Task: to set up contextual advertising in Google Ads
Business field: coding courses that include practice with a real-world project
Duration: 6 months (Dec 2020 – May 2021) (compared to previous period, Jun – Nov 2020)
Total advertising budget: 143,000 UAH
- Very interesting and rather distinctive project. The company has developed proprietary methodology. No theory, only a real-world project.
- The Google Ads account had various search campaigns with different geotargeting for every course. DNS campaigns were also separated by geography and by course.
- Low conversions (phone calls and submitted forms). Before our work, they had received 206 requests with a price of 363 UAH per conversion in half a year. The critical business goal was to increase the number of customers.
Goal: Increase the number of conversions
- Analyzed where conversions came from. After examining the sources, we found out that the majority of conversions were generated by Dynamic Search Campaign and Branding.
- Improved campaigns that had been working well. And reset the campaigns that hadn’t been giving results. We redistributed the budget and mostly allocated it to effective keywords and campaigns.
Firstly, we collected from all campaigns (incl. DSA) the keywords and search queries that had promising conversions. Then, we added them into a separate campaign with the maximum budget, using the maximum conversion value bid strategy. For other keywords, a small budget and manual bids were set. The first campaign brought conversions, but they were expensive and did not meet our expectations.
- Started actively testing various campaigns and strategies. In this way, we set up numerous interesting strategies:
- Added groups with audiences by interest (pulled from Google Analytics), special audiences by keywords (included selected keywords and search queries that gave conversions) to the Display Network. We also tried various geotargetings. As a result, conversions increased from 3 to 4. And the price per conversion dropped to 891 UAH (777 UAH less than in the previous period).
As a test, we decided to try the new (at that time) type of Google campaign called Discovery.
The essence of the Discovery Ads format: This tool enables to show effective personalized ads to the users who are ready to interact with the ad and the brand at the moment when they are just surfing the Internet. For that, Google uses machine learning to determine when users are most likely to engage with a brand. In other words, these clickable image ads are shown to an audience that is potentially interested in ordering a product or service without even knowing it.
After all, that appeared to be the best solution.
In half a year, we received 593 conversions for 59.29 UAH per each.
We created and tested visual materials for different categories of the target audience (general, for women, for men, for people who want to change their profession, etc.). We also tested various audiences: website and YouTube visitors, interest-based and event-based audiences.
And after an active search, we found the ideal strategy – ‘audience-creative’. The most effective option was displaying carousel ads shown to the visitors to sites similar to ours.
Discovery campaigns can be a remarkably successful customer acquisition tool, but you need to test and select different visuals and audiences to find the best combinations.
As a result, we achieved the goal: we increased the number of conversions by 277.63% (on average from 34 to 129 conv / month) with reducing the cost of conversion by 49.31% (from UAH 363 to UAH 184 / conv).