Optimizing remarketing campaigns for an online store

A task – optimization of Google Ads remarketing campaigns to increase the number of orders while maintaining the set price per order.

Complexity – in addition to Google Ads remarketing, the site uses remarketing / retargeting launched in other services, i.e. you need to get sales growth from Google Ads remarketing and, at the same time, prevent drawdowns on orders in other systems that use remarketing / retargeting.

There was also a trend that the launch of Criteo retargeting negatively affected Google Ads remarketing.

ANALYSIS OF EXISTING ADVERTISING CAMPAIGNS

Previously, your account had already run dynamic remarketing ad campaigns for the standard lists that Google recommends for this type of campaign.

But based on historical data, there was a colossal decline in them, and if in the middle of the year it was like a usual “off-season”, then towards the end of the year the situation did not begin to improve (given the forced stop of advertising in October for some time).

It should also be said that in the middle of the year, campaigns were split separately into mobile traffic and PCs. But according to the results of this experiment, we did not manage to get any special positive changes (considering that only adaptive banners worked in the campaigns).

Orders for mobile traffic were only marginally cheaper in comparison with PC traffic, and this did not particularly affect the total number of orders.

ANALYSIS OF USER BEHAVIOR

To optimize remarketing campaigns, we analyzed user behavior before making a purchase, namely:

• the number of days before the purchase since the first visit to the site.

• the number of website visits prior to making a purchase.

• Analysis of ad platforms with the highest conversion rate for remarketing in active ad campaigns.

Based on the results of the analysis, it was decided to focus in Google Ads on those buyers who were ready to place an order within a week from the day of the first visit to the site, as the most interested in buying (judging by Google Analytics data).

The rest of the visitors, whom we might not “catch”, had to interact with Criteo remarketing.

During the launch phase, 6 audiences were created: three for each time period for three different stages in the sales funnel. Also, the previous time range was excluded from each subsequent step. Bids for each audience were set based on the “stage” of the funnel and the time from the last visit to the site, ie. the fewer days since the last visit and the hotter the user was, the higher the rate was.

Pay attention! In this case, dynamic remarketing was configured exclusively using the Google Ads pixel, and not Google Analytics, in order to avoid a delay of up to 24 hours in the transfer of data between systems – this is extremely important when working with user audiences in short time intervals.

The campaigns were launched in the “preseason”, ie. – late January-early February. Separately, we have left campaigns with standard dynamic remarketing audiences, with “new” audiences excluded from them.

RESULTS

So, according to the results of the first month of the test, remarketing to “standard” audiences has shown its complete uselessness:

• the number of orders decreased by one third (33.33%);

• the cost per order increased by 81%;

• the number of clicks increased by 23.5% (pre-season – traffic growth on the site);

• income was down 46%.

• Conversion rate dropped by 46%.

That is, despite the fact that demand and, accordingly, traffic began to grow, the number of orders decreased, and significantly.

At the same time, due to new audiences, in general, the results of remarketing campaigns have improved:

• Increase in the number of orders by 300%;

• Reduced order value by 47%;

• general growth of other indicators.

Of course, one could blame everything on the “preseason”. Therefore – the experiment was continued until mid-March, after which the original campaigns with standard audiences were stopped due to their complete uselessness.

For clarity, below are graphs of comparison for the year, with old campaigns and after creating new segmented ones.

However, these changes did not affect the performance of Criteo’s remarketing campaigns.