Introduction to Customer Acquisition Cost

There are many advanced marketing concepts I’d like to cover in this blog, like incrementality, cannibalization, and attribution.

To talk about those, I want to establish a baseline so everyone understands the issues. Marketing is full of jargon and ideas explained in confusing ways. And honestly lots of bullshit and snake oil too. I want to help fight that.

Who am I to write about this? I helped run growth early at Dropbox, growing 12X in 2 years. I built a social graph analysis startup dedicated to helping social invites. Lyft acquired that startup, and I helped build tools to run their $X00M acquisition budget. So let’s dive in:

Customer Acquisition Cost.

This is often abbreviated as CAC. CAC measures how much you need to spend in order to acquire a new customer.

Measuring CAC is a ratio: 

money spent on acquisition / number of customers acquired

Typically this is measured for a time period, like a week or month, so that you can see how the value is changing. You can also average everything together. Measuring one time period means you need to measure cohorts, so that “number of customers” is a specific signup period, as with the dollars spent. Already this can be complex where the money you spend has a lag from when you see signups, but again let’s keep this simple.

Let’s explain CAC through examples from different marketing channels. This isn’t even close to a complete list.

Social Invite: very low CAC.
Think back to how you first started using Facebook or LinkedIn. You probably got an invite from a friend or family member. It was sent over email or SMS. The company didn’t spend any money directly on getting this invite out there. So how do you calculate CAC? This gets complicated, but technically the product and engineering work were part of that process and those cost money in headcount. Those costs aren’t typically counted. You do look at the fees to vendors that help you send email. These costs might average to pennies per user.

Search engine results: very low CAC.
Instead of hearing about Facebook from a friend, you search a friend’s name in google and find their profile on Facebook. You sign up. How much did Facebook spend on that signup? Well, like invites, the product and engineering work is part of this, but typically not counted. There are also vendors which have direct cost.

Word of mouth, invites, and search results are often padding for other costlier channels. They increase your customers without increasing spend, so they can dramatically lower your CAC.

It sounds amazing, but getting it right is hard. Experiments to see what changes Google's black box are complicated and take time. Maybe significant search traffic requires a very large number of pages, like from user generated content. Building a product people love enough to share is very hard.

Ads: highly variable CAC.
Let’s make this stupidly simple: you spend money and potential customers see your message. This can be lots of money for lots of people, lots of money for a few valuable people, or anywhere in between. This means very many kinds of companies can leverage ads. Generally, the more automation you build here the better. You can test different targeting, refresh creative, and run multiple campaigns. This kind of engineering is too often ignored as companies dedicate their engineering to building product.

Content: moderate to high CAC.
Write a blog post, and people can find that post through search and social media. Maybe they subscribe to your blog so you learn their email. Then you do email content to nurture them to become a user. What are the costs here? Hosting a blog can be free, though there are many tools worth paying for. The real cost here is your marketing staff writing the content. This is often hidden because you might count your payroll as a different bucket of costs than what goes into your CAC. But just think: if you wanted to 10X the content production, what would need to change?

Sales teams: high CAC
Sales teams are the clearest headcount acquisition costs because their compensation is often tied to their performance in bringing in a new customer. If you close a $X deal, you’ll get $Y commission. That is an explicit ratio of customers to dollars spent. As with everything, this can get more complicated. Do your numbers drift? What about different market segments, like small vs large companies? What about a new rep’s performance in their first month vs first year? Note that this channel isn’t available for all products because reps can’t make a living on commission on low cost products. You’ve never gotten a call about someone selling you toothpaste, right?

This is all just scratching the surface. If you’re doing multiple channels here, how do you know which one caused the user to signup? If you dial up spend in one channel and get more users, would that user have signed up anyway? Is your marginal spend worth it? What should you measure if your company is just you and your cofounder?