Contributors : @wolfgang.feger
Determine which channels and ad tactics are effective for your business.
Define the strategic and tactical success metrics for your paid ads experiments and use them to specify the experiment goals.
Focus on the strategic goals that are critical from your business perspectives such as transactions, leads, revenue, or ROI. Use bounce rate as the secondary tactical goal to gauge the quality of the traffic. Use engagement metrics such as CTR to evaluate the ad copy and the graphic assets in your ads.
Decide which channels you would like to test and set up test campaigns.
Google Search Ads is a great channel for testing as it allows capturing user intent, always include it in your experiment plan. Facebook and Instagram Ads are great tools for proactively targeting users and should be considered in your testing efforts. If you have a B2B product or service you should consider a test LinkedIn campaign.
Create the test campaign in each of the channels.
Allocate a similar budget to all the channels.
Test the copy of your Google text ads with two versions of the ad texts, focusing on different approaches, for example, emotional vs informational.
Collect at least 100 impressions and evaluate the ads. Focus on the click-through rate and conversion data. Pick the winner and make a variation of the winning ad:
- Fine-tune the winning ad instead of creating a completely new message.
- In this phase test one element at a time - for example, a headline or description line.
Repeat the process to generate further learnings.
Test the copy and graphic assets in your Instagram, Facebook or LinkedIn ads to learn which ads generate the best engagement.
Start with testing different design assets; test two different images or an image versus a video format. Determine the winning graphic format based on CTR and conversion data. Test winning graphic assets with two different ad copy versions starting with two very different approaches. Once you determine the winning copy, fine-tune it with additional tests. In this phase test one element at a time for example, test two different headlines and leave everything else the same.
Test different landing pages in all your channels to reduce the bounce rate and improve the conversion rate.
Link two identical ads to two different landing pages. Some ideas for testing:
- Product category page vs product detail page.
- Homepage vs category page.
- Different variations of a product detail page set up in an A/B testing tool.
Determine the winning variation focusing on the conversion data and bounce rate.
Experiment with different audiences and keywords to optimize your targeting.
Create two different campaigns targeting different audiences with the same ad. For example, you might use Facebook interest targeting and test different user interests. You might also test interest targeting versus custom or lookalike audiences. You might want to test different age groups and can also target two different geographic locations in two different campaigns. On Google, you can target the same ad with different keywords.
Compare the test campaigns across channels to inform the strategic decisions about your future marketing channel mix.
Focus on the primary success metrics to determine the most successful channels in the campaigns. Scale the campaigns in the channels with the best performance and conduct further small-scale tests in channels with worse performance. Don’t invest in channels that prove ineffective in further tests.
Use the insights from your tests to adjust and optimize your ad tactics.
Target the keywords, audiences and geographic locations that generated the best ROI. Use the copy and graphic images that generated a higher CTR. Use the landing pages that generated the lower bounce rate and better conversion rate.