Simulation
Cross-channel memory
A scripted path from an ad to this site, an email, and a return. See what is carried and what is deliberately forgotten.
Understanding the labHow to use it and why you'd want to
Why you'd want to
A scripted ad, site, email, and return show which categories are carried and which ids are forgotten. Nothing here is a live profile, and after Accept the page queues lab_interaction.
How to use it
- Step through the fixed ad, site, email, and return.
- A category can carry forward while an id is dropped.
- The story stays inside the script. It does not read this visit.
- After Accept, a view is queued, then a demo click.
Simulation. A scripted journey. The carried fields are categories. Raw ids and the email address are marked forgotten on purpose. Nothing is sent to LinkedIn or an inbox.
Journey
LinkedIn ad
Channel class: paid social
Carried
Raw click id
Deliberately forgotten
Advertising id
Deliberately forgotten
Why am I seeing this? Paid social is a channel class, not a click id. Measurement is a topic, not a page URL with parameters. The email step does not keep an address. The return visit still does not restore the forgotten ids.
WalkthroughHow we built itShowHide
Example vendor setup is marked below. eVar and event numbers are examples. This lab does not call Adobe.
The cross-channel memory lab walks a scripted ad, site, email, and return, showing which categories are carried and which ids are forgotten, and queues lab_interaction after Accept. The lab does not call Adobe.
Look up a fixed step
memoryAt returns the rows for ad, site, email, or return. The rows are constants. The function does not read this browser.
Carry a class, drop an id
Channel and topic can be carried. The raw click id, the advertising id, and the email address are marked forgotten. The return step does not restore them.
Stay inside the script
The banner says the journey is scripted. Nothing is sent to LinkedIn or an inbox. The buttons only change the step on the page.
Queue a view, then a demo click
useLabView queues lab_interaction with action view and label page. A step button calls trackLab with action run_demo and the step id as label. Both go through trackLabEvent, which does not push until consent is granted.
From the code
The ad and site rows. Forgotten means the id is not kept.
const ROWS: Record<ChannelStep, MemoryRow[]> = {
ad: [
{ id: "channel", state: "carried" },
{ id: "click-id", state: "forgotten" },
{ id: "ad-id", state: "forgotten" },
],
site: [
{ id: "channel", state: "carried" },
{ id: "topic", state: "carried" },
{ id: "click-id", state: "forgotten" },
{ id: "ad-id", state: "forgotten" },
],
email: [The view fires once after Accept. A demo click uses the same event name.
export function useLabView(labId: string) {
const { consent } = useConsent()
const sent = useRef(false)
useEffect(() => {
if (sent.current || consent !== "granted") return
sent.current = true
trackLabEvent("lab_interaction", { lab_id: labId, action: "view", label: "page" })
}, [consent, labId])
}
export function trackLab(labId: string, label: string) {
trackLabEvent("lab_interaction", { lab_id: labId, action: "run_demo", label })
}Google and Adobe
Google queues lab_interaction after Accept. Adobe is an example of the same fields.
Queued only after Accept. The first push is lab_interaction with lab_id cross-channel-memory, action view, and label page. A step button queues the same event with action run_demo and label ad, site, email, or return. Carried and forgotten fields are not parameters. This site does not call Adobe.
Simplified shape of the object queued after Accept. Register any new parameter as an event-scoped custom dimension. This site does not call gtag('set').
dataLayer.push({
event: "lab_interaction",
lab_id: "cross-channel-memory",
action: "view",
label: "page",
})Example implementation. This site does not call Adobe. eVar62, prop62, and event62 are example variable numbers. In Launch, a data element reads the value and a rule sends it when the value changes. In the report suite, enable eVar62 with Most Recent allocation and expire it on the hit. A processing rule is unnecessary if the data element maps straight to that eVar.
Example alloy payload. _example is a placeholder schema, not a real tenant.
alloy("sendEvent", {
xdm: {
eventType: "web.webinteraction.linkClicks",
_example: {
channelMemory: "paid-social",
},
},
})Example hit. eVar62 is hit-scoped in this example. prop62 dies at the same hit.
s.eVar62 = "paid-social"
s.prop62 = "paid-social"
s.events = "event62"
s.linkTrackVars = "eVar62,prop62,events"
s.linkTrackEvents = "event62"
s.tl(true, "o", "lab_interaction")Translation
Each lab field, the GA4 name, and the example Adobe variable. Scope is the part that usually gets lost in the translation.
Lab id
GA4. Parameter lab_id on lab_interaction. The value is cross-channel-memory.
Adobe. Example eVar62, expire on the hit.
Scope. Event on the hit. It names the page, not a user, and it does not cover the visit.
Action
GA4. Parameter action. view on load, run_demo when a step button is pressed.
Adobe. Example prop62, which dies at the hit.
Scope. Event. Do not store view or run_demo as a user property for the visit. It is an event on the hit, not a user property, and it does not last for the visit.
Label
GA4. Parameter label. page for the view, or the step id for a demo click.
Adobe. Example eVar62 if you mapped the step. The carried channel is not this field.
Scope. Event on the hit. The step label is not a user trait and it ends with the hit, not the visit.
Forgotten id
GA4. Not sent. The row state stays on the page.
Adobe. Example prop62. Do not put a click id there. Not sent.
Scope. Not an event, not a user property, and not a hit or visit field. The script marks it forgotten.
Recommended next
Why these picksReading path transitions…