An Overview on Setting a COS or ROAS Target
There's a term, "Cost of Sale" (COS), that's used to describe advertising costs as a percent of revenue, which is also the reciprocal of the Return on Investment (ROI, also known as Return on AdSpend, or ROAS). Before I started working with new clients, I naively thought that fellow retailers already had a specific COS target in mind when they're looking at their AdWords campaigns.
So, when I started having conversations under the new context of helping them with their accounts, and I asked "Ok, and what's your COS target?", I started hearing what is now a common reply:
How should we pick our AdWords COS?
While there are probably as many approaches as marketers out there, and every company is different, here is a general approach that should help guide the conversation.
Start at your P&L
As a company, do you know how much of your top-line revenue you plan to spend on marketing? Is it 10%, or 20%, or 30%? This is going to be a function of your gross margin, and your overhead costs.
Paid vs. UnPaid Mix Ratio
Let's say you're aiming for a 10% marketing expense at the company level. Now, let's take a look at your inbound marketing mix. Take traffic sources that are free or very inexpensive (organic traffic for sure, earned social, maybe email), and lump them all together. Now, take all of the more expensive traffic sources (PPC, marketplaces, CSEs, affiliates, paid social, etc.), and lump them together, too.
On a revenue basis, how do these two piles compare to one another?
Your free or cheap channels are subsidizing your paid or expensive channels, after all, so it's important to have an idea how these two mix out against one another. If half of your revenue is coming from organic Google traffic, for example, then the other half of your revenue could support twice the COS, in theory.
Using our 10% marketing expense target for your hypothetical company, let's now assume that 50% of revenue is coming from cheap or free inbound sources, like organic Google traffic. This means that the other half of your traffic could run at 20% COS, and your overall company marketing expenses would be spot-on at 10%.
Isolating Shopping Behavior Groups
From here, the strategy starts looking like a fractal--the same pattern repeated further and further down. Now, let's look just at your Google AdWords account. If you're running Branded campaigns (those associated directly with your brand, your domain, your coupons, etc.), then these are going to be producing vastly different performance numbers than the rest of the account. That's because they're isolating a different group of shoppers, who are exhibiting a different type of behavior. These are shoppers who are already familiar with your brand, and knowing that you sell products, are coming to your site to buy them. Few things are further down the funnel than that.
Imagine that your Branded Campaigns make up half of the conversions and revenue recorded in your AdWords account, and your Shopping, Text, etc. campaigns make up the other half. Further imagine that the Branded Campaigns are only spending 5% of revenue--that's one fourth the Account-level COS target of 20%.
If you're only watching COS at the Account level, though, this means that the rest of the Campaigns could be running as high as 35%, while the Branded Campaign mixes it back out to 20% at the Account level. Despite the fact that we want to be subsidizing our high cost efforts with our lower cost efforts, that's probably far more drastic than we intended. If our Gross Margin is 30%, for example, we are actively losing money on those non-Branded campaigns!
Set a Ceiling
What do we do, then? The math works out, and we mix out to the company's overall targets, but we know there's something wrong when a large subset of transactions are individually non-profitable.
The easiest solution is to set a Ceiling COS. This is a guideline that can be used either on Campaigns, or AdGroups, or even on transaction-level costs in other channels.
What's the COS where you wouldn't want to take the sale?
The way I describe this point is to consider a single order (which I almost never do, due to statistical insignificance, but this is mostly a metaphor), and then imagine the cost level where you wouldn't actually want the sale. There's break-even with margin, obviously. Then there's $1 better than break-even, but if all of your orders were at that level, you'd never pay for your overhead. So it's probably a shred more profitable than that...
Like the first question, of company-level marketing expense planning, this is going to be something you'll need to identify for yourself, given your deep understanding of your own business.
Once you have that number, though, implementing it is pretty simple. No matter how far down you slice your traffic sources, you simply level out at what ever that maximum allowable COS would be, and optimize around it.
So, let's say that our ceiling COS is 25%, then we'd be running the non-Branded Campaign at 25%, the Branded Campaign is 5%, and (assuming the 50/50 split of revenue between the two) the Account would mix out to 15%--less than the 20% target, but we know that when you peel it back, that's where it should be. It's not the same as if the Shopping Campaign itself were running at 15%, where you know you're missing opportunities. Rather, this is a result of properly monitoring and managing your expense mixes.
Revisit and Revitalize
Once you start shifting your budgets around to accommodate a new strategy like this one, it will start to shift your ratios, though. For example, if you were to increase your AdWords COS (and bias it based on your Branded Campaigns, of course), then you'd be getting more paid traffic, and that might shift your Paid vs. Un-Paid ratio at the root level, and you'll want to adjust again.
So, it ends up being a moving target, but you'll find that it will stabilize over time. Just be sure you keep an eye on it as you go, and it should be very successful.
Have questions? Drop me a line via the Contact form--I love to talk shop.