eCommerce Case study: 51.2% increase in sale with the same ad budget on Google Ads
Next, we will present a broad optimization steps we have taken on our client's Google Ads account.
The goal our client set for us was to reduce costs and increase ROAS from promotion to over 10.
By the way, ROAS means Return On Ad Spend, which means Return on Advertising Budget, or how much money you make from sales compared to how much money you pay on your ads. For example, if you make sales of $10,000 from paying $1,000 on ads, your ROAS is 10,000 / 1,000 = 10. Essentially on a ROAS of 10 you earn $10 lei for every $1 spent on advertising. A nice and predictable money printing machine, right?
And since we work on performance, we only get paid if we can optimize your account to a point where it is really profitable and you are happy to pay our performance percentage.
Read till the end see if we got paid or not :)
Fast forward to the end result:
Cost decrease by 51.2%. Here's what the chart looks like:
Prior to optimization, the client account had an average conversion cost of approx. $22.44:
After optimization, the cost per conversion decreased to $11.26 so 51.2% lower:
Done with the math? Let's move on.
As for ROAS was around 5X:
And after optimization it increased to over 11:
In fact, increasing the ROAS (AKA the profitability of the account) looks like this:
It is important to mention what happened to the conversions.
We can see here just a slight decrease in conversions which tells us we've successfully eliminated only the keywords that were under performing.
But in our case the conversions had a very slight decrease, which shows that only those conversions that were unprofitable were reduced:
How did we achieve these optimizations?
We have applied our battle tested automated rules to achieve the following outcomes:
- Increase the bids on the words that perform but are not displayed on the first positions
- Lowering bids on words that don't convert and costs a lot of money
- Lower the bids on the words that convert, but they are not profitable
I also wanted the bid increase and decrease automation to be proportional with the performance of the words, So for a word with ROAS of 7 (below target) I lowered the bid less than i would for a word with ROAS of 2. So the worse it performs, the more I decrease the bid.
The first decision we took was around the time interval during which we should evaluate this keywords. Because performance varies according to any given period of time.
Below is an example of a report tracking the cost per conversion (from another account). Note how the cost differs depending on the time interval:
This type of overview allows us to evaluate the keywords at their true performance over different time intervals. I will explain why this is important.
We wanted to see what words perform and to what extent, over different periods of time. The better a word performs over a period of time, the more we will increase its bid. The poorer it performs over several periods of time, the more we will lower the bid.
By "time interval" we mean the standard Google Ads intervals, namely 7 days, 30 days, 90 days.
To identify how the words performed on each interval, we used a labeling system developed in house. For example, if a word had a ROAS lower then 7 for the last 7 days we apply a label with "O: ROAS <7 (7 days)".
If the same word has ROAS of <7 and 30 and 90 days, it will still be labeled with "O: ROAS <7 (30 days)" and with "O: ROAS <7 (90 days)", respectively .
In addition to ROAS, we put labels for several combinations of metrics and intervals, 63 in total. this is our secret sauce that allows for no human error and makes us so confident working on a performance structure exclusively.
At the same time, this labeling system allowed us to see how much we spent on words that under performed in the last 7, 30 or 90 days.
If we have a high cost on a word that is not profitable enough neither for 30 days nor for 90 days, we will reduce the bid for that keyword the most. This was the first optimization step and as a result we've seen the biggest cost savings.
A keyword may have performed well for 90 and 30 days, but poorly over the last 7 days. Which case we wouldn't decrease the bid here. - it may be a temporary variation, for example the week before the salary.
Or maybe it went wrong for 90 days, but good for 30 and 7, in which case we would raise the bid by a small percentage because the recent performance is more relevant than the distant one, but still we will take note that previously performed poorly and monitor it closer as we go.
Then over the next few weeks we've increased and decreased bids slowly based on the percentage tags we've applied to each keyword analyzed this way.
The following is an example of a bid decrease percentage applied to keywords with under benchmark ROAS:
- -15% for ROAS <7 in the last 90 days
- -15% for ROAS <7 in the last 30 days
- -15% for ROAS <7 in the last 7 days
So if a word has become unprofitable at all intervals, I dropped its bead by approx. 45% (-15% for each interval). If it also had a high ad placement on google's first page , equal or higher to 2, I reduced the bid by another 15%. So a total decrease of the bid by approx. 60%.
Why so much?
Because the higher the positions the more they spend. And if they are also under performing, over multiple intervals, this are the first words that need to be optimized, and must be optimized most intensively.
If a keyword has become unprofitable and it displays on position 1.5 (average), that keyword is wasting budget both because it is not profitable and because it has many clicks due to the high position on the page.
Therefore we reduced the bids for each under performing metric we've found. Three times for the three intervals of poor performance and once for the high position. Four elements of non-performance = four bid reductions.
As a result of the bid reduction, expensive clicks have dropped considerably and the cost per click reduced to half:
We did the opposite for the words that were profitable and shown on a lower page position, to which we increased the bids on several occasions. So 15% increase for each positive metric they had. These actions led to profitable conversions that have very well offset the unprofitable conversions lost by our initial decease of the bids.
What other effects did the optimizations have?
At an account level, this graph shows the best how other metrics aligned (blue ROAS, yellow cost per conversion):
We conclude with the most important outcome of this optimizations:
The most important outcome, beyond the specific Google Ads metrics, is the increase of sales in the first month post-optimization by $12,700, keeping the optimized level of cost per conversion. As a percentage, the growth is 51.62% compared to the period before the optimization.
$12,790 was the new revenue generated of this increase in sales. Ad it was pure profit for our client, because he did not spent any extra money to get it:
The increase in sales resulting from the optimized results will be invested in the development of the account and a fraction of it to reward our performance.
We love creating more win win situations for potential customers. if you want similar results, let's schedule a call and see if we can help.