Spending wisely should always be an objective regardless of the economic conditions. But this is especially true for marketers who control a large share of the budget at many companies — and who are the most at risk for wasting money.
The lessons below are based on my experience as CMO of an online shopping destination. As we grew, our executive team learned that startups have two choices when it comes to growth. The first is to grow quickly by spending a lot on marketing. The second is focus on long-term sustainable growth with less marketing spend. And sometimes, particularly during difficult economic times, it’s necessary to transition from one to the other; that is, to increase marketing efficiency.
Why Return on Marketing Spend Is Counterintuitive
Initially, when our company was in rapid growth mode, we had a particular target ROAS (return on ad spend), which we calculated by dividing the tracked revenue generated from a particular campaign by the amount we spent on that campaign. We assumed that any money we could spend above this target ROAS was good, since we would then hit our revenue and marketing efficiency goals.
However, as we learned, ROAS will decrease as you scale up your marketing. In other words: you cannot generate the same ROAS at a million-dollar spend that you can generate at $100,000 spend; the ROAS will definitely be higher at lower levels. Let me demonstrate this using the below table and graph.
For this example’s sake, we will start by spending $1,000 per day and keep ramping up to see where we hit the peak in terms of ROAS. We keep scaling up by $1,000 per day and realize we can continue to increase the ROAS, and we hit the peak at $5,000 with a 2x ROAS. For simplicity’s sake, let’s assume the ROAS keeps falling by 0.10 for every $1,000 increase in spend.
Let us assume the target ROAS for a company is 1x, so the company can continue to spend in marketing at any spend level that brings it a return of above 1x.
In this case, the company could continue to spend $15,000 in marketing as it is generating $15,000 in returns and it is meeting its goal. But is that efficient? Here is how this chart looks when we graph it:
We can clearly tell by the chart above it is not efficient to spend the maximum amount. As you can see, there is a point where the incremental spend does not bring you incremental returns — that is, where the “Return” graph starts declining and dips below the spend line. In this case, we can clearly see the level is $12,000.
At this level of spend, the ROAS on differential is still positive, which means we continue to generate a positive return in this bucket than we were in the previous bucket. If you look at this simplified graph, we will actually be generating the same amount of money at $13,000 then we will at $12,000. Why spend the extra $1,000 at all then? And then the gap just keeps getting wider after that; it shows we will generate the same amount of return at $14,000 than we were at $11,000, and so on.
How to Adjust Marketing Spend to Be More Capital Efficient
Early on, we learned we had to cut off spending at the right inflection point, or we could end up wasting thousands of dollars a month with no marginal benefit. In this fictional example, a $12,000 spend with 1.3x return seems to be the best strategy for our example company.
Don’t just look at the overall goal and targets and continue to spend at those levels without digging in each campaign and determining what is our inflection point.
But this table is not as simple as it seems. To really increase marketing efficiency, it’s important to look at the ROAS differentials, which tell us how much incremental revenue we are getting out of the additional spend.
Specifically, note the inflection point of $8,000. At this level, we are still generating an additional $1,000 for every additional $1,000 spent. In the above case, at $12,000 level, we were generating only $200 for every $1,000 spent — but when calculated on an overall level, we were still generating 1.3x return.
If the company’s strategy is to be more conservative, we could look at the $8,000 level and realize that after this level, for every $1,000 we spend we will actually be generating less than the spend. We could also determine that our optimal spend level is at $8,000 with 1.7x return — not the $12,000 level, which only generates 1.3x return.
While this is a simple example, the underlying principle is substantiated by data my team has accumulated over the last three years. Many companies will have this data, but it’s my goal with this post to encourage them to study that data from a new perspective.
A version of this post originally appeared on TechCrunch, here.