Infinite Scroll

Overview

Introduction

This page explains how to integrate Taboola recommendations into a web page with infinite scroll (aka continuous scroll), using the param values provided by Taboola.

πŸ“˜

Infinite scroll (aka continuous scroll)

This integration applies to any page in which the user can continuously scroll for additional articles (even if the number of articles is limited).

🚧

Prefilled scripts

If you received prefilled scripts from Taboola, you can copy and paste them 'as is', instead of editing param values.

🚧

New integration?

If this is a new integration, please provide Taboola with a timeframe for launch, so we can plan post-launch QA.

Terminology & flow

πŸ“˜

Taboola loader

Loads your publisher-specific configurations and fetches your Taboola recommendations.

In order not to impact page load time, loader.js is loaded asynchronously.

πŸ“˜

The _taboola command queue

A global array whose existence is ensured by the following code:

window._taboola = window._taboola || [];

You communicate with the Taboola loader by 'pushing' commands to this queue.

This allows you to pass commands 'safely', regardless of whether the Taboola loader has loaded yet.

πŸ“˜

General flow

The initial page

  1. Add the Taboola loader
  2. Pass the page details to the _taboola command queue.
  3. For each placement on the page:
    1. Create a <div> container with a unique ID.
    2. Pass the placement details to the _taboola command queue.
  4. Pass the flush command.

Each new page

For each new page that scrolls into view:

  1. Pass the page details to the _taboola command queue.
  2. For each placement on the page:
    1. Create a <div> container with a unique ID.
    2. Pass the placement details to the _taboola command queue.
  3. Pass the flush command.

Do not pre-fetch recommendations - wait for the new page to scroll into view.

πŸ“˜

Initial page vs subsequent pages

The integration steps for the initial page and subsequent pages are almost identical.

Supported browsers

🚧

Taboola's integration is supported for any browser that is fully compatible with ES5 (ECMAScript 5) or higher. This includes all modern web browsers.

Older browsers (such as IE 11) may encounter problems with video ads.

The initial page

Add the Taboola loader

🚧

The Taboola loader should run once only, when the initial page is loaded.

Fill in your own Publisher ID, as provided by Taboola (or use the prefilled script that was sent to you).

Add the following

<script>

    // *Global* command queue for the page
    window._taboola = window._taboola || [];

    !function (e, f, u, i) {
        if (!document.getElementById(i)) {
            e.async = 1;
            e.src = u;
            e.id = i;
            f.parentNode.insertBefore(e, f);
        }
    // Fill in your Publisher ID (an alphabetic string), as provided by Taboola:
    }(document.createElement('script'), document.getElementsByTagName('script')[0], '//cdn.taboola.com/libtrc/<publisher-id>/loader.js', 'tb_loader_script');

    if (window.performance && typeof window.performance.mark == 'function') {
        window.performance.mark('tbl_ic');
    }

</script>
ParamInstructions
<publisher-id>Fill in your Publisher ID (an alphabetic string), as provided by Taboola.

For example, if your Publisher ID is '<<publisherID>>', then the URL should read:

//cdn.taboola.com/libtrc/<<publisherID>>/loader.js

Pass the page details

Pass the page details to the _taboola command queue:

// *Global* command queue for the page
window._taboola = window._taboola || [];
  
// Pass the *page* details to the command queue:
_taboola.push({<page-type>:'auto', url:'<url>'});

πŸ“˜

The values shown below are sample values only. Make sure to fill in your own values, as provided by Taboola.

ParamInstructionsNotes
<page-type>Replace this with the relevant page type (as provided by Taboola) - e.g. for article, use:
article:'auto'
Possible values:
video, article, photo, search, category, home.

----
Taboola optimizes recommendations based on page_type. Make sure that the page-type correctly reflects the current page type.
'auto'To obtain an automated ID, leave this param 'as is':
article:'auto'
Specifies the ID of this page.
<url>Pass the fully-qualified, canonical URL of the initial page.E.g. If the current URL is:

https://www.example.com/news/456/new-york-jets?filter=sports&utm_source=taboola

Then pass:

https://www.example.com/news/456/new-york-jets

🚧

Guidelines

  1. Use your own param values, as provided by Taboola.
  2. Pass the fully-qualified, canonical URL of the initial page.
  3. If you copy the above code snippet to another page, make sure that page_typecorrectly reflects the new page type.
  4. Do not pre-fetch recommendations for pages further down in the continuous scroll.

Pass the placement details

For each placement on the page:

  1. Insert a <div> container, in the location that the placement should display:
<div id="<container_id>"></div>

πŸ“˜

Assign each <div> an ID that is unique across the entire continuous scroll. - e.g.:

container_id = "<<containerId>>-" + <some_unique_value>;


Tip: Consider using an ID that is similar to the placement name.

  1. Pass the placement details to the Taboola command queue:
// *Global* command queue for the page
window._taboola = window._taboola || [];

// For each placement, pass *your* param values, as provided by Taboola:
_taboola.push({mode: '<mode>', container: '<container_id>', placement: '<placement>', target_type: '<target_type>'});

πŸ“˜

Params

The values shown below are sample values only. Make sure to fill in your own param values, as provided by Taboola.


  • mode - The UI Mode ID for this placement, as provided by Taboola:

    E.g. '<<mode>>'

  • container - The ID of the <div> (chosen by you) that will contain this placement's content:

    E.g. '<<containerId>>-1'

    Must be unique across the entire continuous scroll.
    Consider using an ID that is similar to the placement name.

  • placement - The placement name, as provided by Taboola:

    E.g. '<<placementName>>'

    Must be unique within a given article - but repeated for each article in the continuous scroll.

  • target_type - The target type, as provided by Taboola:

    E.g. '<<targetType>>'

    Taboola optimizes recommendations based on target_type.
    Do not update this value, unless your Taboola Account Manager instructs you to do so.

πŸ“˜

Custom Segments

To target a custom segment - or simply track its performance - you can optionally pass a cseg param.

For more information, see Custom Segments - or reach out to your Taboola Account Manager.

🚧

Guidelines

  1. Use your own param values, as provided by Taboola.
  2. The ID of the <div> must be unique across the entire continuous scroll.
  3. The placement must be unique within a given article - but repeated for each article in the continuous scroll.
  4. Do not update target_type, unless your Taboola Account Manager instructs you to do so.
  5. Do not pre-fetch recommendations for pages further down in the continuous scroll.

Flush the queue

Immediately after passing details for the last placement, pass the flush command:

// *Global* command queue for the page
window._taboola = window._taboola || [];

// Pass the `flush` command:
_taboola.push({flush: true});

🚧

  • The flush command instructs Taboola to fetch any placements in the command queue (and not wait any further).
  • Pass the flush command once only - after passing all placements for that page.
  • If you have unique requirements, reach out to Taboola to discuss alternative flows.

πŸ‘

Need a hand?

Feel free to reach out on our Community Discussion page!

A subsequent page

For each new page that scrolls into view, repeat the process.

Pass the page details

// *Global* command queue for the page
window._taboola = window._taboola || [];
  
// Pass the *page* details to the command queue:
_taboola.push({<page-type>:'auto', url:'<url>'});
ParamInstructionsNotes
<page-type>Replace this with the relevant page type (as provided by Taboola) - e.g. article:

article:'auto'
Possible values:
video, article, photo, search, category, home.

----
Taboola optimizes recommendations based on page_type. Make sure that the page-type correctly reflects the current page type.
'auto'(Recommended)
To obtain an automated ID, leave this param 'as is':
article:'auto'
----
Else, insert your own, internal ID:
article:'123'
Specifies the ID of this page, for the given page type.
<url>Pass the fully-qualified, canonical URL of the new page.E.g. If the current URL is:

https://www.example.com/nfl/123/nfl-roster?filter=news&utm_source=taboola

Then pass:

https://www.example.com/nfl/123/nfl-roster

🚧

Guidelines

  1. Use your own param values, as provided by Taboola.
  2. If you copy the above code snippet to another page, make sure that page_typecorrectly reflects the new page type.
  3. Pass the fully-qualified, canonical URL for each page, once the new page scrolls into view - e.g. when you are ready to update the address bar with a new URL.
  4. Do not pre-fetch recommendations for pages further down in the continuous scroll.

πŸ“˜

Best practice

Each time a new page scrolls into view, update the address bar with a new URL

Pass the placement details

For each placement on the page:

  1. Insert a <div> container, in the location that the placement should display:
<div id="<container_id>"></div>

πŸ“˜

Assign each <div> an ID that is unique across the entire continuous scroll. - e.g.:

container_id = "<<containerId>>-" + <some_unique_value>;


Tip: Consider using an ID that is similar to the placement name.

  1. Pass the placement details to the Taboola command queue:
// *Global* command queue for the page
window._taboola = window._taboola || [];

// For each placement, pass *your* param values, as provided by Taboola:
_taboola.push({mode: '<mode>', container: '<container_id>', placement: '<placement>', target_type: '<target_type>'});

πŸ“˜

Params

The values shown below are sample values only. Make sure to fill in your own param values, as provided by Taboola.


  • mode - The UI Mode ID for this placement, as provided by Taboola:

    E.g. '<<mode>>'

  • container - The ID of the <div> (chosen by you) that will contain this placement's content:

    E.g. '<<containerId>>-1'

    Must be unique across the entire continuous scroll.
    Consider using an ID that is similar to the placement name.

  • placement - The placement name, as provided by Taboola:

    E.g. '<<placementName>>'

    Must be unique within a given article - but repeated for each article in the continuous scroll.

  • target_type - The target type, as provided by Taboola:

    E.g. '<<targetType>>'

    Taboola optimizes recommendations based on target_type.
    Do not update this value, unless your Taboola Account Manager instructs you to do so.

πŸ“˜

Custom Segments

To target a custom segment - or simply track its performance - you can optionally pass a cseg param.

For more information, see Custom Segments - or reach out to your Taboola Account Manager.

🚧

Guidelines

  1. Use your own param values, as provided by Taboola.
  2. Do not update target_type, unless your Taboola Account Manager instructs you to do so.
  3. Pass the new placement details once the new page scrolls into view - e.g. when you are ready to update the address bar with a new URL.
  4. Do not pre-fetch recommendations for pages further down in the continuous scroll.

🚧

Updating the address bar

If you update the address bar each time a new page scrolls into view, then the following best practice applies:

  1. When updating the browser history, make sure to replace the existing record with the new URL.
  2. Adding new records to the browser history makes the Back button awkward to use. As a result, users are unlikely to trigger Taboola Explore More.

For more information, reach out to your Taboola Account Manager.

Flush the queue

Immediately after passing details for the last placement, pass the flush command:

// *Global* command queue for the page
window._taboola = window._taboola || [];

// Pass the `flush` command:
_taboola.push({flush: true});

🚧

  • The flush command instructs Taboola to fetch any placements in the command queue (and not wait any further).
  • Pass the flush command once only - after passing all placements for that page.
  • If you have unique requirements, reach out to Taboola to discuss alternative flows.

πŸ‘

Need a hand?

Feel free to reach out on our Community Discussion page!

What's Next?

  1. In order to provide personalized recommendations for impressions in the EU and California, make sure to implement GDPR and CCPA respectively.
  2. If you have not already done so, configure ads.txt.
  3. Contact Taboola so that we can verify your implementation.