Marketing Cloud Personalization

Marketing Cloud Personalization Needs a Foundation Before It Needs Configuring

Marketing Cloud Personalization, formerly Interaction Studio, is real-time interaction management: recognizing a visitor, deciding what to show them as the page renders, and learning from what they do next. Twopir Consulting implements it starting from identity resolution and the catalog — because without those two, the recommendations are generic and the affinity models have nothing to model.

Discuss Personalization Marketing Cloud Consulting
  • 12+ Years of CRM and marketing platform delivery
  • 500+ Clients across six markets
  • 40+ Certified Salesforce, HubSpot and AI specialists
Personalization Stack
THE STACK, BOTTOM-UP Identity resolution Anonymous today, known tomorrow Layer 01 Catalog What can be recommended, and why Layer 02 Behavioral capture Views, dwell, search, cart, app events Layer 03 Affinity and Einstein models What this visitor tends toward Layer 04 Experiences What actually changes on the page Layer 05 TWOPIR PERSONALIZATION PRACTICE Identity First Before any experience is built A Real Catalog Structured, current, and complete Measured Experiences Against a control, not against last month 2πr WHAT CHANGES Recommendations grounded in this visitor Rather than a bestseller list with a name on it One experience across web, app and journey The site and the email stop contradicting each other
Prerequisites

Four Things That Have to Be True Before Personalization Is Worth Buying

Marketing Cloud Personalization is a capable product and it is regularly bought before the organization can use it. The symptom is familiar: the license is live, the recommendations are running, and they recommend roughly what a bestseller list would.

That is not a product failure. It is what happens when a real-time decisioning engine is given no reliable way to tell visitors apart and no structured description of what it is allowed to recommend.

Both are fixable, and both are cheaper to fix before the license than after it.

You can tell visitors apart, and recognize them again

Personalization works on anonymous visitors and gets substantially better when it can tie a session to a known customer — and better again when it can join today's anonymous session to last month's known one. That is identity resolution, and it is an architecture problem rather than a setting. Where identity should be resolved across web, app, commerce and CRM together, the answer is usually Data 360 rather than anything inside the personalization tool.

Your catalog is structured, current and complete

Recommendations are only as good as the description of what can be recommended. If the catalog lacks categories, attributes, availability or relationships between items, the engine has nothing to reason with and falls back to popularity — which is what produces generic output. Catalog work is unglamorous, it is usually owned by a team outside marketing, and it is the single highest-leverage input to the whole system.

Behavioral capture actually covers the journey

The models learn from what visitors do, which means the sitemap configuration — what is tracked, on which pages, with what taxonomy — determines what can ever be learned. Partial instrumentation produces partial models, and the gap is invisible in the interface: the engine reports confidently on the subset it can see.

Someone owns the experiences after launch

Personalization is not a project that completes. Experiences need reviewing, models need watching, and the catalog changes underneath both. Without a named owner the campaigns from launch keep running against a catalog that has moved on, which degrades quietly rather than failing visibly.

Use Cases

Four Things Personalization Is Genuinely Good At

Described in terms of what the visitor experiences rather than what the product does. Each is worth having; none is worth having before the foundation above is in place.

Use Case 01

Recognizing a returning visitor

The most valuable and least discussed capability. Someone who browsed three times last week and has now returned should not be treated as a first-time visitor, and the site should not re-explain what they already looked at.

  • Anonymous session continuity across visits
  • Recognition at sign-in, joining the anonymous history to the known profile
  • Recently viewed and abandoned context carried forward
  • Suppressing acquisition messaging for existing customers
  • A consistent identity between the website and the journey
Use Case 02

Recommendations that reflect this visitor

Einstein recipes combine behavioral signals with catalog structure to decide what to surface. The quality of the output is almost entirely a function of the inputs — which is why the catalog work matters more than the recipe tuning.

  • Einstein recipes configured against real business rules
  • Catalog modelling: attributes, categories and relationships
  • Business constraints — margin, stock, seasonality, exclusions
  • Recommendations in email as well as on the page
  • Measured against a control, so their value is known rather than assumed
Use Case 03

Real-time content decisions on the page

What a visitor sees changes as the page renders, based on who they are and what they have done. Powerful, and the place where poorly governed personalization becomes visible to customers as inconsistency.

  • Banner, module and message personalization by segment or affinity
  • Next-best-action surfaced in-session rather than in tomorrow's email
  • Experience rules that degrade sensibly for unknown visitors
  • Governance: what may be personalized and what must stay fixed
  • Performance budget, so personalization does not cost page speed
Use Case 04

Feeding behavior into journeys and CRM

Personalization sees what happens between emails. That behavior is frequently more current than anything in the CRM, and routing it into journey decisions is where the two products compound rather than merely coexist.

  • Behavioral signals as journey entry and decision criteria
  • Affinity and segment membership available to email personalization
  • Web behavior surfaced to sales and service through CRM
  • Consistent treatment across channels rather than three opinions
  • Data sharing between Personalization and Salesforce CRM configured deliberately
Disambiguation

Four Different Things Are Called Personalization

This is the source of most confused conversations about the product. A team says it wants personalization, everyone agrees, and it emerges later that three people meant three different capabilities with different prerequisites and costs.

Personalization capabilities compared
What people meanWhat it actually isNeedsWhere it lives
"Use their first name"Merge fields and personalization strings at send time.A populated attribute and a fallback for when it is missing.Native in Email Studio and Content Builder. No additional license.
"Different content for different segments"Dynamic content blocks, chosen by rule at send time.A reliable attribute to segment on.Native in Content Builder. Still no additional license.
"React to what they did"Journey decisions based on behavior and engagement.The behavior reaching the platform as an event or an attribute.Journey Builder, fed by the integration layer.
"Change the website as they browse"Real-time interaction management: decisions made as the page renders, models learning continuously.Identity resolution, a structured catalog, and behavioral capture across the journey.Marketing Cloud Personalization. This is the product on this page, and the only one of the four that needs its own license and its own foundation.

Capabilities verified against Salesforce documentation Last reviewed 23 September 2026

12+
Years delivering CRM and marketing platform transformation
500+
Clients served across global markets
40+
Certified specialists across Salesforce, HubSpot & AI
98%
Client retention rate
Failure Modes

Four Reasons a Personalization License Sits Underused

This product has a higher rate of under-realized investment than anything else in the Marketing Cloud family, and the reasons are consistent enough to list.

None of them are about the product. All four are about what was true in the organization when it was bought.

Bought before identity was solved — The engine cannot reliably tell one visitor from another, or connect today's anonymous session to a known customer. It does the best it can with sessions, which is a fraction of what it could do with people.

A catalog with no structure — Items exist but lack categories, attributes and relationships. The models have nothing to reason with, so recommendations converge on popularity — and everyone concludes the AI is disappointing rather than underfed.

Experiences launched and never revisited — Built during implementation, still running two years later against a catalog and an audience that have both moved. Nothing fails; the value simply erodes, and because nothing fails nobody looks.

Never measured against a control — Personalization is switched on everywhere at once, so there is no unpersonalized comparison. When someone eventually asks what it contributed, the honest answer is that nobody set it up to be answerable.

Before You Buy

Five Questions to Answer Before the License

If the answer to more than two of these is uncertain, the honest recommendation is to fix the foundation first. We would rather say that before a license is signed than explain it afterwards.

Can you reliably recognize a returning anonymous visitor, and connect that session to a known customer when they sign in?
Does your product or content catalog carry the categories, attributes and relationships a recommendation engine would need?
Is behavioral capture instrumented across the journey you care about, or only on the pages that were easy?
Who will own the experiences after launch, and how often will they review them?
How will you know whether personalization worked — is there a control group, or only a before-and-after?
Related

Data Cloud

Where identity resolution belongs when it has to span web, app, commerce and CRM rather than one channel.

Journey Builder

Turning behavioral signals into journey entry and decision criteria.

Email Marketing

The other three things people mean by personalization, none of which need this license.

Marketing Cloud Next & Agentforce

Salesforce Personalization is also available as a Marketing Cloud Next add-on.

Marketing Cloud Consulting

The practice page: architecture, edition and add-on selection.

Marketing Cloud Case Study

A worked example of the architecture-first approach on a real client engagement.

Common Questions

Questions About Personalization

Yes. Marketing Cloud Personalization is the current name for the product previously called Interaction Studio, and before that Evergage. The capability is real-time interaction management: recognizing a visitor, deciding what to show them as the page renders, and learning from what they do.

The older names still carry search volume and still appear in internal documentation, which is worth knowing when you are reading anything written more than a couple of years ago.

No, and this is the most common and most expensive confusion in this area. Merge fields, dynamic content blocks and rule-based variation are all native to Content Builder and need no additional license.

Marketing Cloud Personalization is for real-time decisions made as a web page or app screen renders, with models learning continuously from behavior. If the requirement is "different content for different segments in an email", you already have that.

Structure, currency and completeness. Items need categories, meaningful attributes, availability, and relationships between them — because that is what the engine reasons over when deciding what to surface.

Catalog quality is the dominant input to recommendation quality, ahead of recipe tuning by a wide margin. It is also usually owned by a team outside marketing, which is why it needs to be raised early rather than discovered during configuration.

They solve adjacent problems and are frequently confused. Data 360, formerly Data Cloud, resolves identity and unifies profiles across sources for the whole business. Personalization decides what an individual sees, in real time, on a digital property.

Where identity has to be resolved across web, app, commerce and CRM together, that belongs in Data 360, and Personalization consumes the result. Trying to solve enterprise identity inside the personalization tool is a common and expensive detour.

Yes, and it is mostly a governance question rather than a technical one. The rule we work to is that a customer should be able to explain why they saw what they saw — recommendations based on what they browsed are expected; content implying you know something they did not tell you is not.

That means deciding up front what may be personalized and what stays fixed, and honouring consent and preference signals in the experience layer as well as in messaging. It is worth writing down before the first experience is built.

With a control group, established before personalization is switched on. That is the one decision that cannot be made retrospectively, and skipping it is why so many teams cannot answer this question about a product they have owned for two years.

Holding back a proportion of traffic feels like leaving value on the table. It is what turns "engagement is up" into a number you can defend in a renewal conversation.

Next Step

Check the Foundation Before the License

Tell us how you currently recognize a returning visitor and what state your catalog is in. Those two answers usually determine whether Personalization is the right next investment or the one after that.

Discuss Personalization Marketing Cloud Consulting

Serving: US | Canada | UK | UAE | Australia | New Zealand

Relatable? We should definitely talk.

What we'll cover:

From CRM and integrations to custom apps and complex system architecture — we help you scale without chaos.
  • Identify revenue leaks across CRM, integrations, and GTM
  • Design and optimize complex systems (CRM to Custom Apps)
  • Apply practical AI to improve operations and pipeline conversion
  • Eliminate silos and build a unified revenue system
  • Assess GTM performance and key bottlenecks
  • Align teams with clear processes and ownership
  • Define a scalable RevOps model
  • Improve forecasting and reporting
  • Review your HubSpot/Salesforce setup for scale