Oi! Tudo bem? I’m back again with a post on CLV / CAC. Enjoy!!!
The Times article that I referenced in yesterday’s post mentions the recent closure of a Brazilian startup called Shoes4You. Shoes4You had a solid pedigree. It was founded by young Frenchman Olivier Grinda, the younger brother of entrepreneurial wunderkind Fabrice Grinda (founder of OLX). The company, which essentially replicated Shoedazzle’s subscription shoes business model in Brazil, was backed by a premier roster of investors, including Accel, Redpoint, and Flybridge. Based on the company’s highly optimistic Series A press release, you’d think these guys were well positioned for success. Unfortunately, as so often occurs in the startup world, things didn’t go exactly as planned. So what happened?
The Times article states that “costs like legal and tax matters along with underdeveloped e-commerce logistics made operations difficult” for Shoes4You, ultimately contributing to the startup’s untimely demise. But every eCommerce startup in Brazil faces these costs, and many are thriving, so there must have been other factors at play. In my view, the failure of Shoes4You can be traced back to the relationship between the company’s cost of customer acquisition (CAC) and its customer lifetime value (CLV).
Startup 101 tells us that in order to operate your startup on a sustainable basis over the long-term, you need to reach a point where each of your new customers contributes more to the company’s bottom line over his / her expected lifetime (CLV) than it costs your company to acquire that customer (CAC). My guess is that Shoes4Your never quite reached this critical threshold.
An example for the layman
Here is a very simple example to better illustrate the fundamental CLV / CAC equation. Let’s say I’m running a new eCommerce startup that sells widgets. I run a model using data gathered from cohort analysis. My model tells me that each of my customers, once activated, is purchasing on average 4 widgets per year over an expected lifetime of three years. I sell each widget for $40, and each sale costs me $35 in variable costs (COGS, payment gateway fees, cost to provide free shipping, taxes, etc), thereby generating $5 of contribution margin. If I get $5 from each sale, and each customer is buying 4 widgets per year for 3 years, each new customer will generate a customer lifetime value of $60. Nice!
But wait. How much am I paying to acquire each of these customers? Let’s go back to our hypothetical widget example. Suppose I’m flush with cash, fresh off a Series A investment. I’m using the money to acquire customers through several channels: I’m running an in-house Google adwords campaign, doing paid advertising on Facebook, working with an online ad agency that runs remessaging display campaigns on my behalf, and working with another agency that is doing display advertising via several Brazil-specific digital ad networks. Each week, my total online ad spend across all these channels is $2,000, and on average I’m gaining about 30 new paying customers per week. This means that my CAC – the amount I’m paying to generate a new paying customer – is $67.
Uh oh. My CLV analysis is telling me that each new customer generates $60 in contribution margin over his lifetime. But it’s costing me $67 to acquire each new customer. This means that for each new customer I bring on board, I’m losing $7!! Not good. When CAC > CLV, you know your startup is in trouble.
So what do I do now? Step 1: I need to dramatically reduce my online marketing spend, potentially even bring it to a halt. At this stage, every customer I bring on board is a money loser, so I need to take a step back and reassess a few things before I keep throwing money out the window. Step 2: I need to think about why my CLV / CAC relationship is out of wack, and come up with a plan of attack to right the ship. A couple things merit consideration here. Let’s look at each in turn:
Customer Acquisition Cost
- Track repeat purchase behavior. On the CAC side, there are things you can do to reduce customer acquisition cost. Run experiments using cohort analysis to determine the ROI on each of your customer acquisition channels. Group all the new buyers who came onto your platform in a given week via a specific channel into a cohort, and track the purchase behavior of that specific cohort over time. You may see that those who came onto your platform via Facebook have a repurchase frequency of 5x a year, versus only 2x a year for those who became first time buyers via your remessaging campaign. This would suggest you should allocate more dollars to facebook ads and reduce your remessaging spend.
- Track channel-specific CAC. Similarly, run experiments to determine channel-specific CAC for each of your customer acquisition channels. You may notice that for every $2000 spent on ad networks, you’re gaining 40 new paying customer, versus only 20 customers for every $2000 spent on google adwords. This would suggest reallocating dollars from adwords to ad networks.
- Optimize the customer purchase funnel. Another idea is to run A/B tests to optimize your customer purchase funnel. Is your website designed such that customers who enter the homepage have an easy, effective, and streamlined means of going from first contact with a product page through to the payment page and onto the actual purchase? Run tests to reduce bounce rate and make sure the highest possible percentage of website visitors make it all the way to the end of the purchase funnel. This will help you get the most bang for your online marketing buck and thereby reduce your CAC.
- The power of experiments. The reality is things are a bit more nuanced than this, but the basic idea is that you can run experiments to determine the relative effectiveness of each of your customer acquisition channels, and optimize accordingly. You can also optimize your customer purchase funnel to reduce bounce rate and enhance purchase probability (and thereby the acquisition of a new customer). All of these things should help reduce your overall CAC, thereby helping come closer to the desired relationship between CLV and CAC.
Customer Lifetime Value
- Improve margins. On the CLV side, there are a couple things we can do. First things first, let’s try to improve margins (in other words, there are multiple ways we can increase the profit contribution resulting from each individual sale we make, which thus increases CLV). First, we can try to reduce COGS by finding new, lower-cost suppliers, or by purchasing in bulk from our existing suppliers. To a certain extent, this will happen naturally as we scale, but it’s definitely something we should be focused on. Second, we can experiment with changes to our shipping policy, potentially moving from a full free shipping model to a partial or subsidized shipping model based on the size of customer orders and their geographic location (be careful to gauge the impact on sales to ensure changes to shipping policies are not completely offset by reductions in customer orders). Finally, (but not exhaustively) have you been offering too many customer discounts, thereby reducing gross margins? Perhaps we can reduce our average customer discount without generating a corresponding decrease in sales… You get the idea right? There are various things we can do to improve margins. Test them all. Remember, better margins —> higher CLV
- Increase basket size. You can also work to increase basket size. Going back to our example, our customers are buying one widget per purchase, four times a year. Can we get them to buy two widgets each time they make a purchase, such that they are buying 8 widgets a year instead of four? The impact on CLV resulting from a greater basket size should be pretty self-explanatory. But how can we get there? Run experiments! Try giving customers free shipping only if they purchase more than one widget at a time, and see what happens. You might be pleasantly surprised!
- Enhance repeat purchase rate. Another way to improve CLV is to improve your customer repeat purchase rate. Remember, CLV is in large part a function of how many times per year your customers purchase your product. How can we improve this? In a nutshell, step into the shoes of the consumer and do everything possible to improve the quality of the customer experience and the quality of your end product. This is a bit of an intangible, and will vary depending on your business model, but the bottom line is that we need to be customer-centric so that our customers buy more frequently.
- Get rid of crappy customers. Whoaaaa wait a second, don’t we love all our customer equally? NO. We do not. Some of our customers have high basket sizes, buy high margin widgets (real world example: designer t-shirts), and have high repeat purchase rates. We LOVE these customers. Other customers rarely buy, and when they do, they buy only one widget, and it’s a low margin widget (real world example: generic underwear). Let’s get rid of the crappy customers! How can we do this? It’s called customer segmentation, my man! Label the email addresses of all crappy customers as “crappy” and the email addresses of all rock star customers as “rock stars”. When one of our crappy customers logs on to make a purchase, make their lives difficult by forcing them to pay for shipping, lengthening the customer purchase funnel, etc… Over time we end up with a much higher quality customer population, thereby increasing our overall CLV.
Back to business
Okay, so let’s assume I’ve slowly made the changes above, and as I’ve done so, I’ve been using cohort analysis and A/B testing to gauge the effectiveness of my strategic tweaks. Over time, I should begin to see my CAC go down, and my projected CLV go up (eventually I’ll have enough longitudinal data to retrospectively calculate my actual CLV). Before long, you’ll get to a point where you feel confident that your CLV > CAC. When you reach this point, guess what? Time to put the pedal to the metal!!
When CLV comfortably exceeds CAC (ideally, it should exceed CAC by 2.5x according to Jeremey Liew of Lightspeed), you are making money off each new customer you bring on board. If you already have cash in the bank, or your business model throws off high amounts of organic cash, go ahead and crank the online ad spend using those funds. If you were in lean startup mode, it may now be time to shift into high gear. Go back to your VCs (or find new ones), show them your new numbers, explicitly tell them your CLV is way higher than your CAC, and raise some friggin’ dough (you should be able to do so at an attractive valuation with your new numbers). When the money is in the bank, go full throttle on the online marketing spend… It’s time to grow grow grow, baby!
I’m definitely not an expert on CLV, but hopefully the insights outlined above serve as a helpful starting point. To learn more about this fascinating and important topic, I suggest checking out the following supplement resources:
- How to estimate lifetime value for an ecommerce business; Sample cohort analysis. Awesome explanatory post by Jeremy Liew of Lightspeed. Contains a link to a sample cohort analysis. Super helpful. Thanks Jeremy!
- How can you improve LTV and CAC? Another classic by Liew.
- Startup Killer: the Cost of Customer Acquisition. Great post by David Skok. I drew one of the charts I used above from this post.
- After the Techcrunch Bump. Josh Kopelman opines on the importance of cohort analysis.
- STARTUP STAGES AND GROWTH THROUGH UNIT ECONOMICS. Nice collection of additional resources on related subjects such as product market fit, cohort analysis, and unit economics. By a startup guy named Michael.
Phewww. Just pumped out this post in two hours, record time! Going to take a little break now. Don’t forget, I’ve got an awesome post coming out soon on the differences between the São Paulo and Mexico City ecosystems. You don’t want to miss it, it’s going to be very solid… So check back mañana! Oh, and if you missed yesterday’s post where I talk about my Wharton journey, check it out! It’s been very well received.