Why you can only eat that sandwich once…

Or how Data companies and Restaurants are more similar than we imagine

When I met my wife, she wanted to open a restaurant.

Great, I thought — I love eating…

I mean, everyone does. Restaurants are right there will healthcare and transportation. People will always continue to get ill, fly/drive, and, yes, eat food.

We had money to invest, but I did not know anything about restaurants. It was time for some learning.

I took on a couple of evening entry-level jobs at two restaurants. Got fired from one on my 3rd day, but kept the other one.

Two months later, I asked my wife what kind of restaurant she wanted to build? 

One for 10-20 seats with a special that changes every day/week, or one for 50-100+ seats that can potentially grow into a franchise and prepares mostly the same food.

She said she wanted to do the former, and we abandoned the idea because I wasn’t ready to lose $$.

If you know anything about restaurants, you already know where this is going. Restaurants are low margin businesses. Boutique concepts work out very rarely. It is a lot of work with giant real estate bills and major staffing headaches.

A data company is a lot like a restaurant business. You really need to decide what you’re trying to create:

(1) A boutique that creates a daily special for a small select group of enterprise patrons…

(2) Or a low margin solution that is designed to work at scale…

A lot of data founders deceive themselves thinking that they can do (1) and grow into (2). And maybe even have both high margins as well as scale. They imagine companies such as Tableau. They dream about Teslas. And basically equate Data Technologies to everything else we know to work in the tech world: start with expensive non-scalable solutions, and transition to scale over time.

But there is a problem. Data business is a lot more like a restaurant business, than a pure software business.

And while you can start with a non-scalable solution, just as in the case of a restaurant, there is a high chance you will end up with one

There is a unit of utility there. In restaurants it is food (meat in a burger, cheese in a pizza). In data companies the utility unit is a record of a transaction or some user activity. 

Yes, that burger can end up costing $2, $10, or even $50. But you can never sell a $50 burger at a restaurant that also sells them for $2 . The $50 burger can never be sold at scale.

If you’re building a Dashboarding, ETL, or Data Cataloging solution, don’t make a mistake thinking you can sell $50 burgers to millions of people. Figure out what you’re doing first: a $10 dev subscription at scale or a >$50K subscription for enterprise customers.

A burger is still just a piece of meat squeezed between two buns - no matter how many times you look at it. Same goes for that webpage event you captured in your data pipeline.

---

📝 About the Author 

In my former life I was a Data Scientist. I studied Computer Science and Economics for undergrad and then Statistics in graduate school, eventually ending up at MiT, and, among other places, Looker, a business intelligence company, which Google acquired for its Google Cloud.