How Subscription and Retail Stores Help Direct-To-Consumer Companies (Or Don’t)

Josh Hix
11 min readMay 8, 2018

Inc recently published a piece on the explosion of Direct-to-Consumer (DTC) companies, along with some of the challenges these businesses face. I’m one of the co-founders and CEO of Plated, a 6 year old DTC meal kit business we built from 2 founders on a couch to several hundred million dollars in revenue before being acquired by the Albertsons Companies, and I couldn’t agree more with most of what the author wrote.

However, I think there are two areas worth expanding on a bit: The idea that “CAC is the New Rent”, and that Subscription may be a solution for the challenges of needing to pay this “rent”.

First, a few quick definitions and some (example) numbers we’ll build into a Toy Model to illustrate our thinking (shout out to our fantastic Head of Strategic Finance Steven Cruz for the modeling and editing help):

Customer Acquisition Cost, or CAC

There’s no universal definition for this, so let’s build an example: If we spend $90 on Facebook and 10 people sign up as a result (we can track the clicks from Facebook to our website, so we understand the cause and effect), our CAC would be $9 ($90 / 10 new customer).

However, we also give a discount to 1st time customers, and we lose money on that first order (this isn’t true for all businesses, but it is for many subscriptions). Let’s assume this costs us $1.

Therefore, our fully loaded CAC is $9 in media + $1 in discount cost = $10 per customer.

Lifetime Value, or LTV

Let’s assume we know with certainty that on average these customers will spend $100 dollars over 2 years with us (some will spend $10, some $1,000, but $100 on average in our example).

Let’s also assume that after 1 year, 25% of customers are still buying from us and turn into long term, loyal customers (though they’ll still quit our subscription at a rate of roughly .25% per month over time as well after those first 12 months— people move and enter new phases of life). We’ve seen data on many subscription commerce businesses across varied categories over the years (including meal kit businesses, other food subscriptions, personal care subscriptions, and others), and the vast majority have retention in this range (more on this topic below)

Let’s further assume we have a 40% gross margin — so we make $40 in profit per $100 of revenue.

We therefore have a 4:1 LTV:CAC once a group of customers (a cohort) has been with us for 2 years. Said another way, we make $40 in profit for every $10 we invest in acquiring a customer (again, on average and after 2 years).

The $30 we make in net profit per customer ($40 in profit — $10 to acquire them) goes to help pay for rent and salaries for our team, and one day to make a profit for our investors.

Auction based channels

Many online (and even some offline) media channels are auction based. Let’s stick with Facebook as our example (and we’re making some simplifying assumptions here, the algorithms that run these auctions are highly complex and usually top secret).

The advertiser (you) puts in a “bid” for media — in our example, you say you’re willing to pay up to $100 to generate visits to your website, and if you’ll remember you end up with 10 new customers as a result. You could also bid $50, or $200, or any other amount within reason. As a general rule though, the more you’re spending, the fewer visits or customers you’ll generate with each incremental dollar of spend. And though the x and y scale will differ for each company (as a function of variables like website conversion rate, brand awareness, etc), in general the curves looks like this:

Non-auction channels

In contrast, some media channels are non-auction. In other words, it costs $x to buy a unit of media — a page in a magazine, or a 30 second television spot, and you can (within limits) buy as many of these as you want, without the price going up per unit. The cost / customer curve would look like this:


When a customer quits our subscription, we say they’ve churned. If 5%, or 5 of our original 100 customers, quits in the first month we would say we have a 5% monthly churn rate.


Once a customer has quit, we’ll try to convince them to sign up again — maybe they’ve just gone on holiday for the summer. Most people will never take us up on it, but a few will — these are reactivations. We’ll assume we can reactivate 0.75% of people who quit.

With these definitions and the beginnings of our Toy Model in place, let’s talk through the 2 ideas highlighted above.

Idea 1: CAC is the new rent

The original, ecommerce 1.0 idea back in the late 90’s / early 00’s was that you could avoid rent (and the markup of a retailer) by selling online, and therefore build a more profitable business. Generally speaking though, what’s happened is that without a corner storefront for customers to walk by, they forget about your brand and rarely make repeat purchases.

To avoid the fate of the forgotten, ecommerce businesses end up spending large amounts on advertising (on Facebook, Google and other channels) to remind customers they exist and therefore to purchase (either for the first time or as a repeat purchaser). This advertising ends up looking like rent from an overall financial perspective — it’s expensive, and you can’t ever stop spending on it. Hence, as

says, “CAC is the new rent.” But the real trouble is that CAC (advertising) is actually worse than rent in some ways — it gets less efficient the more you spend. In our Toy Model, if we spend $1.06MM per month instead of $265,000, we get 43K customers instead of 20K. In other words, we spend 4x on marketing and only get 2.15x the number of customers — we’ve experienced this dynamic ourselves in real life, and seen many other companies struggle with it first hand as well. If we were selling entirely through retail stores and increased our spend 4x, we should be acquiring 4x the number of customers (we’re definitely making some simplifying assumptions here, ignoring some of the additional management overhead of retail for one, but we believe the concept to be directionally accurate).

At the core of this problem is the reality that much of online spend on customer acquisition happens in auction based channels, unlike in retail (which is a non-auction channel) where costs are linear. In other words, your 1,000th store costs roughly the same in rent as your 10th, while the 1,000th customer you acquire is decidedly more expensive than your 10th. As an added bonus, for a lot of businesses retail serves as more than simply advertising and customer acquisition — in other words, more than just an auction based CAC alternative. Retail is clearly harder to get started in than simply starting a website and spending $100 per week on Facebook (about Plated’s original budget) — so what’s the solution? Let’s return to that at the end.

Idea 2: It’s really expensive to acquire a lot of customers online, and subscriptions are the answer

If you believe idea #1 — you have to pay “rent” in the form of advertising, or CAC, to drive every purchase customers make from a DTC brand — then why not just offer a subscription so that once you acquire a customer you never have to advertise to them to encourage them to buy again? You bring them through the front door of your “store” once, and they keep buying indefinitely. It’s a tried and true business model that works well for media (and other) companies.

The problem here is that most customers in subscription commerce models stop buying despite (or in some cases because of) the subscription. As mentioned in our definitions above, we’ve seen data on a lot of subscription commerce businesses over the years, and for virtually all of them only ~25% of subscribers turn into “loyal” customers and keep buying over the long term (think 1 or more years). As a general rule, there’s simply too much “friction” for most people in receiving a physical product every week / month / quarter, and they quit — this is different from media subscriptions, where most people are willing to pay a monthly fee even if they happen to not watch, read or listen much that month (and prices for these businesses tend to be lower as well). Said another way, for roughly ¼ of people who like your product enough to try it, a subscription is the perfect way to buy — for others though, you’ll ultimately need to let them buy in additional ways to make them happy.

But, a subscription-only commerce business still appears to be a better business model than an a transactional one, so you forge ahead. Before you know it, you’re making $30 from every customer you sign up (on average — and this is a healthy LTV:CAC ratio), and your business is growing. VC’s are tripping over themselves to invest in your company. Where’s the problem? Because you’re losing so many customers, your business reaches a natural “ceiling” where you cannot grow beyond (see below for Revenue and EBITDA charts). AND, you have to keep advertising to reactivate customers, otherwise your business will shrink. This isn’t exactly the panacea you thought it would be — and in fact looks a lot like traditional retail, if not worse. In our Toy Model, our business runs at 4% EBITDA, and requires $11.5MM in capital to get to profitability.

And now that you’ve raised money, your investors are constantly encouraging you to Go Faster! and sign up all the “loyal” subscription customers before your competition gets to them. But going faster makes things worse, not better.

Let’s use our Toy Model for illustration. If you “go slowly”, your revenue and profitability look like this:

If you give in to your investors (and let’s be honest, your secret desires to grow fast and take over the world) and go fast — here’s what it looks like to acquire, say, 5x as many customers per month, using online auction-based channels:

Again, these are example numbers, but the dynamic illustrated is very real. And this is based on a subscription model where 25% of the customers you acquire become repeat customers for many many years — the challenges are even worse if you need to advertise to convince people to make their 2nd, 3rd and more purchases (i.e. you have a transactional business).

So here you are, facing down a ~4% EBITDA business that can’t grow beyond $42.5MM in revenue, you still need to raise and invest $11.5MM to get there, and the faster you try to grow the worst the situation is. What’s the solution?

Epilogue: 1+1 = 3 (Omnichannel) — this is how you “go fast” and “go profitably”

While we may have painted a daunting picture of DTC business models, there are clear advantages as well. Relative to traditional retail, where you have no real idea who’s buying your product, DTC models allow you to natively understand your customer, serve them better, and improve your products based on that learning. That alone can be enough of an advantage to start and sustain a business for years. And relative to retail stores you run yourself, you can grow and prove out an online DTC brand much, much more quickly than building new stores. In the early days when you only need to acquire small numbers of customers to grow very rapidly, your CACs can be low and allow you to operate an efficient marketing engine. Returning to our two CAC diagrams from earlier, in the shaded area (while your business is still early on and you’re still refining and iterating on the product) it makes some sense to be online only, where you can test iterate very quickly. The challenges often come in the later years.

For many products, and for many customers, retail is a feature, not a bug — they want to be able to touch and feel a product before buying online, or to buy at the last minute in a store. Many of these customers will also never subscribe to your product or service — they want to buy more flexibly, or on demand.

In our view, Omnichannel — the idea that you should allow your customer to buy from you anywhere, any way they want, online or offline, is the clear answer for most products. Omnichannel done well should make customers cheaper to acquire — both because having a retail presence will raise your brand awareness and lower CAC’s over time, and because many customers want to try your product in a low risk way before subscribing later. For roughly ¼ of people who try your product, a subscription is perfect. For other customers, they’ll never subscribe — and you’ll only be able to serve them with a retail or transactional offering.

Retail stores can also serve as miniature warehouses of sorts, making your delivery to customers faster when they do order online. They can also serve as pickup points for the “click and collect” customer.

In short, an Omnichannel business model done well should allow you to both “go slow” in your online media, keeping CACs at a healthy and efficient level and to vastly improve your customer experience online and offline which leads to higher customer retention over time. And this leads to a healthier, more profitable business.

It probably won’t surprise a lot of people, but the best customer experience usually wins — and for commerce businesses, that means engaging your customer online and offline, because that’s where all of us spend our time now!

Model Scenarios for reference:

“Go Fast” Toy Model Version: Customer Acquisition set at 25,000 new subscribers per month.

“Go Slow” Toy Model Version: Customer Acquisition set at 5,000 new subscribers per month.

Editable Toy Model Template



Josh Hix

Engineer & Serial Entrepreneur. Co-founder, CEO @Plated.