Ellis Agency. DELIVERABLE · EA-2026-0002
PHASE 1 · KICKOFF & DISCOVERY

Analytics
Measurement Plan

Goals, conversion funnels, audiences & the event taxonomy for the guilford.edu refresh — grounded in a live GA4 baseline and built to prove post-launch lift, with the long-missing giving funnel stood up for the first time.

CLIENTGuilford College
ENGAGEMENTWebsite & Analytics Refresh
PREPAREDJun 2026 · Ellis Agency
STATUSInternal · Confidential
Ellis Agency.At a glance
00

The plan, in one page

This plan layers custom funnels, dashboards & automated reporting onto Guilford's existing Google Analytics — enhance, don't rebuild. It standardizes a messy event setup, stands up the giving funnel that doesn't exist today, and sets pre-launch benchmarks so the refresh's impact is provable.

120k
sessions / 90 days
(live GA4 baseline)
$0
giving conversions
tracked today
6+
standardized key events
(+ micro events)
5
conversion funnels
defined
WHAT THIS DELIVERABLE IS

The measurement plan for the GA4 implementation: goals & KPIs by audience priority, GA4 segments/audiences, the standardized key-event spec, an event taxonomy tied to the page-type templates, the five funnels, the cross-domain/GTM technical plan, and the Looker dashboard + automated weekly/daily reports. Benchmarks come from a live pull of property 359308554 (last 90 days, pulled Jun 22).

Biggest finding

There is zero giving/donation conversion tracking today. Every one of the 7,044 conversions in the baseline is enrollment-flavored (application + RFI clicks). Yet donors are the priority-2 audience and the kickoff wants a far-more-prominent Give CTA. Standing up the cross-domain giving funnel — donation_startdonation_complete — is the marquee new capability of this plan.

CONFIRMED BY THE DATA

The kickoff intuition that ~60% of traffic is direct/repeat holds: Direct is 31.2%, returning sessions 31.2%, and the top pages are dominated by known-audience destinations (/current-students, /faculty-and-staff, /account/login). Do not narrow the site.

Governing directives — every part of the plan honors these

Audience priority — lockedEnrollment + giving firstKeep current students in view Automated weekly/daily reportsEnhance, don’t rebuildSeasonal flexibilityFuture personalization
Guilford College · Analytics Measurement PlanConfidential · Internal02 / 13
Ellis Agency.§1 · Pre-launch baseline
01

Pre-launch baseline (benchmarks)

A live pull of GA4 property 359308554, last 90 days (pulled Jun 22). These are the benchmarks the post-launch plan proves lift against — annotate the launch date in GA4 for a clean before/after.

120,036
sessions
78,234
total users
45.0%
engagement rate
~2:34
avg. session
7,044
conversions
(100% enrollment)

Traffic mix — by channel

ChannelSessionsShare
Organic Search62,21651.8%
Direct37,50731.2%
Display (Encoura)6,1135.1%
Referral4,8564.0%
Organic Social3,7213.1%
Paid Other3,0802.6%

New vs returning

New sessions73,403 · ~61%
Returning sessions37,483 · 31.2%

Device split

Desktop76,683 · 63.9%
Mobile42,393 · 35.3%
READ OFF THE BASELINE

Organic + Direct are 83% of traffic; paid (Encoura display + paid social) is a measured, attributable slice to keep clean. The Direct + returning weight, plus the top-page mix, is the current-student/applicant loyalty signal — the site must keep serving them.

Guilford College · Analytics Measurement PlanConfidential · Internal03 / 13
Ellis Agency.§1 · Pre-launch baseline

Top pages (views / sessions)

PathViewsSess.Role
/44,53737,370Front door — every audience
/current-students15,94612,506Returning / known audience
/faculty-and-staff8,1936,542Faculty & staff wayfinding
/admissions/apply7,8407,378Enrollment funnel
/directory7,6272,184People directory
/program-finder6,1794,854Program interest
/admitted-students4,4683,939Yield / known audience
/account/login4,0272,700Current-student re-entry
/apply/status2,6771,665Applicant re-entry

Current conversions (key events)

EventConv.Type
guilford_college_application2,842Enrollment
guilford_application_button1,889Enrollment
guilford_app_click_to_apply_site1,068Enrollment
click_to_email483Micro / contact
request_information_click481RFI (click, not submit)
application_page_visit183Enrollment
click_to_call98Micro / contact

Plus un-keyed: form_start 15.8k · form_submit 12.7k (generic, not segmented by purpose) · site_search/view_search_results 1.3k.

Gaps the plan closes — read straight off the baseline

No giving trackingDonors are priority-2, but giving has zero conversions. Stand up a cross-domain donation funnel (giving.guilford.edu / GiveCampus).
Vendor / ad-hoc eventsguilford_* ×3 overlap for one action. Standardize (apply_start/apply_submit…) with a reconciliation path so the baseline still ties out.
Generic formsform_start/form_submit fire 15.8k/12.7k but carry no form_type — unusable as funnel steps until segmented.
Off-domain & unjoinedApply portal, catalog, events & giving are unmeasured; portal.mist.com/referral (1,834) is an un-joined hop. The T6 frame is the injection vehicle.
Guilford College · Analytics Measurement PlanConfidential · Internal04 / 13
Ellis Agency.§2 · Goals & KPIs
02

Goals → KPIs, by audience priority

Each KPI ties to a real baseline number and a post-launch target. Targets are framed as relative lift; absolute targets get set at the next pass against a full-year pull (admissions seasonality). "New" KPIs have no baseline — standing them up is the goal.

PRIORITY 1 — PROSPECTIVE STUDENTS & FAMILIES (ENROLLMENT INTENT)
KPIBaseline (90d)Target / threshold
Apply starts≈1,889+1,068 (overlapping legacy)+15% on a clean single apply_start
Apply submits≈2,842 guilford_college_application+10%; start→submit ratio measurable for the first time
RFI submits481 + 478 (click-level only)establish true submit metric; +20% vs click proxy
Visits scheduled0 tracked (/visit traffic exists)stand up; baseline first 90d, then +10%
Program interestnot segmented (finder 6,179 pv)stand up; leading indicator of apply_start
PRIORITY 2 — EXTERNAL CONSTITUENTS (GIVING ENGAGEMENT) · BIGGEST GAP
KPIBaseline (90d)Target / threshold
Give CTR0 trackedstand up; homepage + footer Give-CTR baseline
Donation starts0 trackedstand up cross-domain; baseline first 90d
Donation completes (+ value)0 trackedstand up; report gift count + revenue — headline new capability
Alumni / community engagementpartial (file_download 660)stand up email_signup + event_rsvp
Guilford College · Analytics Measurement PlanConfidential · Internal05 / 13
Ellis Agency.§2 · Goals & KPIs
PRIORITY 3 — CURRENT STUDENTS (WAYFINDING) · KEEP IN VIEW, DON’T NARROW
KPIBaseline (90d)Target / threshold
Wayfinding success/current-students 15,946 pv; high direct/repeathold/grow engagement on these paths; don’t regress
Site-search efficacyview_search_results 1,349 (no zero-result signal)stand up site_search {results_count}; zero-result < 15%
PRIORITY 4 — FACULTY / STAFF (WAYFINDING)
KPIBaseline (90d)Target / threshold
F/S re-entry/faculty-and-staff 8,193 pv; /work-here 2,741; /directory 7,627hold engagement; don’t regress wayfinding
CROSS-CUTTING — REACH, QUALITY & COMPARABILITY (ALL AUDIENCES)
KPIBaseline (90d)Target / threshold
Sessions120,036hold/grow; watch for migration dips (annotate)
Engagement rate45.0%≥ 47% (consistency / IA / mobile)
% Direct + % ReturningDirect 31.2% · Returning 31.2%hold — signals retained loyalty
Mobile engagementMobile 35.3% / Desktop 63.9%grow mobile engagement (mobile-first refresh)
CADENCE & OWNERS

Automated weekly (account lead + Guilford marketing) and daily (internal ops watch) Looker email. Every KPI gets a named owner; any KPI without an owner and a decision attached gets cut.

Guilford College · Analytics Measurement PlanConfidential · Internal06 / 13
Ellis Agency.§3 · Audiences
03

Audiences → GA4 segments / audiences

Built in Ty's locked order (confirmed by Jean), each twice-purpose: a reporting segment now, and a publishable GA4 Audience for future personalization & remarketing. Defined by the baseline behaviors actually observed in the pull.

1

Prospective students & families

DEFINED BY /admissions*, /apply/*, /program-finder, program/major pages, /visit*; or firing apply_start / request_info_submit / program_interest.

USE Sub-audiences (preserve): HS · Transfer · Graduate · Adult · International. Remarketing seed: apply_start-without-submit.
2

External: donors → alumni → community

DEFINED BY Give-CTA clickers, donation_start (cross-domain), giving-landing viewers; alumni/community via news engagement, email_signup, event_rsvp.

USE Baseline anchors: none giving-side (the gap). Remarketing seed: donation_start-without-complete (highest-ROI new audience).
3

Current students

DEFINED BY Direct + returning sessions on /current-students, /account/login, /apply/status; portal.mist.com referrers.

USE The "~60% direct/repeat" cohort — confirmed. Personalization candidate (returning → student quick-links module).
4

Faculty / staff

DEFINED BY Sessions on /faculty-and-staff, /work-here, /directory; internal-tool referrers.

USE Guard wayfinding; lowest emphasis but a real, recurring audience.
CROSS-CUT (OPERATIONAL) Mobile vs Desktop · New vs Returning · Encoura-sourced (display 6,113 + paid social 3,080) for paid-media attribution.
Guilford College · Analytics Measurement PlanConfidential · Internal07 / 13
Ellis Agency.§4 · Conversions
04

Conversions (GA4 key events) + reconciliation

A clean, standardized key-event set replaces the overlapping vendor names. The primary events drive the north-star and the funnels; micro events are leading indicators.

Key eventTriggerKey parametersReconciles / retires
apply_startClick into the application from any Apply CTA / apply portal handoffcta_type · template · audience_priority · apply_levelReplaces guilford_application_button / app_click_to_apply / application_page_visit
apply_submitApplication submitted (apply-portal confirmation, cross-domain)apply_level · source_template · deviceReplaces guilford_college_application
request_info_submitRFI form submit (form_submit {form_type:rfi})program_interest · audience_priorityReplaces request_information_click / request_info
visit_scheduledVisit / tour scheduled (form or scheduler confirmation)visit_type · date · templateNEW — no current equivalent
donation_startGive flow begins on giving.guilford.edu / GiveCampus (cross-domain)gift_designation · source_template · entry_ctaNEW — biggest gap
donation_completeDonation confirmation page (cross-domain)value · currency · gift_type · gift_designationNEW — biggest gap

Micro / supporting events

cta_clickprogram_interestviewbook_viewemail_signupevent_rsvpvideo_play / progresssite_searchfile_downloadclick_to_email / call
RECONCILIATION — SO THE BASELINE TIES FORWARD

Fire both the legacy guilford_* events and the new standardized events in parallel for ~30 days; build a legacy↔new mapping in Looker so the pre-launch baseline stitches cleanly to post-launch. Then retire the legacy events from "key" status (keep collecting briefly for audit). Never delete history — annotate the cutover date.

Guilford College · Analytics Measurement PlanConfidential · Internal08 / 13
Ellis Agency.§5 · Event taxonomy
05

Event taxonomy tied to the templates

Instrument once, globally. One standardized event schema is baked into the shared component kit that powers P0 + T1–T5 and the portable T6 frame. A CTA, card, filter or form emits the same event with the same params no matter which template renders it — the direct synergy with the page-type inventory.

Shared componentEventNote
CTA system / bandscta_click (+ apply_start / donation_start when it’s the funnel entry)Give CTR + all CTA reporting
Card gridscard_clickcard_type · label · position
Filter / sort + Load Morefilter_use · load_morefilter_type · value · results_count
Formsform_start · form_submit {form_type}fixes the generic 15.8k/12.7k problem
Article body / mediavideo_play · video_progress · file_download · scrollengagement on the highest-volume type
Stat / promo bannercta_click · viewbook_viewgiving + viewbook promos
Global searchsite_searchsearch_term · results_count

Canonical parameter dictionary

ParameterValues / formatPurpose
templateP0, T1–T6which template fired it
page_typehomepage · hub · article · listing · profile · program · event · interior · landing · externalmaps to the 10 page types
audience_priority1–4powers audience-priority dashboards
cta_typeapply · give · visit · rfi · login · learnCTA classification (Give CTR)
positionhero · nav · body · sidebar · footer · bannerdid the prominent Give placement work?
devicemobile · desktop · tabletmobile-first validation
form_typeapply · rfi · visit · contact · newsletter · event_rsvpthe fix for generic forms
value · gift_type · gift_designationamount · one_time/recurring · fundgiving revenue reporting
DATALAYER CONTRACT The component kit pushes a uniform dataLayer object ({event, …params}); GTM maps it to GA4. The T6 frame ships the same push + GTM snippet, so non-Drupal systems emit identical events with zero per-system work.
Guilford College · Analytics Measurement PlanConfidential · Internal09 / 13
Ellis Agency.§6 · Funnels
06

Conversion funnels

Five funnels expressed as the standardized events. Most ratios are first-measurable post-launch — that's the point. Highlighted steps are key events.

A

Enrollment (primary)

session_startcta_click {apply}Admissions hubapply_startapply_submit

BASELINE /admissions/apply 7,378 sessions; legacy app events ≈2,842 conv. True start→submit ratio not currently computable — standardization makes it computable.

B

Giving (cross-domain)

entry / herocta_click {give}giving landingdonation_startdonation_complete

NET-NEW Zero giving tracking today. Stand up the full cross-domain funnel; first 90d post-launch is the giving baseline. Prominent homepage Give CTA (P0) is the top-of-funnel lever.

C

Visit

entry/visit · campus-mapcta_click {visit}schedulervisit_scheduled

NET-NEW /visit + /visit/campus-map rank in the top ~15 pages but no visit_scheduled event exists — net-new conversion.

D

Request info (RFI)

entrycta_click {rfi}form_start {rfi}request_info_submit

BASELINE Click proxies 481 + 478; generic form_start 15,767 / form_submit 12,708. Segmenting forms gives the first true RFI start→submit rate.

E

Current-student wayfinding (success = fast exit)

entry (direct/return)/current-students · logincta_click {p3} / site_searchtask complete

BASELINE /current-students 12,506 sessions; login 2,700; Direct 31.2% + Returning 31.2%. Metric is engagement + low search-refinement; guardrail is no regression post-refresh.

Guilford College · Analytics Measurement PlanConfidential · Internal10 / 13
Ellis Agency.§7 · Cross-domain & technical
07

Cross-domain & technical plan

What makes off-domain conversions (giving, apply_submit) measurable and attributable — and how the refresh's component kit + T6 frame carry consistent tracking everywhere.

01Audit first, enhance — don’t rebuild. Inventory the current GTM container / gtag, existing key events & streams before changing anything; map legacy → new so no history is lost.
02Standardize via GTM. Move event firing into one GTM container fed by the component-kit dataLayer — one container, one mapping, all templates.
03Cross-domain measurement across guilford.edu, the apply portal (/apply/*, portal.mist.com), giving.guilford.edu / GiveCampus, catalog.guilford.edu, and the events platform.
04Referral exclusions for those hosts so cross-domain hops don’t fragment sessions or self-attribute (today portal.mist.com/referral = 1,834 is an un-joined hop).
05T6 is the injection vehicle. The portable frame ships the GTM snippet + the same dataLayer push, so non-Drupal systems emit identical events with no per-system engineering.
06Consent Mode v2 / ADA-privacy. Implement analytics_storage / ad_storage consent, keyboard-accessible banner (WCAG AA), and flag consent-driven gaps loudly so trends don’t lie.
HARD DEPENDENCY — GATES OFF-DOMAIN MEASUREMENT

Deploying one GTM container + the T6 dataLayer/GTM snippet on the apply portal, giving platform, catalog & events requires IT sign-off (Gloria Thornton) plus GiveCampus tag access. Without it, the giving funnel and a true apply_submit can't be measured.

Guilford College · Analytics Measurement PlanConfidential · Internal11 / 13
Ellis Agency.§8 · Dashboards & reporting
08

Dashboards + automated reports

A Looker Studio dashboard organized by audience priority, plus automated weekly + daily email — satisfying the kickoff's "automated reporting" ask.

Enrollment
apply_start/submit, RFI submit, visit_scheduled, start→submit funnel, by template & device
Giving (net-new)
Give CTR, donation_start/complete, revenue (value), funnel
Audience reach
sessions, new vs returning, % Direct + % Returning, channel mix (Encoura called out)
Content / top pages
top pages & sections (news volume), engagement rate, duration
Wayfinding / search
current-student + faculty/staff path engagement, site_search + zero-result rate
Mobile vs desktop
split + engagement — validating the mobile-first refresh
Distribution. Dashboard URL goes to the project link library (kind analytics) and the deliverable embedUrl as a live embed.
Automated email. Looker scheduled delivery — weekly (account lead + Guilford marketing), daily (internal ops watch).
Portal metrics auto-pull. The GA4 seam is live (gaPropertyId 359308554); manual metrics stay the source of truth until adopted.
Clean before/after. Annotate launch + tracking-cutover dates in GA4 so comparisons are honest.
SUCCESS THRESHOLDS (POST-LAUNCH, VS RECONCILED BASELINE)

Enrollment: apply_start +15%, apply_submit +10%, RFI submit +20%. Giving: Give CTR + donation funnel established from zero. Quality: engagement ≥ 47%, mobile up, no wayfinding regression. Comparability: standardized events live, legacy reconciled, forms segmented.

Guilford College · Analytics Measurement PlanConfidential · Internal12 / 13
Ellis Agency.Summary & next steps

Where this leaves us

OPEN DEPENDENCIES

GiveCampus tag access & confirmation-page reachability · legacy guilford_* event ownership (Encoura) for retirement · GTM container ownership · apply-portal submit signal · visit-scheduler platform · full-year pull for seasonality-anchored absolute targets · Guilford consent/privacy posture.

Ellis Agency. Guilford College · Website & Analytics Refresh · EA-2026-0002 · Internal
Guilford College · Analytics Measurement PlanConfidential · Internal13 / 13