A common issue that marketing teams face is the proper way to measure the success of their campaigns. This is a crucial issue because how you measure success determines where you invest your limited resources.
This post discusses some of the most commonly used metrics when determining ad spend and some of their drawbacks before showing how customer lifetime value avoids these issues.
The fictional story behind cost per acquisition
Let's start with the worst metric that is unfortunately quite common to see in use: Cost Per Acquisition, or CPA.
CPA is measured by dividing the advertising spend - let's say $100,000 - by the number of customers acquired through the campaign - let's say 10,000. In our example, CPA is $10.
Many marketing teams like to use CPA because it seems relatively straightforward in a world where you want more for less. A campaign with a CPA of $8 is better than one with a CPA of $10 - I'm spending my advertising resources 20% more efficiently! Or am I?
The drawback with CPA is that it's blind to the revenue side of the equation. What if the $10 CPA campaign brought in customers spending an average of $80 on a product after clicking your ad, where the $8 CPA campaign brought in customers spending $8? Your money was spend 20% more efficiently, but the campaign goes from very profitable to not profitable at all.
This drawback of CPA leads companies to consider another popular - and admittedly better - metric: return on ad spend, or ROAS.
Return on ad spend - incrementally better
ROAS is calculated as the revenue earned from a campaign divided by the ad spend. If a campaign brings in $100,000 and costs $20,000, ROAS is $5 (earning $5 for every $1 spent).
By bringing in revenue and spend, you get a much better idea of which campaigns are most profitable. So what's the drawback?
First, ROAS doesn't do a good job of understanding the long-term value of the customers you acquire. Often, revenue for ROAS is tracked for the session where the customer clicks, or sometimes using an attribution window tracking the customer's revenue for a fixed period like 7 days or 30 days.
These numbers are totaled, thrown into a spreadsheet, presented to leadership, and then forgotten about. What a customer does after this campaign doesn't matter. A campaign bringing in one-time customers is seen as just as valuable as one bringing in long-term, loyal, high-spend customers as long as they spent the same amount in the short period after clicking the ad. This is not the way to make long-term decisions on resource allocation.
Second, ROAS suffers from ignoring incrementality. In advertising, incrementality measures the customers / revenue earned from a campaign that you would not have earned otherwise; many customers who purchased because of your campaign would have purchased anyway.
In the above example, if $60,000 of the $100,000 campaign revenue were acquired from customers who would have purchased anyway (for example, if the sponsored ad on Google results wasn't there, they'd have just clicked the next organic link to your site), then the true incremental ROAS would be $40,000 ($100k total revenue minus $60k that would have been earned anyway) divided by $20,000, or a ROAS of $2. In some cases, ROAS falls below $1, meaning the campaign was a net detractor to the bottom-line.
Many marketing teams at this point throw up their hands and say, "It's the best we've got! Measuring incrementality is complex and expensive and we trust our agency/vendors when they tell us their campaigns are a value-add." Alarm bells should be ringing here. Trusting the agency/vendor who earns more when you spend more is like trusting your sheep to the wolves on blind faith.
Using Customer Lifetime Value for determining campaign success
The flaws in CPA and ROAS come from their limitations in the scope of what they look at. CPA doesn't consider revenue, and ROAS considers neither long-term value nor incrementality. Customer Lifetime Value (CLV) provides all of the above.
Revenue is a primary metric in calculating CLV as CLV is just the expected revenue from a customer over their time with you. (Even better, you can subtract out costs to have CLV calculated purely from profit.)
As is implied in the name, CLV is built for a long-term view. What a customer does in the next week isn't as important as their behavior with you in the many months to come, getting rid of the bias for short-term results over long-term profitability.
Since CLV is a cohort-based analysis - where you calculate CLV for all customers who start in a given month/quarter - you have a baked-in, no-nonsense way of calculating incrementality that doesn't rely on blind faith in the vendors. Simply calculate CLV for customers who are acquired by your ad (or for existing customers who purchase because of your ad), and compare them to the baseline for their cohort.
The advantage of the last item - calculating incrementality - is a huge benefit. Not only can you see how revenue changes for customers influenced by your ad, but also how the probability that the customer will stay active with you changed. Did your campaign increase the lifetime of customers with your company? If so, this is a wonderful point to make in the next budgeting meeting!
I hope this short article was helpful in illuminating the downsides of common metrics used to evaluate advertising success and spend and the significant upside when using customer lifetime value instead.
Comments