How to ensure a prebid.js on a new site is working correctly with Prebid Analytics
With Prebid Analytics by Roxot, you can test new sites, bidders, config updates, placements, and more without waiting days for the data to appear. Moreover, you will test on real users in contrast to dev tools consoles where you can only test on yourself. This provides more realistic results as your site, bidder, or placement is being tested in real visitor's environment.
When testing a prebid.js setup on a new site with Prebid Analytics, make sure you have completed all analytics installation steps: added the site to the system, implemented a new prebid.js with Prebid Analytics adapter to the site, and updated the site config to activate the tool.
Start your QA process with Real-time Dashboard that shows events across your websites as they happen. In the drop-down menu choose a site you want to test.
Pay attention to Total Requests and Total Impressions. Total Requests metric indicates how many ad requests are sent from your site. If the number of requests is growing, the chances are your site sends requests properly. Total Impressions metrics, in turn, shows you the number of rendered prebid ads. Growing number of impressions means that bids are sent to your ad server, some of them outbid other demand sources such as AdX, and ads are rendered correctly. On the screenshot below, there were 0 impressions for 2543 requests. The problem might be with the ad server & prebid.js integration. In this case, double check prebid line items in your ad server & run the test again.
After conducting general tests, dig deeper and check if all ad units and bidders work correctly. Go to the Site Dashboard and scroll down to the Performance by Ad Unit widget. Make sure all of the configured ad units are displayed in the widget. If an ad unit is missing, most probably, it's configured incorrectly. Next, check Requests and Fill Rate for each ad unit. The requests metric will show you if an ad unit sends requests to your bidders. Fill Rate is a percentage of successfully conducted auctions (auctions with at least one bid). It indicates how bidders respond to ad requests. As you can see on the screenshot, ad-unit-bottom does send requests but its Fill Rate is 0% - there are no bids to run an auction. When only one ad unit has 0% fill rate, check how it's configured.
The final step in testing your prebid.js setup on a site is to ensure that every bidder works as expected. If you are still on the Site Dashboard, scroll up to the Bidders Performance widget and analyze Requests, Bid Rate, and Impressions. Again, 0 under Request means that requests are not sent to the bidder. Check the config on your site. If bidder's Bid Rate is 0%, contact your account manager and work on troubleshooting together. If you see impressions for the site and every ad unit, most probably you won't have any issues with bidder's ads being rendered.
Make a habit of checking how your site performs. Frequent examinations will help you quickly spot an issue and avoid losing revenue due to configuration errors.
The first prebid analytics tool for more efficient ad operations
Made on