This package provides functionality to display ads and track user behaviour for ft.com.
This package is available on npm.
npm install --save @financial-times/ads-display
If you need to get some metadata about the page or user, you can use getAdsData()
.
import { getAdsData } from '@financial-times/ads-display'
getAdsData({
user: true, // fetch data about current user
page: { // fetch data about current article or stream page
type: 'article' // 'article' or 'stream'
id: 'id-of-the-page' // id of the stream or article page. Usually found in appContext
}
}).then(adsData => {
// Do stuff with the data.
})
Developer note: You can also override the host and version of the ads-api endpoints which this function fetches from by providing a host
property. This is useful if you want to test locally.
getAdsData({
...,
host: 'http://local.ft.com:3002/v2'
})
The default host is https://ads-api.ft.com/v2
See Client-side API section for more details
Both client and server side components for displaying ads on a page.
You can define ad slots and pass data to the client for initialising the ads library
import ReactDOM from 'react-dom/server'
import { Slot, AdsOptionsEmbed } from '@financial-times/ads-display'
app.use(appContextMiddleware)
app.get('/my-page-with-ads', (req, res, next) => {
const adOptions = {
appName: res.locals.appContext.appName,
adUnit: ['ft.com', `articles/${req.articleId}`],
// page level targeting, added to every ad request on the page
targeting: {
user: {},
article: {}, //or
stream: {}
},
...
}
const slotProps = {
name: 'leaderboard',
formatSmall: false,
formatsLarge: 'SuperLeaderboard,Leaderboard,Responsive',
formatsExtra: 'Billboard,SuperLeaderboard,Leaderboard,Responsive',
// slot level targeting, added only for this ad slot
targeting: {
specialAdTargeting: true,
pos: 'top'
}
}
const adHTML = ReactDOM.renderToStaticMarkup(
<AdsOptionsEmbed options={adOptions} />
<Slot {...slotProps} />
)
res.render(adHTML)
})
If you haven't defined your ad slots on the server, you will first need to define one or more ad slots on the page. You can learn more about this here
Once you have your ad slots defined in your HTML, you will need to initialise the ads library.
This is the minimum you need to get some ads.
import { displayAds } from '@financial-times/ads-display'
import * as flags from '@financial-times/anvil-ui-ft-flags'
const flagsClient = flags.init()
const options = {
appName: 'my-app'
}
displayAds.init(options, flagsClient);
On ft.com, we will want to pass in a page rootID to each ad call. ads-display exports some utility functions, amongst which a method getRootId()
which can be called when initialising the ads.
import { displayAds, adsUtils } from '@financial-times/ads-display'
import * as flags from '@financial-times/anvil-ui-ft-flags'
const flagsClient = flags.init()
// For a pageKit app, we can get information from appContext
const { appName, abTestState } = appContext
const options = {
appName,
abTestState,
rootId: adsUtils.getRootID()
}
// On ft.com we only show ads if the flag is enabled
if(flags.get('ads')) {
displayAds.init(options, flagsClient);
}
In most use cases on ft.com and the FT app, we will want to get some data about the user and the page and append this to our ad requests to provide more targeted ads.
import { getAdsData, displayAds } from '@financial-times/ads-display'
import * as flags from '@financial-times/anvil-ui-ft-flags'
const flagsClient = flags.init()
const { appName, contentId, abTestState } = appContext
if(flags.get('ads')) {
getAdsData({
user: true,
page: {
type: 'article',
id: contentId
}
}).then(adsData => {
const options = {
appName,
abTestState
targeting: adsData.metadata, // for all pages on ft.com
adUnit: adsData.adUnit, // for article or stream pages
};
displayAds.init(options, flagsClient);
});
}
Here is an example of using ads-display on the ft.com article page. This ties in all parts of ads-display, by first fetching some data, then enabling both the display ads and the behaviour tracking with permutive
import {
getAdsData,
adsUtils,
displayAds,
adsPermutive
} from '@financial-times/ads-display'
import * as flags from '@financial-times/anvil-ui-ft-flags'
// Assume pagekit app where appContext is defined.
const flagsClient = flags.init()
if(flags.get('ads')) {
getAdsData({
user: true,
page: {
type: 'article', // or 'stream'
id: appContext.get('contentId')
}
}).then(adsData => {
// initialise ads
displayAds.init({
appName: appContext.get('appName'),
abTestState: appContext.get('abTestState'),
rootId: getRootId(),
targeting: adsData.metadata,
adUnit: adsData.adUnit,
}, flagsClient);
if (flags.get('AdsPermutive')) {
// Initialise permutive
adsPermutive.setup();
adsPermutive.loadPermutiveScript();
// Identify the user
adsPermutive.identifyUser({
uuid: adsData.metadata.user.uuid,
spoorId: adsData.metadata.user.spoorId
})
// Send a page view event
const pageView = adsPermutive.fromAdsDataToPageView({
...adsData.metadata,
rootId: getRootID(),
type: appContext.get('appName')
});
adsPermutive.trackPageView(pageView);
}
});
}
These are all the options you can pass to displayAds.init()
server side via AdsOptionsEmbed
-
appName
(required)<String>
- Name of the app -
adUnit
: (optional)<Array>
- An array to use for the GPT ad unit. Each item in the array will be concatenated with a/
. E.g.options.adUnit = ['ft.com', 'UK']
becomesft.com/UK
. If not provided, will default tounclassified
. -
adsLayout
: Used for setting thenLayout
value in the ad call, useful for ad ops for targeting purposes, to test different types of ads formats and for setting up demos -
abTestState
(optional)<String>
- The state of all the mvt tests running on the page. Usually found inappContext
for pageKit apps. -
rootId
(optional)<String>
- The root id of an ft.com page. You can get this using thegetRootId()
function. -
targeting
(optional)<Object>
- An object containing the metadata from the ads data manager to be used as targeting for the ad calls. -
sandbox
(optional)<Boolean>
- Determines wether to load test ads from the origami ad unit in Google Ad Manager -
disableMonitoring
(optional)<Boolean>
- Whentrue
,ads-display
will not send any ads lifecycle metrics to Spoor API. -
formats
(optional)<Object>
- Object containing custom ad slot formats. For example:{ MyAdSlot: { sizes: [[970,400], [970,250]] } }
-
lazyLoadMargins
(optional)<Object>
- Object containing lazy loading margins to apply below a certain viewport side. For example:{ 1000: '10%', // if the viewport is smaller than 1000px then apply a 10% margin 400: '5%' // if the viewport is smaller than 400px then apply a 5% margin }
-
waitForMoat
: (optional)<Boolean>
- Will wait 1 second max for the moat prebid api to be loaded before initialising any ad slot so that we can get invalid traffic data with the ad requests
Permutive is a behaviour tracking tool that is used to segment people based on their activity on the site. See adsPermutive for more details. It is proxied here by ads-display for convenience. Here's an example:
import { adsPermutive } from '@financial-times/ads-display';
adsPermutive.setup();
adsPermutive.loadPermutiveScript();
adsPermutive.identifyUser({
guid: '1234',
spoorId: '12345'
});
adsPermutive.trackPageView({
page: {
type: 'article',
rootId: '1234'
}
});
For ft.com, there is a helper function that calls all the functions, with some data about the page and user (usually fetched from getAdsData()
), type and rootId.
import { adsUtils } from '@financial-times/ads-display';
adsUtils.enablePermutiveFtCom({
metadata: { user: {...}, page: {...} }, // data from `getAdsData()`
type: 'article',
rootId: adsUtils.getRootID()
})
import { displayAds } from '@financial-times/ads-display
This function initalises o-ads, loads the required third party scripts, and scans the page for ad slots to initialise
Check weather the request is deemed as valid traffic (using an external script from moat). Talks directly to the global window.googletag
object to set an extra targeting parameter m_data
= 0
(valid) or 1
(invalid) for every ad call. Filtering of ads is done on the Google Ad Manager side based on these targeting parameters. There is a race condition between displayAds.validateTraffic()
and displayAds.init()
so the first ad call may not have the right value for the m_data
parameter.
-
deferScript
: Defer loading of the moat script, true by default -
callback
: Called when the moat script is loaded
import { adsUtils } from '@financial-times/ads-display
Returns the root id of the current ft.com page
A helper function to run all the necessary functions for loading and sending data to permutive on ft.com pages.
import { getAdsData } from '@financial-times/ads-display
or to minimize bunle size (depending on your build config), you can import directly:
ES Module: import { getAdsData } from '@financial-times/ads-display/dist/esm/adsData
CommonJS: import { getAdsData } from '@financial-times/ads-display/dist/cjs/adsData
Returns a data object in the following format:
{
metadata: {
user: {
// key -> value user props
},
article: {
// key -> value article page props
}
stream: {
// key -> value stream page props
}
},
adUnit: [] // ad unit for stream or article pages,
}
Format the user data as key => value pairs for use in ad requests. The keys are mapped from the ads-api /v2/user response to the equivalent keys that the ad ops server expects in the ad calls.
import { formatUserData } from '@financial-times/ads-display
or to minimize bunle size (depending on your build config), you can import directly:
ES Module: import { formatUserData } from '@financial-times/ads-display/dist/esm/adsData
CommonJS: import { formatUserData } from '@financial-times/ads-display/dist/cjs/adsData
getAdsData({ user: true }).then(adsData => {
const keyValuePairs = formatUserData(adsData.metadata.user);
...
})
Format the article data as key => value pairs for use in ad requests. The keys are mapped from the ads-api /v2/user response to the equivalent keys that the ad ops server expects in the ad calls.
import { formatArticleData } from '@financial-times/ads-display
or to minimize bunle size (depending on your build config), you can import directly:
ES Module: import { formatArticleData } from '@financial-times/ads-display/dist/esm/adsData
CommonJS: import { formatArticleData } from '@financial-times/ads-display/dist/cjs/adsData
getAdsData({
page: {
type: 'article',
id: appContext.get('contentId')
}
}).then(adsData => {
const keyValuePairs = formatArticleData(adsData.metadata.article);
...
})
import { adsPermutive } from '@financial-times/ads-display
Exposes all the public functions from adsPermutive
AdsOptionsEmbed
is a JSX component that renders a script with the passed options embedded in a json blob, they are then retrieved automatically client side by ads-display
- options ads-display options
Slot
is a JSX component that renders an Ad slot compatible with o-ads
-
name
<String>
: Ad slot name -
formats
<Array|String>
: Available formats to use -
targeting
<Object>
: Slot level targeting object ex:{ pos: 'top', mvtGroup: 'specialChristmasAd'...}
-
formatsDefault
<Array|String>
List of formats to use as a default -
formatsSmall
<Array|String>
List of formats to use for small sized screens -
formatsMedium
<Array|String>
List of formats to use for medium sized screens -
formatsLarge
<Array|String>
List of formats to use for large sized screens -
formatsExtra
<Array|String>
List of formats to use for extra large sized screens -
classNames
<Array|String>
CSS class names -
style
<String>
: Custom CSS styles
Slots
props for the most common ads on FT sites such as Leaderboard, mid, right handrail
The following o-ads
events will be tracked:
page-initialised
slot-requested
slot-rendered
slot-collapsed
This tracking can be disabled through the disableMonitoring
option.
Each of those o-ads
events will be sent as an oTracking event to Spoor API and will contain several time marks as specified in the metrics config object alongside additional properties to identify the associated slot. The events will be sent to Spoor for all users.
Ads tracking relies on third party scripts to add extra tracking, reporting, counter bot-traffic measure and more.
It's the client site google library that o-ads is built on top of to load ads on the site
Where?
GPT is inserted by o-ads
, which is a dependency of this package
Why? It is necessary to load ads
Moat is an analytics tool that has an invalid traffic functionality, brand safety and extra tracking that google doesn't get
Where?
Moat is inserted on the page when nAds.init()
is called
Why? As part of our ad offering, these functionalities are required
Permutive is the behaviour tracking software that segments users based on their actions on site.
Where?
permutive is inserted by @financial-times/adsPermutive
if the adsPermutive
flag is on
Why? As part of our ad offering, these functionalities are required
If you are developing on ads-display, you may find it useful to run the demo page. This mocks a consuming app where you would normally import ads-display from.
make demo
- run a demo server on http
make demo-https
- run the demos on https. This is useful if you want to test under https://local.ft.com
to make sure there are no cross origin request errors.
There are two demos available:
- promise-demo.html
- async-await-demo.html
These are examples of how to implement ads-display using async/await method or promises.
In order to release a new version please refer to the page-kit release guidelines