exact steps were."

The chefs who finally clicked with the team came from Da Dong Roast Duck Restaurant and the French restaurant TRB Bites. During development they drew on their own experience to make three dishes a day.

The only rule was that a dish had to be both suitable for pre-preparation and good to eat. As Pan sees it, "turning a dish into a kit can't cost you the freshness of the ingredients. Steamed fish needs a live fish killed on the spot; cleaning octopus is expensive. Neither suits a semi-prepared kit."

Once the dishes and a first set of steps were settled, the team ran a tasting inside Xiachufang. More than 40 employees gave feedback in a WeChat group, the recipes were revised, and more than 200 dishes were fixed. The kits themselves are made in a central kitchen of just over 100 square meters.

"It's something like a factory floor. Small, but it has everything, and the hygiene standards are very strict. We found a central-kitchen consultant with more than a decade of experience to plan the layout, the work rules and so on."

But a head chef can't just invent a dish, write a recipe, and have the central kitchen turn it out in bulk. In between comes a "translation": someone has to write an operating sheet the kitchen can follow.

Think of the central kitchen as a Foxconn assembly line, except that what's assembled is food, not electronics.

"Every dish has to be broken into a lot of operating steps. If a kit holds eight small items, one of those items alone may take three steps."

Take the bag of meatballs in one kit. Producing it in the central kitchen takes several steps. The meat room grinds chicken breast into a paste and chops chicken cartilage fine; the two are mixed and sent to the hot kitchen, which squeezes the mixture into balls, poaches them in boiling water, and vacuum-seals them once they cool.

At the extreme, one vacuum-sealed packet a user receives has to pass through every room in the central kitchen — vegetable washing, cold kitchen, meat room, hot kitchen — before it can be boxed. Then the packers put all the raw material for one dish into a single meal box, with nothing misplaced or missing.

With all that in place, the remaining work is choosing each week's menu and writing the recipes.

"I want every step of cooking to go smoothly for the user," Pan says. Having written so much recipe content at Xiachufang, she holds herself to a stricter standard here.

As a Child, She Saw the Need in Her Mother

Pan Xiaoyue has imagined something like semi-prepared ingredients since she was small, and she thinks her current venture may have been fated.

She saw the demand for washed and cut vegetables in her own mother.

"My mom loves buying pre-cut vegetables at the supermarket. She doesn't much enjoy cooking, and shopping for food is often a worry for her — what to make today, and how to keep varying it. Pre-cut vegetables are convenient, so she loves buying them."

As a child, while other girls dreamed of opening a café or a flower shop, she dreamed of opening a restaurant that sold pre-cut vegetables: a customer eats a dish there and, if she likes it, takes the ingredients home to cook herself. She was simply born liking to be in the kitchen, she says, fixing good things for herself.

Later Xiaoyue came to Beijing to study screenwriting at the Central Academy of Drama. Unlike most of her classmates, she liked the subject but chose not to enter the industry, given the state of China's screenwriting market. She worked as a magazine editor, and helped plan the Blue Harbour branch of One Way Space. After all that circling she came back to the thing she liked best, and stayed at Xiachufang for five or six years.

Then, last October, Xiachufang's founder, Wang Xusheng, came to her with his own thoughts on semi-prepared ingredients, and the two turned out to agree. She holds to a few principles of her own in building the business: users must be left the pleasure of cooking, and the product can't become purely utilitarian.

Whether she was publishing her own cookbook, working at Xiachufang or starting this company, Pan Xiaoyue says her purpose has never changed: "I hope more people will be willing to cook for themselves and feel the pleasure of it."

But This Is Not an Easy Market

321 Cooking came to the semi-prepared ingredients market late. By early 2015 many similar delivery companies had appeared in China. The larger ones included Qingnian Caijun, which had long run storefronts on delivery platforms such as Ele.me, and more arty ventures like Nuanshi and Xinwei. Most of those startups have since pivoted or closed.

Pan Xiaoyue agrees the meal-kit business is hard to run. But with the industry out of favor, she also sees openings of her own: no company in China has yet put forward subscription-based delivery of semi-prepared ingredients. Companies like Qingnian Caijun, besides burning cash too fast, have deeper flaws.

"If you sell dish by dish, your order value is very low, which pushes your operating costs up — and you still have to add a delivery charge. That leaves you stuck for a long time where it's hard to make your money back."

Delivery charges are a cost O2O founders find very hard to squeeze. But Pan believes the other kind of ingredient-delivery business has the opposite problem: order values that are too high.

"A company like Xinwei started with semi-prepared Western dishes. They also hired very good photographers, so one look at the pictures made you hungry. But the order value was high, which drove the repeat rate down — people only bought on special occasions, at Christmas, say, which is consumption tied to an occasion. So your repurchase cycle gets very long."

So 321 Cooking chose a subscription model popular abroad. One of the examples they looked at was Blue Apron, valued at close to $2 billion. That company also delivers kits of semi-prepared ingredients on a monthly fee, letting ordinary people get restaurant-grade seasonings and good produce and, with a little cooking, put attractive, tasty dishes on the table.

A model like that will inevitably struggle to transplant. Conditions in China are nothing like those abroad, and an ingredient subscription also takes on the Chinese habit of going to the market for groceries.

Pan thinks so too. "It really is much harder. For one thing, people here have no habit of subscribing and buying this way. For another, there are far too many competitors here and users have a lot of choices — buying groceries or ordering delivery is much more convenient in China than abroad."

Still, she believes in the model. "I think this will be a big trend. People have no subscription habit, but subscriptions do exist for flowers and yogurt, and some people are out ahead and can be drawn in. That group will keep growing."

Sending at least four dishes at a time lifts the order value somewhat, and Pan wants the service to become a frequent purchase, so the price isn't high: it sits between eating at a restaurant and buying groceries to cook at home. By her figures, two months in and with no marketing, 321 Cooking takes 80 to 100 subscriptions a week, more than 50% of them repeat orders, and more than 10% of users choose a monthly plan on their own.

The Hardest Part May Be Finding New Users

What weighs on Pan most is bringing in new users. On the day of this interview the young team had just hired its first marketer. Next, 321 Cooking plans to pitch its kit subscription to specific companies, though how that will go isn't clear yet.

Ingredients account for 40% of the cost of 321 Cooking's Sanxian line, and logistics and packaging for another 13%; add labor, operations and the central kitchen, and the margin is thin. Pan has done the arithmetic: on the central kitchen's current supply model, five times the current order volume would fill the kitchen, but breaking even as a company would take eight times.

That means 321 Cooking will find it hard to make its money back before it moves up to a bigger central kitchen. Order growth is therefore everything.

321 Cooking has flaws that may keep new users out. A first-time subscriber, for instance, must order next week's ingredients a full week ahead. The central kitchen isn't efficient enough yet, so a long window is left for it to buy and train.

Not being able to pick the weekly menu is a problem for some users too. Pan says they may add some choice later — picking 4 dishes out of 6, say.

321 Cooking wants to do more as well: kits built around the Lunar New Year's Eve dinner, so a family gets the warmth of eating at home while the parents are freed from the kitchen; or work with famous chefs who supply menus and generate some talk.

These ideas sound interesting, but they all rest on the number of new subscribers continuing to rise. Otherwise it ends up as one more cash-burning game.

Xiachufang, the investor behind 321 Cooking, is itself a classic slow company. Its arty community drew large numbers of users six years ago, but making money has been a serious problem. How do you compete on e-commerce with other recipe sites, or with the big e-commerce platforms? How do you build a supply chain? These are open questions.

As for what comes next, Pan says she is prepared for capital never to take an interest — prepared to let users grow slowly, the way you would running a small restaurant.

"I'm a pessimist. I'm not optimistic about the numbers, because trust is hard to build. I'm fairly confident that once a user has tried our product she'll keep subscribing. But taking that first step is hard for her. I don't want to assume things will go too well."

Q = QDaily

A = Pan Xiaoyue