# Optimizing the Cost of Customer Acquisition: Modeling Metrics to Drive Startup Success

I admit it, I’m a weird guy. How many others have a PhD in financial modeling combined with an MBA in finance and accounting?  I’ve never put any of this on my business card (not even my CPA) because degrees are not necessarily the indicators of who will succeed in business. After all, the one thing Bill Gates, Mark Zuckerberg, Michael Dell, Ted Waitt and Larry Ellison have in common, besides 4 of the 5 being billionaires before turning 30, was they never graduated from college. Don’t get me wrong: I’m notadvocating dropping out of school. In fact, in this post, I will leverage my finance and modeling background to discuss further how to better predict the future success of companies. When we evaluate companies, we like to see that:

LTV (lifetime value) > CAC (customer acquisition cost) by “a lot”

In my February 20th post, I showed how to go from LTR (lifetime revenue of a customer) to LTV (variable profits).  In this post we’ll discuss how to calculate CAC (customer acquisition cost), and in subsequent posts, will discuss estimating LTR and other related issues.  Ultimately, we will discuss what it means to have a healthy LTV/CAC ratio, which is a key measure of whether a business can scale profitably and efficiently.

Cost of Customer Acquisition

Determining CAC requires considering all costs to generate leads and then calculating the cost per lead. Once this is established, the next step is determining dollars/lead spent on converting a lead to a customer and this plus the cost per lead becomes the total cost of driving leads towards monetization. If this amount is divided by the percentage of leads that convert, we have the cost of acquiring a customer or CAC.  We’ll discuss measuring CAC for an enterprise SaaS business in a future post. For now we’ll review how an ecommerce site, “ABC Company”, calculates CAC.

Most online commerce sites have a multi-step process for acquiring a paying customer (See Figure 1). The steps are:

1. Spend money each month to drive potential customers to the site and get them to provide their email address (e.g. register). This spend divided by the number of registrations is the cost of acquiring an email.
2. Using onsite and email campaigns to offer users the opportunity to get something for free in exchange for providing their credit card numbers.  Freemium models offer a free version of the product, and customers pay if they want the premium version.
3. Convert those that receive the free product to paying subscribers. For those with a freemium model, there is a bit of a catch-22. On the one hand, the free version has to be good enough to establish strong value in the minds of users. On the other hand, it can’t be so good that the need to convert to the premium (paid) version seems unnecessary.

Figure 1: Conversion Funnel

A variety of channels can be used to drive email registration.  These include SEM (search engine marketing), SEO (search engine optimization), Facebook advertising, banners, bloggers, traditional print advertising, PR, and even TV advertising. The best method of analyzing the results is by cohort.  A cohort is a group of subjects/customers who have shared a particular event during a particular time span.  Users generated from each channel in a particular month are a cohort.  Average email registration cost for each cohort is total dollars spent on that channel divided by the number of registrations.  Since registration may not occur in the same month, it’s important to track the cohort over time.  Figure 2 shows a calculation of registration cost per cohort and overall cost for November 2013 for the ABC Company.

Figure 2: November 2013 Cohort Analysis for ABC Company

In November, 2013 the ABC Company spent an average of \$4.00 to acquire a registered user.  Excluding TV, almost 95% of registered users were acquired through channels that yielded a cost of under \$2.38 per registration. The action indicated by the analysis is to re-allocate the TV budget to other sources, particularly Facebook advertising and bloggers. ABC did re-allocate spend and brought their average cost of acquiring an email down to \$2.50.

ABC wants registered users to try a free month of its online product (i.e. become a free subscriber or “Sub”).  In exchange, it asks for the sub’s credit card information.  By dividing the percent who become free subs into the cost of getting a registered customer, the yield is the base cost of a free customer.  For example, if ABC is able to convert 12.5% of registered users in an average cohort to free subs, and the cost of the email is \$2.50, then the base cost of the free sub is \$2.50/.125, or \$20.00. Determining the percent that convert from the cohort requires tracking it over time as it can take many months for some members to convert.

The cost of the free month plus the base cost determines the total cost of a free sub.  Where the product is virtual this cost may be zero (making it easier to give one month away). For physical products like clothing, food, cosmetics or socks, the cost includes items in the box, packaging and labor for putting the box together and shipping. Since ABC charges \$4.50 in shipping for the “free box” and has a total cost of \$12.50 to create and ship the box, the net cost is \$8.00. Adding this to their \$20.00 base cost gives \$28.00 total cost for a free sub.

The last step in calculating CAC is determining what portion of free subs convert to paying subs and dividing total cost of a free sub by this percentage. Table 3 tracks each of ABCs cohorts of free subs to see what portion convert to paying subscribers. It shows that over time about 75% covert. If we take the cost of a free sub and divide that by the 75% conversion rate, ABCs CAC is \$37.33.

Figure 3: Cumulative Conversion from Free Sub to Paying Sub

As with almost anything online, A/B testing (presenting alternatives to different users to determine which alternative drives more conversions/registrations/purchases) helps improve the CAC. Examples of things to test include:

1. Various sources of site visitors
2. Whether or not to have a free month in a subscription model
3. Whether or not to ask for a credit card as a condition of getting a free month
4. What to put in the free product in a freemium model
5. Optimal pricing for a subscription. When the product is virtual this becomes a question of optimizing total revenue versus the CAC.
6. Alternate pitches to get users to take actions

Every step in the process should be analyzed to optimize the CAC as lowering this cost will allow a company to scale more quickly.