Cost Per Acquisition, or CPA, is an important (if also flawed) metric that tries to calculate the cost of “buying” a new customer. In simple terms, the math is:
CPA = Total Ad Spend / Number of New Customers
The obvious goal is to drive CPA down so you can acquire more customers for every ad dollar you’re spending. But CPA fails to properly account for one very important detail: revenue.
You might know (roughly) what it costs you to buy new customers, and that’s something—but the bigger question is how much those new customers buy and whether all of that adds up to a profit in the end. The trouble with measuring only CPA is that it fails to take the bigger-picture dollar gains and losses into account.
Return on Ad Spend, or ROAS, is different from CPA. ROAS measures the amount of revenue (NOT profit) your business earns from each ad dollar you spend, as opposed to measuring the number of customers earned.
ROAS = Revenue from Ad Spend / Total Ad Spend
So, for example: if you spend $100 on advertising and generate $300 in revenue, your ROAS is 3x (or just “3”). For every dollar you spend on ads, you earn three dollars in return. That’s good. Do more of that.
Take these two ad campaigns as a simple side-by-side contrast:
Campaign 1
Ad Spend: $500
Conversions: 5
CPA: $100
Revenue: $100
ROAS: 0.2(x)
Campaign 2
Ad Spend: $500
Conversions: 5
CPA: $100
Revenue: $2000
ROAS: 4(x)
See the difference?
In both cases, the CPA is $100. For Campaign 2, a ROAS of 4.0 tells us that this campaign is making the monies, which is muy bueno. But for Campaign 1, a ROAS of 0.2 says that this campaign is actually losing monies, cual no es bueno.
Question: When would it be acceptable to continue running an ad campaign with a ROAS below 1 (i.e. below breakeven)?
Answer: If you’re in Aggressive Growth Mode and you’re willing to lose money to acquire customers who will (someday) carry you to profitability. The more that’s true, the more your ROAS can drop below 1.
Just remember to watch your cash position carefully—because (for example) if you’re buying traffic from Zuck, he’s likely to ping your credit card before you can get to break even. The question becomes: how long can you wait for your ads to be profitable? 30 days? 60 days? 90 days? That’s on you and your math.
If campaigns are intended to make money (i.e. profit) for the business, a good rule of thumb is to aim for a ROAS of 3 or greater.
ROAS is Revenue from Ad Spend / Total Ad Spend. Trouble is, top-line revenue figures are (by nature) a little raw. How can we tweak the Revenue half of the formula so that ROAS figures tell us more?
Adjusted View: Amend the revenue number to include predictable costs. If we’re being specific, the total amount of money you pulled in (without paying any expenses yet) would be gross revenue. After we subtract some stuff, we’ll call it adjusted revenue. It’s not “profit” because we’re not trying to account for all expenses here; we’re just trying to ballpark useful patterns in data.
So what should we subtract out? The best catch-all would be “your cost of goods sold,” or COGS. That’s the sum of all expenses which you had to pay to have the sellable stuff in the first place—so you may as well take that sum (or your quick estimate of it) back out of the gross revenue figure.
📊 Quick Example: Acme has spent $1K on their new ad campaign for Doohickeys, which has sold 100 units and earned $4K in gross revenue. The “simple” ROAS would be 4(x).
But Acme also knows that each Doohickey costs them about $20 total from scratch to warehouse (“$20 landed unit COGS”), an expense which can’t go unpaid. So Acme’s adjusted revenue is $2K and the adjusted ROAS is 2(x) — less impressive-looking, but more meaningful analytically.
Long-Term View: Estimate the lifetime value (LTV) of the new customers. Customer LTV is slippery (if not impossible) as an exact metric, but getting a ballpark estimate is actually pretty simple.
You just need two pieces of general information: your average order value (AOV), and the average number of orders placed by each of your (existing) customers. Multiply those together and voilà, you have a quick Customer LTV.
Then, if you multiply that LTV figure by the number of customers acquired by ads, you have some idea of how much value those new customers might actually bring you over the long haul of things.
⚠️ Important Caution: This trick is less an “adjustment” than a swing towards the opposite problem. The basic ROAS formula doesn’t “know” that many acquired customers will order again, and without advertising costs the second time around. But conversely, the Long-Term View of things doesn’t know that (A) today’s averages are not tomorrow’s guarantees and (B) cash flow is like oxygen: a short period of deprivation can kill you even if you’re healthy. Observing your customers’ LTVs is one thing; depending upon them is quite another, and a terrible idea.
0 Comments