Get your Education tracking to 95%+ accuracy guaranteed in writing.
Universities, independent schools, bootcamps and online academies all run the same broken loop: paid media optimizes on RFI form fills, the application happens on a portal or on Common App where no pixel reaches, and enrollment gets confirmed in the SIS months later. I connect inquiry, application, deposit and census-date enrollment into one instrumented funnel with values that reflect what each program is actually worth.
Free30 minutesNo obligation
Education · what breaks
Why enrollment marketing loses the trail — Education
- 01
The application lives where your pixel doesn't
Common App, a hosted Slate portal, or an SIS behind single sign-on. The browser session ends at the handoff, and so does any chance of browser-based attribution.
- 02
Every program gets the same conversion value
A nursing master's and a two-week certificate both report as one RFI, so bidding drifts toward whichever program produces the most cheap inquiries rather than the most net tuition.
- 03
The lag outruns every attribution window
An undergrad inquiry in September enrolls the following August. No ad platform window spans that, so the platforms optimize on the only thing they can see, which is form fills.
- 04
Melt is invisible to the ad accounts
Deposits get reported as enrollments in the marketing deck, then a share of those students never appear at census. That gap never travels back into bidding, so it repeats next cycle.
Education · the fix
What changes when the SIS defines the conversion — Education
One funnel, six named stages
Inquiry, application started, application submitted, admitted, deposited, and enrolled at census. Each is its own event with its own value instead of a single lumped lead.
Program-level values in the bid
Expected net tuition for that program times the historical yield from that stage, so a graduate program outbids a short certificate because it genuinely should.
Off-domain applications still attributed
The click ID captured at first touch rides into Slate or your CRM as a hidden field, so a Common App submission weeks later still imports against the right campaign.
Intake planning based on real yield
You can see which channels produced applicants who actually matriculated last cycle, not which produced the most inquiries during the October push.
Education · architecture
Inquiry, application, deposit, enrolled at census — Education
Three systems own this funnel: a CRM for inquiries, a hosted or third-party portal for applications, and the SIS for enrollment. We stitch them with a click ID and reconcile to the registrar.
- 1
Click ID captured on the RFI, before the handoff
gclid, gbraid, wbraid, msclkid and fbclid land in hidden fields on the request-for-information form and onto the inquiry record at creation.
- 2
The CRM carries attribution off-domain
Slate, Element451, Salesforce Education Cloud or Enrollment Rx holds the click ID, so a Common App submission weeks later still has something to match on.
- 3
Portal and apply-subdomain hops repaired
Cross-domain linking across apply.yourschool.edu and hosted portals, with identity carried into the application record instead of restarting the session.
- 4
The SIS decides who actually enrolled
Banner, Colleague, Anthology Student, Jenzabar or PowerSchool at census date is the enrollment record. A paid deposit is not enrollment.
- 5
Program-level values go back out
Expected net tuition for that program times the historical yield from that stage becomes the value carried on each stage event.
Education · events
Six stages, and one optimization target per program type — Education
Application submitted is usually the bidding target; enrolled is reporting, except for bootcamps and short programs where it lands in window.
| Event | Fires when | Key parameters | Sent from / role |
|---|---|---|---|
| rfi_submitted | A request-for-information form posts | gclid, em (hashed), ph (hashed), program_group, event_id | Browser + server, deduped by event_id · reporting only |
| event_registered | An open house, info session or webinar registration confirms | gclid, em (hashed), event_type, event_id | Server · reporting only |
| application_started | An application is created in Slate, the portal or Common App | gclid or hashed contact, program_group, application_id | Server · optimization (secondary) |
| application_submitted | The application is marked submitted in the CRM or portal | gclid or hashed contact, value (program EV × stage yield), currency, conversion_date_time | Server offline import · optimization (primary) |
| admitted | The admission decision is released | gclid or hashed contact, value (program EV × yield), currency | Server offline import · reporting only |
| deposit_paid | The enrollment deposit or seat fee posts | gclid or hashed contact, value (expected net tuition), currency, transaction_id | Server offline import · reporting only |
| enrolled_census | The SIS confirms enrollment at the census date | gclid or hashed contact, value (net tuition), currency, transaction_id | Server offline import · reporting; optimization for short programs |
| melt | A deposited student does not appear at census | application_id, reason_code | Held in your systems · reverses the deposit value |
Education · prove
Reconciled to the registrar, one intake cycle at a time — Education
The registrar's census file is the number. Everything upstream of it is an estimate until it has been matched against that file.
Source of truth
Enrolled-at-census records in Banner, Colleague, Anthology Student, Jenzabar or PowerSchool — the registrar's file
- Pull the census enrollment file for the intake with program, start term and the originating inquiry or application ID.
- Join back through the CRM to the click ID captured at RFI, and report the unmatched share by program and channel.
- Compare application_submitted and enrolled counts to platform-reported conversions by conversion date.
- Measure melt — deposits minus census enrollments, per program — and adjust the stage values before the next intake.
- Confirm no education-record field, aid status or specific program of study left the container. Group codes only.
- Check that K-12 and parent-facing properties are still measurement-only, with remarketing disabled.
Education · free audit
What we check in the free audit
- The application lives where your pixel doesn't
- Every program gets the same conversion value
- The lag outruns every attribution window
- Melt is invisible to the ad accounts
Checked against
Enrolled-at-census records in Banner, Colleague, Anthology Student, Jenzabar or PowerSchool — the registrar's file
Free30 minutesNo obligation
Education · when not to buy
When an institution shouldn't build this — Education
Four honest reasons to hold off.
- You fill cohorts from organic search, word of mouth and school counselors, and paid is a small experiment. Fix form tracking and stop.
- One program, one intake a year, one price. Program-level values have nothing to differentiate and a single lead action reports it fine.
- Your SIS team won't release a census extract, even aggregated. Without the registrar's number there is nothing to reconcile against.
- Admissions and marketing don't agree on what counts as an inquiry. Settle the definition first; the schema only encodes what you decide.
Education · what we deliver
What we build for an enrollment funnel — Education
Six-stage enrollment event schema
Each stage defined and documented with the system that owns it and the exact rule for when it fires, so marketing and admissions are reading the same funnel.
Slate, Element451 or Education Cloud pipeline
Source, click ID and campaign written onto the inquiry record at creation, then stage changes pushed back out to the server container as they happen.
SIS reconciliation against Banner, Colleague or Anthology
Enrolled-at-census records matched against platform-reported conversions, so the number in the deck comes from the registrar rather than the ad account.
Program-level value table
Expected value per program per stage, built from your last two cycles of yield and net tuition, then refreshed each intake as the mix shifts.
Cross-domain and portal handoff repair
Click IDs and session identity carried across apply.yourschool.edu, hosted application portals, and off-domain applications you don't control.
Minor-audience and consent handling
Consent Mode v2 wired, remarketing scoped so under-18 audiences aren't built, and tag behavior split between K-12 or parent-facing sites and adult program pages.
Written guarantee
95%+ tracking accuracy
Measured against Enrolled-at-census records in Banner, Colleague, Anthology Student, Jenzabar or PowerSchool — the registrar's file. Miss it and we keep working at no cost.
FAQ
Education questions, answered
For browser tracking, yes — you can't put a tag on commonapp.org. What works is capturing the click ID at the inquiry stage on your own site, storing it on the Slate or CRM record, and importing the submission and eventual enrollment as offline conversions against it later.
FERPA governs education records, so once someone is a student the constraints tighten. In practice we send hashed contact identifiers and stage names, never program-of-study detail, financial aid status, or anything drawn from an education record. Practitioner guidance, not legal advice — have your registrar and general counsel review the field list.
Personalized advertising to minors is restricted on Google and Meta, and remarketing lists that could include under-18 users are a problem. We keep K-12 and parent-facing properties on a separate tagging setup with remarketing disabled and conversion measurement only.
It will if last cycle's data drives this cycle's peak unchecked. We keep stage events consistent year over year so the platforms have comparable history to work from, and plan the ramp deliberately instead of restarting learning at the start of every intake.
On the traditional undergrad timeline, no — use it for reporting. Application submitted is usually the right bidding target, because it lands weeks after the click and correlates with matriculation. Bootcamps and short online programs are different: enrollment is close enough to the click that you can often optimize on it directly.
That's the normal case: a CRM for inquiries, a hosted portal for applications, the SIS for enrollment. Deciding which system is the source of truth at each stage and reconciling the differences is the whole job — it's why the Source-of-Truth Reconciliation Method exists.
If you fill your cohorts from organic search, word of mouth and school counselors, and paid is a small experiment, don't build this yet. A single-intake independent school with a modest budget is usually better off fixing form tracking and stopping there. This is for institutions running multiple programs and real media spend.
More on Education
- GuideOffline Conversion Tracking: Send Closed Deals Back to Google Ads and Meta
- GuidePhone Call Conversion Tracking: Stop Losing Your Best Leads to the Phone
- EducationConversion tracking for education providers, measured in enrollments
- Google Ads · Application → enrollmentHow to track application to enrollment for Google Ads
- Meta · Application → enrollmentHow to track application to enrollment for Meta Ads
Free · 30 min
Connect your ad spend to census-date enrollment
WhatsAppFree30 minutesNo obligation
