DoubleClick for Publishers

Lines 3-15 is boiler-plate that asynchronously loads the GPT JavaScript library used by DFP, using SSL/HTTPS if necessary. This is where the command queue is built which handles the list of functions (generally, ad calls) to be handled asynchronously. Publishers will not need to make changes to this section of the code.

(Optional) Lines 16-20 set page-level attributes for ad slots that serve AdSense. If set here, this will override any styles set in the DFP or AdSense. This is not commonly used since you can set this in DFP. For more information about how to set AdSense parameters in GPT, please review the documentation for the .set() function in the Help Center.

Lines 17, 22, 39, 46: If using asynchronous rendering, wrap all javascript -- following the initial boilerplate script -- in the googletag.cmd.push function on line 17. This will add the function to the command queue to be processed asynchronously with respect to the page load. Read more about the differences between synchronous and asynchronous rendering.

Line 23: "/1234/travel/asia/food" specifies the network code (1234) and targeted ad unit (travel/asia/food). You can find your network code in the "Admin" tab of DFP.

[728, 90] sets the ad slot creative size. Indicate here which creative size or sizes will be eligible to serve to this ad slot. Indicate multiple sizes using the syntax: [[width1, height1], [width2, height2], [width3, height3]]. All sizes here should also be associated with the targeted ad unit -- this will just allow you to narrow the list down based on the specific slot.

Lines 23, 26, 37, 40, 44, and 47: "div-gpt-ad-123456789-0" is how we match the ad slots defined in the head to the ad slots on the page (the div tags in the body where the creatives serve). They can be named anything so long as they match, but our tag generator uses "div-gpt-ad-[random number]-0", "div-gpt-ad-[random number]-1" and so on. This is just a tidy way of making them similar but differentiated.

While random numbers are used here, this is not how GPT uniquely identifies an ad request. That is done behind the scenes with the GPT library. Additionally, these names may be kept the same from page to page so long as there aren't multiple instances of the same div name on the same page.

Lines 25, 28, 29, and 30: Set slot-level key-value targeting with the .setTargeting() function. You can associate multiple values with one key, as in the first example: ("key". ["value1", "value2", "value3"]). If you want to target multiple keys, call the function multiple times as shown in the second case (gender=male and age=20-30). For more information about how to set key-value parameters in GPT, please review the documentation for the .setTargeting() function as well as our article on "Setting targeting and sizes with GPT". While random numbers are used here, this is not how GPT uniquely identifies an ad request. That is done behind the scenes with the GPT library. Additionally, these names may be kept the same from page to page so long as there aren't multiple instances of the same div name on the same page.

Line 30: googletag.pubads().setTargeting("topic","basketball"); sets the page-level key-value targeting. Note that when custom targeting is configured using page-level custom targeting, all ad slots will inherit this custom targeting. Like slot-level custom targeting, you can associate multiple values with one key: ("key". ["value1", "value2", "value3"]). For more information about how to set key-value parameters in GPT, please review the documentation for the .setTargeting() function as well as our article on Setting targeting and sizes with GPT.

Line 31: googletag.pubads().enableSingleRequest(); enables Single Request Architecture (SRA). Include this line if you want to call all ad slots on the page in one call (allows for guaranteed roadblocks and, by consolidating ads into one call, may have performance benefits for your page). Why would anyone not use SRA? Currently, SRA does not support DoubleClick tag (InRed tag) rewind, nor Google Programmable Ads (GPA).

Enable Synchronous rendering for the ad slots below (elements following ad calls will not load until the creative has fully rendered for that call). If you do not include this function the loading of body elements will not be blocked on the ad slots before them.

Line 29: (Optional) style="width: 728px; height: 90px" is the size that the element will take before the creative is rendered. If using multi-size tags, it's best to omit this (in that case, the size of the element will take on the size of the selected creative once rendered) or -- if you choose to use it -- make both dimensions large enough to contain the largest eligible creative. But for single size ad tags, this can be used to expand the container element until the creative loads so that other page elements do not shift when the creative renders.

Lines 3-10 is boiler-plate that asynchronously loads the GPT JavaScript library used by DFP, using SSL/HTTPS if necessary. This is where the command queue is built which handles the list of functions (generally, ad calls) to be handled asynchronously. Publishers will not need to make changes to this section of the code.

(Optional) Lines 11-13 set page-level attributes for ad slots that serve AdSense. If set here, this will override any styles set in the DFP or AdSense. This is not commonly used since you can set this in DFP. For more information about how to set AdSense parameters in GPT, please review the documentation for the .set() function in the Help Center.

Lines 15 and 18: "/1234/travel/asia/food" specifies the network code (1234) and targeted ad unit (travel/asia/food). You can find your network code in the "Admin" tab of DFP.

[728, 90] sets the ad slot creative size. Indicate here which creative size or sizes will be eligible to serve to this ad slot. Indicate multiple sizes using the syntax: [[width1, height1], [width2, height2], [width3, height3]]. All sizes here should also be associated with the targeted ad unit -- this will just allow you to narrow the list down based on the specific slot.

Lines 15, 18, 29, 31, 34, and 36: "div-gpt-ad-123456789-0" is how we match the ad slots defined in the head to the ad slots on the page (the div tags in the body where the creatives serve). They can be named anything so long as they match, but our tag generator uses "div-gpt-ad-[random number]-0", "div-gpt-ad-[random number]-1" and so on. This is just a tidy way of making them similar but differentiated.

While random numbers are used here, this is not how GPT uniquely identifies an ad request. That is done behind the scenes with the GPT library. Additionally, these names may be kept the same from page to page so long as there aren't multiple instances of the same div name on the same page.

Lines 17, 20, 21, and 22 (Optional) Set slot-level key-value targeting with the .setTargeting() function. You can associate multiple values with one key, as in the first example: ("key". ["value1", "value2", "value3"]). If you want to target multiple keys, call the function multiple times as shown in the second case (gender=male and age=20-30). For more information about how to set key-value parameters in GPT, please review the documentation for the .setTargeting() function as well as our article on "Setting targeting and sizes with GPT". While random numbers are used here, this is not how GPT uniquely identifies an ad request. That is done behind the scenes with the GPT library. Additionally, these names may be kept the same from page to page so long as there aren't multiple instances of the same div name on the same page.

Line 22: googletag.pubads().setTargeting("topic","basketball"); sets the page-level key-value targeting. Note that when custom targeting is configured using page-level custom targeting, all ad slots will inherit this custom targeting. Like slot-level custom targeting, you can associate multiple values with one key: ("key". ["value1", "value2", "value3"]). For more information about how to set key-value parameters in GPT, please review the documentation for the .setTargeting() function as well as our article on Setting targeting and sizes with GPT.

Line 23: googletag.pubads().enableSingleRequest(); enables Single Request Architecture (SRA). Include this line if you want to call all ad slots on the page in one call (allows for guaranteed roadblocks and, by consolidating ads into one call, may have performance benefits for your page). Why would anyone not use SRA? Currently, SRA does not support DoubleClick tag (InRed tag) rewind, nor Google Programmable Ads (GPA).

Line 24: To render the creatives synchronously, add googletag.pubads().enableSyncRendering();. This instructs the service to wait for each ad request to complete before continuing with rendering the page.

Line 29: (Optional) style="width: 728px; height: 90px" is the size that the element will take before the creative is rendered. If using multi-size tags, it's best to omit this (in that case, the size of the element will take on the size of the selected creative once rendered) or -- if you choose to use it -- make both dimensions large enough to contain the largest eligible creative. But for single size ad tags, this can be used to expand the container element until the creative loads so that other page elements do not shift when the creative renders.