Average ROAS comes with certain risks when you split your budget among your marketing channels. See here how to solve them with 3 realistic examples.
Calculating the average ROAS comes with certain risks and limitations when you split your budget among your different marketing channels.
This article aims to explore these limitations and offer hands-on solutions. By the end of the article, you'll know the following:
- The importance of ROAS in your marketing performance.
- How to calculate it to steer your advertisement budget.
- The difference between average ROAS and incremental ROAS. And what it means for your marketing budget.
- 3 realistic scenarios on decision-making based on ROAS and incremental ROAS.
Why is ROAS a good marketing metric: an example
In recent years, advertisers started embracing attribution modelling to make their investment decisions more data-driven and accountable. With unified measurement techniques, we can create a data-driven foundation for strategic and tactical budget allocation and media scenario planning.
And ROAS is one of the greatest tools for data-driven budget optimisation. Indeed, Return on Ad Spend, commonly known as ROAS, is one of the most used marketing metrics. Especially when it comes to marketing budget decisions. Even more than ROI (Return on Investment) when we focus on ad costs and returns.
While ROI is still relevant for your business, it cannot help you optimise your marketing activities on a campaign level. It's more like the bigger picture of how Revenue compares to your general marketing spend.
But when making campaign budget decisions, you need a more accurate metric that connects business value to your specific ad campaigns. This way, you have the correct data to split your budget among your channels.
Attribution results can be a significant first step towards data-driven media planning. However, using only the historical media impact and effectiveness is not enough.
The following stylised (but realistic) example showcases how the abovementioned works in practice. Assuming we have a budget decision to make for an upcoming campaign. For this campaign, we need to decide on a budget split between the following channels: TV and (Online) Video.
To decide on this budget split (in a data-driven way), we look at the historical attribution results for comparable campaigns:
The historical campaign results suggest TV and Video, on average, had the same investment levels during these historical campaigns. However, the average Revenue generated by Video is higher (€6.550) compared to the average Revenue generated by TV (€2.400). This also makes the ROAS for Video (€2,63) higher than the ROAS for TV (€0,96).
So, the average ROAS for Video is higher than the average ROAS for TV. That means that Video always performs better than TV for this example. Right? WRONG!
The risks of average ROAS: how to avoid the "flaw of averages" when making budget decisions.
When a marketer looks at this example, it seems reasonable to invest more towards the investment that has the higher ROAS. And this is a classic example of what we call the "flaw of averages".
Sam Savage (Harvard Business Review) suggests that "plans based on assumptions about average conditions usually go wrong". In a humorous example, Savage narrates the story of a statistician who drowned by crossing. This happened because he calculated that the river's depth was, on average, three feet.
This humorous story really outlines the limitations of averages in statistics. And how making decisions based on average values can have serious consequences. In that sense, ROAS is not any different.
In our example, this translates to the assumption that investing more in Video is a good decision. Because video has a higher ROAS on average. But again, this is not a wise choice. Actually, we need to investigate historical investments and analyse their incremental impact on the KPI of interest.
Calculate ROAS with a sensitivity analysis of your investments.
When you run your advertisement budget based on the average ROAS, you assume each future investment will have a similar ROAS. Moreover, this suggests a linear relationship between these investments and the generated Revenue. And you will continue to have the same impact (return) for each additional euro you spend, no matter what.
In reality, we often observe non-linear relationships between investments and Revenue. So, it is fundamental to investigate historical investment levels and compare these to the observed incremental Revenue. To do so, we reconstruct investment sensitivity curves to represent the expected Revenue for different investment levels.
We use investment sensitivity curves to spot where additional investment has the most impact. And this should be the basis for scenario analysis calculations. Incremental ROAS calculations also detect the budget sweet spot for future investments. Investment sensitivity curves are also known as 'shape effects' or 'investment elasticity curves'.
How to optimise your marketing budget with incrementality testing.
Going back to our example, we have created the following investment sensitivity curves:
Investment sensitivity curves provide the expected Revenue for different investment levels. Looking at the Video curve of our example, we observe a diminishing curve. The first investments are relatively effective. However, when the budgets increase, we expect a diminishing return.
Looking at the TV curve, we observe an S-shaped curve. This needs an initial investment level to start making a revenue impact. After this, the channel becomes more efficient until an investment level of around €3.000. Above this level, the channel also shows a diminishing effect.
These curves align with the patterns we often measure in channels like TV and Video. For more traditional channels such as TV and Radio, minimum investment levels are often required to observe an impact. Moreover, these channels usually do not allow investing in smaller budgets.
To better understand this, let's deep-dve into the investment sensitivity in our example with 3 possible scenarios. Assume that you need to split the budget between TV and Video for an upcoming advertising campaign. How would you use ROAS for channel investments?
Scenario 1: historical budget split
In the historical investment levels, we observed a budget split of €2.500 in TV and €2.500 in Video:
This scenario is marked by the plotted purple bullets on the graphs. The expected Revenue is visible on the chart too.
In this scenario, we would expect a return in Revenue of €2.400 for TV and €6.550 for Video. And it would result in a total revenue of €8.950. The ROAS in this scenario would be €1,79, considered a good ROAS.
Scenario 2: increased Video spend
In the second scenario, we switch the budget split between TV and Video. Specifically, we move €500 of the TV budget towards Video. This scenario results from the "flawed" conclusion that we must put the budget into the channel with the highest average ROAS. In this case, the scenario would result in a budget split of €2.000 in TV and €3.000 in Video:
This scenario is marked with the plotted purple bullets on the graphs. And the arrows indicate the budget shifts compared to the historical budget split.
Video increased the investment to €3.000 (+€500), resulting in an expected revenue of €6.850. While the TV investment decreased to €2.000, resulting in an expected revenue of €800.
The total Revenue of this scenario would be €7.650. And the ROAS would be €1,53. This ROAS is lower than the one of the initial investment.
Scenario 3: increase TV spend
In the third scenario, we switch the budget split between TV and Video. In this case, we move €500 of the Video budget towards TV. This is the scenario that would result from analysing the investment sensitivity curves. This scenario would see a budget split of €3.000 in TV and €2.000 in Video:
This scenario is marked with the plotted purple bullets on the graphs. The arrows indicate the budget shifts compared to the historical budget split.
In the scenario, we increased the TV investment to €3.000 (+€500), resulting in an expected revenue of €4.900.
Hence, we decreased the Video investment to €2.000, resulting in an anticipated revenue of €6250. This scenario results in a total revenue of €11.150. And a ROAS of €2,23. This is the highest ROAS in all 3 scenarios.
If you compare the above three scenarios, you see they all had the same total investment. Using the historical budget split, we achieve a revenue of €8.950.
Investing more towards video (the channel with the higher ROAS) seems logical. Although, the investment sensitivity analysis suggests otherwise. Particularly, we found that Video is already quite saturated within these spend levels. And therefore, this scenario results in a 15% decrease in Revenue compared to the historical budget split.
On the other hand, the TV channel is not yet saturated on these spend levels. So, there is still room for additional investment. This is why investing more in TV would lead to a potential increase of 25% in Revenue for our example.
Takeaway: split your advertising spending using incremental ROAS
Working for leading advertisers, we have learned that attribution results can be a significant first step towards data-driven media planning. However, we are not quite there yet with only the historical media impact and effectiveness. To further optimise your advertisement budget, you need to use marketing metrics optimally.
The article's example shows that the average ROAS can be a "flawed" metric. And it can lead to poor budget decisions for your future campaigns. Instead, investment sensitivity provides a better overview of the incremental impact of each of your future investments.
In this example, analysing investment sensitivity curves led to a potential increase of 25% in Revenue. While ignoring the analysis could lead to a possible decrease of 15% in Revenue.
Of course, ROAS is not the only metric that can help you optimise your marketing budget. Other factors can play a major role too.
Your industry is one of them. For instance, there are distinct steps you would follow to optimise the budget for insurance brands. In contrast, e-commerce brands should approach their marketing budget differently.
Would you like to know more about optimising your marketing spend? See how predictive marketing helps you forecast your campaign performance and revenue.
