«That's what we've always done.»
What if we tell you that there is a more profitable solution?
How do Norwegian online shops track profitability on digital advertising today?
ROAS is a reasonably known concept for most people who work in online commerce and advertising of products on the web.
ROAS, or “return on ad hand”, tries to track the profitability of paid activities. This is done online by looking at how much we get left for the ad crowns used.
The biggest challenge of using the “return on ad-pend” as a KPI for success is that for most people it is very challenging to measure this very thing. Where we want to measure “return on ad sling”, many of us measure “revenue on ad sling”.
By applying to revenue, or sales revenues, from Google Analytics, for example, we simplify what most people actually are a very complex calculation.
What is a profitable ROAS?
That's the big question.
Many people believe that a high ROAS is a profitable ROAS--the higher ROAS is the, the more money we earn. But is that really true?
If we look at the table below, most people might agree that: Product B is much more profitable than Product ASo let's say that this is the same thing as that. We spend the same amount of money promoting both of these. Product A and B, but the sales revenue varies greatly.
Obviously Product B much more profitable than Product A- What? It has a much higher ROAS!
Or...?
This is where it's easy to get tricked. And when you look at the table above, you're just going to ignore an important part of the math.
This is very rare that it is as simple as sales revenue that corresponds to actual income. For example, see the table below:
Product A and B may have different purchase prices, pick-up costs and freight costs. And this is two very different things. margins. So if we take this into account in our math, it might look pretty different.
Product B might no longer look so profitable? It's despite a high ROAS.
We see, however, that we make money selling product A -- despite a lower ROAS.
We can therefore call out the “myth” that a high ROAS is always a profitable ROAS.
How do companies set ROAS targets?
They optimize for an average ROAS
Many companies try to solve the above challenge by appreciating something.
By looking at their entire range of products and calculating the coverage of the different product categories or branded products, the company can know about the revenue on ad-hand they need to go to zero with the advertising of these products.
A company may know that they need at least $5 in sales revenue for each crown used for advertising bags (500 % ROAS) and at least $2,5 in sales revenue for each crown used for advertising scarf (250 % ROAS).
Based on this, they average ROAS across the different product categories of 375%.
The method is easy to relate to and easy to communicate with partners. We understand why someone chooses this solution -- but did you know that it also poses a great risk? Both to tape money and for: miss potential income.
With an average ROAS of 375%, products will look equally profitable regardless of whether the margin is 40% or 10%. In this case, it may mean that the company potentially loses money when they sell a bag and that they do not sell as many scarf as they could (and still have been profitable).
They set different ROAS targets for different categories
The example above might look a little extreme and “single” but the truth is that this is how a good number of companies operate.
A better and more reflective approach to different margins in a range of products is to set ROAS targets for each category and setting own budgets and objectives, and optimising different categories.
This gives advertisers a more accurate picture of what campaigns are actually profitable and not, but it is under a very important assumption: that your customers actually buy what they click on.
In surprisingly many cases, a online store will find that customers who click on the same product advertisement do not actually end up with the same shopping basket.
Not only do they choose different payment solutions and shipping options (which affect the profitability of the shopping basket), but they also trade different products:
Customer 1 clicks on the ad for the purse, goes right to shopping basket and only trades the purse she clicked on.
Customer 2 clicks on the ad for the purse, but looks further around in the online store and ends up with a completely different product.
Customer 3 clicks on the ad for the purse, looks around and ends up buying both the purse she clicked on and two scarf!
These three customers had exact same starting with, but ended with three different shopping curves with very different profitability for the company. This is not taken into account when we work with ROAS.
POAS: The profitable solution you are looking for
So how do we solve all of these challenges? Tracking profit by order rather than revenue is a good place to start.
The future ad market is marked by increased degree of automation. What is the consequence? The way we work on optimization needs to change.
Good data and communication with advertising systems has always been important, but digital marketing agency We see a clear shift in the importance of data in the future.
With the introduction of black box promotion types such as Performance Max and the data base restriction through third parties, own enterprise data more important than ever.
Those who are left to win in the long term - it is advertisers who clearly manage to use data to communicate their goals to machine learning systems.
So what should you do?
Instead of reporting and optimising on the revenue reports and optimising directly on gross profit and thus POAS (profit on ad hand) instead of ROAS.
Probability tracking in practice
In INEVO we cooperate with ProfitMetrics to base us on profit tracking in advertising and actively use POAS as KPI.
We assume a direct API connection from CMS or an XML feed at all product prices (and any sales prices), as well as purchase prices.
And we're also tracking actual purchases in the shopping basket so that we can deduct extra costs, payment costs, shipping costs etc. for each order. For example:
This gives us a gross profit by order that we can directly feed into the advertising systems like Google Ads and Facebook for further optimization - as well as analytical systems like Google Analytics.
By focusing on POAS you can (or the machine learning) by: just the sum you need to maximise real-time profit rather than base you on an average ROAS.
In the advertising system it will then be extremely clear what campaigns we make money on and where we actually lose money when we deduct the advertising costs from gross profit:
Profit tracking results
What results can you expect from the implementation of profit tracking?
It will of course vary.
Within a month with active optimisation of profit rather than income, one of our customers experienced a 122% increase In the net profit and a 18% decrease in costs.
Profit tracking challenges
POAS is not a magic KPI. And here, too, you'll find weaknesses.
In our opinion, the disadvantages are not enough to avoid making the shift from income focus to profit focus in advertising. Nevertheless, it is important to be aware of the following challenges:
1 - The data you get out is just as good as the data you put in.
This is, after our experience, the largest threshold for companies that want to get started with profit-timulation.
Many companies do not have a good enough overview of the purchase prices and costs associated with their products - far fewer have this information directly in their advertising systems.
A profit focus could be implemented by setting more general rates for both margins and other order costs.
But remember! The data you get out is just as good as the data you put intoSo let's say that this is the same thing as that.
Therefore, we recommend investing in the right data in place.
2 - A shift from income focus to profit focus will require a change in focus within the organisation
In many organisations, growth and success are measured based on income growth.
Therefore, there may be a challenge for you to handle advertising if you report and optimise for profit, while the rest of the organisation focuses on sales revenue.
The increase in profit does not have to be equivalent to an increase in sales revenue and vice versa.
A successful migration from the intext focus to the profit focus requires a change in thinking and focus, across teams and within the organisation.
3 - Switching from the ropeue to profit is a “sock” for machine learning
Smart biting and machine learning work from historical data. When we switch from Income to profit as input for bidding, this will, without exception, give the system a “sock”.
Not only is the profit sum lower than the revenue previously reported, but the POAS target is often much lower than the ROAS target system before has optimised for.
Going from a target ROAS of 900% to 120% will challenge machine learning and therefore an uncomplaint is to make the actual switch in advertising systems.
Our recommendation is to start by collecting data as a secondary goal before choosing to do the shift. You should also prepare for at least one conversion cycle to be over the learning phase.
Get started with POAS
«So, we've always done that, so maybe it's time to try something new?»
Do you think profit tracking and optimization is something for you and your online store? Then it's little to wait for! The sooner you get started, the greater the competitive advantage will be in today's advertising market. We'll help you get started. Contact us!