Description of the situation

The client of iMarketings.lv is an e-commerce company that sells LED products in several countries of the European Union through various paid advertising channels. One of them is Meta (Facebook) Ads, which accounts for a fifth or ~20% of the total traffic each month.

Comparing the results of Meta (Facebook) and Google Ads, we observed that Meta (Facebook) Ads provides a lower cost per acquisition (CPA), while Google Ads provides a higher average order value (AOV).

We considered it a challenge and wanted to achieve the highest purchase value in Meta (Facebook) advertising as well. So, we started an experiment, the results of which we would like to share. In this case, our goal was to increase the average order value (AOV) of Meta (Facebook) Ads using the features offered by Meta (Facebook) Ads Manager.

Solution

To improve the average order value from Meta (Facebook) advertising, we planned to reach Meta (Facebook) users who are more likely to purchase more expensive and/or multiple products, which would lead to a higher average order value. Such users can be reached through value optimisation. This technological capability is able to identify and reach customers who on average spend more than other users.

This client already had several audiences set up on their Meta (Facebook) Ads account, which were performing well over time. In particular, they always achieved the client-defined target KPI (ROAS 300%), one of which was a lookalike 10% audience made up of existing buyers (purchase optimisation). In order to find out whether value optimisation (optimisation of the average order value) leads to an increase in AOV, we decided to use this particular audience for the experiment. To perform the experiment, we duplicated this audience and set it to value optimisation, resulting in two identical audiences whose only difference is the optimisation (conversions and value).

Result

The period of the conversion vs value optimisation experiment was 1 month. The experiment was conducted using a lookalike audience of 10% and a daily budget of € 25 for each optimisation. In this case we did not use high budgets. The aim was to conservatively assess whether the value optimisation would provide a positive result.

The following were the results after 1 month:

By using value optimisation, we have achieved great results: the average order value (AOV) increased by 93% and ROAS by 58% compared to conversion optimisation.

Conclusions

Our experiment showed that value optimisation does reach users who are more likely to buy more expensive and/or more items, as demonstrated by the significant growth of AOV. By using value optimisation, we not only gained a higher average order value (AOV), but also the ability to further increase budgets for both the audience used in the experiment and other audiences to ensure even more customer growth.