Personalization: recommendations & path forward

What we can do, the right tool for each, and where to start

We pressure-tested all 13 personalization ideas with engineering. The headline: every one is achievable — the question was never if, but which tool and how much lift. ContextHub is one option, but for most of these the right tool is Adobe Target (which the team already knows well), with a few belonging to the Magento commerce engine or already built into OnDemand. Below, the ideas are grouped fastest-first so we can decide where to start.

✓

Proven: Target can read ContextHub

The key technical question — can Adobe Target read visit data stored in ContextHub across visits — is answered. Trone and Jacqueline built a working proof-of-concept for #1 New vs. returning messaging. This de-risks the whole history-based family (#1, #6, #2): the hardest architectural question is solved, and what remains is building the individual experiences on top of a validated pipeline.

4
Quick wins
3
Doable — one confirmation
2
Needs discovery
4
Better handled elsewhere

13 ideas total · nothing is off the table — it's a matter of sequencing and dev investment

Tools:
Target front-end, framework-agnostic
ContextHub AEM, first-touch / components
Magento commerce engine
Built-in already on OD
Lift: SM MD L–XL
Quick wins — low lift, no major dependencies. Strong candidates to start with.
IdeaToolLiftWhat it takes
06Category interest carry-through
Track the test categories someone browses and prioritize them on return — e.g. reorder the top row of tests by user type.
Target SM–MD
Target swaps the top row of tests based on the type of user. Simple if we can use the existing uservisited object.
Depends on OD confirming the uservisited object
03Browser / device targeting
Mobile tap-to-call vs. desktop "find a location."
ContextHubor responsive code SM
We can sense browser/device through ContextHub, or do it just as easily with responsive design. Lowest-hanging fruit.
04Time-of-day messaging
"Same-day results" only during lab hours; weekend variants.
ContextHub SM
Easy — ContextHub reads the customer's browser time. Only open question is whether the placement is a component we can personalize.
07Seasonal / campaign banners
Flu season, back-to-school, New Year wellness pushes.
ContextHub SM
Easy ContextHub use case — needs no additional data or integrations.
Doable — needs one confirmation — achievable in Target; each hinges on confirming we can access one data source.
IdeaToolLiftWhat it takes
01New vs. returning messaging
Welcome-back cues, "pick up where you left off."
Target MD–L
Target stores visit history and can build an audience off it directly. Proof-of-concept is working — the remaining lift is building out the welcome-back experiences and QA.
✓ POC proven — Target reads ContextHub
02Page-hit urgency nudges
After 3+ visits to a PDP without buying, show a "still deciding?" prompt + social proof.
Target+ Magento remarketing? MD–L
A Target audience, since ContextHub only sees first-touch (Target handles everything after). Worth also checking what remarketing the Magento engine can do.
08Referral-source messaging
Tailor hero/CTA by UTM — paid vs. organic vs. email.
ContextHub SM–MD
Build custom ContextHub segments for paid vs. organic — straightforward once we confirm we have access to the UTM data and can leverage it.
Confirm UTM data access
Needs discovery — possible, but real questions to resolve with OD before committing.
IdeaToolLiftWhat it takes
05Geo-based PSC finder
Auto-surface the nearest Patient Service Center on PDP / checkout.
ContextHub+ Target + custom L–XL
ContextHub can do geolocation, but surfacing it at checkout raises open questions: does a geo lookup exist, is it an exposed endpoint, can we process customer location data, and is checkout a component we can touch (likely React, not a component).
10Logged-in user personalization
Greet by name, show past orders, reorder reminders.
Custom + OD L–XL
Needs real discovery with OD on what personal data is allowed. Today even the logged-in check carries no PII — it's just a randomized value that signals "logged in." HIPAA is the gating factor.
Better handled elsewhere — not a ContextHub job; routes to Magento, an existing OD feature, or a custom build.
IdeaToolLiftWhere it goes
09Cart abandonment recovery
Return-visit "complete your order" banner.
Magento —
Not a ContextHub candidate, but there are real possibilities within the Magento commerce engine.
11Recently viewed tests
Persistent "recently viewed" module.
Built-in —
OnDemand already has a widget for this — no new build needed.
12Dynamic bundling / cross-sell
"Frequently ordered together" on PDP and cart.
Magento —
Depends on purchase data collected — commerce-engine territory, not ContextHub.
13Health profile quiz → recs
Intake quiz stores answers, personalizes the grid.
Custom build —
Not a ContextHub candidate — would be a custom build.
Why Target leads here

Most of OnDemand — especially the cart and checkout — is a React app, not editable AEM components. ContextHub can only personalize content AEM can see and is locked one-to-one to the template (it can't even reorder elements). Target rewrites the front end regardless of framework, which is why it's the better fit for the bulk of this list. ContextHub still earns its place on the simpler component-level swaps and first-visit signals.

What we need from OD to unlock more

Most of the list above hinges on three things we can only confirm with the OnDemand team. These are the decisions that turn "possible" into "scheduled" — and where dev investment, if any, would go.

✓

The uservisited object — PROVEN

OnDemand already tracks page visits in browser storage, and we've now confirmed Target can read it across visits (working POC on #1). This was the single biggest lever — and it's solved. Remaining work is building each experience on the validated pipeline.

→ Unlocked: #1 New vs. returning, #6 Category interest, most of #2
2

Magento commerce capabilities

Cart, recently-viewed, bundling, and remarketing likely live in the Magento engine, not ContextHub or Target. We need a short discovery on what it can do out of the box and what's turned on.

→ Unlocks #9 Cart recovery, #2 remarketing, #12 Bundling
3

Data access & privacy rules

What customer data can we use, given HIPAA? Specifically: is UTM data available, does a geo-lookup endpoint exist and can we process location, and what personal data is permitted for logged-in personalization?

→ Unlocks #8 UTM, #5 Geo PSC, #10 Logged-in

See it live

Working prototypes built from OnDemand's real styles, so we can see the quick wins in context rather than imagine them. (Visual mockups — not yet production code.)