Hoi Polloi Tech — MarTech, Digital Analytics, Personalisation & Cookie Governance

Measurement, personalisation and consent work delivered for and alongside

Citigroup

PwC

Marsh McLennan

Marsh

Horizon BCBSNJ

The Hartford

Capital One

The Vanguard Group

Microsoft

Intel

Lenovo

BMW

Hyundai

Rolls-Royce

Tetra Pak

MARS

PepsiCo

Subway

Sazerac

W. R. Grace

Roche

Wiley

Cox Communications

BYU Broadcasting

TCS

Infosys

LTIMindtree

C5i

Engagements delivered in principal and lead architect roles — for end clients directly, and alongside global delivery partners.
All names and marks are the property of their respective owners and are shown to describe work delivered — they imply no endorsement, partnership or current relationship.

The 2026 shift

Three things changed at once. Most stacks changed for none of them.

Agentic AI arrived in the marketing stack, third-party identity finished collapsing, and consent stopped being a legal checkbox and started being a measurement problem. Each of them punishes a weak data foundation — and they arrived together.

Agentic

At Summit 2026 Adobe introduced CX Enterprise Coworker, an agent layer that plans and executes work across Experience Platform, Real-Time CDP, Customer Journey Analytics and Journey Optimizer. Journey Agent and Data Insight Agent go further still. All of them reason over your XDM schemas and identity graph. Where that foundation is wrong, agents don't fail loudly — they scale the error.

Source: Adobe, April 2026

15–30%

The share of conversion signal organisations are recovering by moving to server-side collection — events routed through infrastructure you control before distribution, rather than fired from the browser into a dozen vendors. It is now the default architecture for resilient measurement, not an optimisation.

Source: Digital Applied, 2026

67%

The proportion of Consent Mode v2 implementations reported to fall short of compliance standards. Consent is no longer a banner your legal team signs off — it is a signal that has to propagate correctly through every tag, every server call and every downstream model. Most implementations quietly don't.

Source: Tatvic, GTM 2026 guide

What we do

One discipline, four layers.

We are not a full-service transformation firm and we don't pretend to be. We do one thing — the customer data and measurement layer — at a depth that generalists cannot reach.

Digital Analytics & Measurement Architecture

Measurement designed around the decisions the business actually makes, not the events that happened to be easy to capture. Thirteen years on Adobe Analytics and seven on Customer Journey Analytics — implementation and migration, Web SDK and Edge Network, server-side collection, data layer design, and the unglamorous discipline of making numbers reconcile across Adobe, Google and the back end.

Adobe AnalyticsCustomer Journey AnalyticsWeb SDK / EdgeAdobe LaunchTealium iQServer-side collectionGA4Data layer design

Personalisation & Journey Orchestration

Segments that resolve to real people, journeys that respond in the moment, and experimentation with enough statistical honesty to be worth acting on. Adobe Target across A/B, MVT, experience targeting, Auto-Target and Recommendations; Journey Optimizer for orchestration; Real-Time CDP for the identity and activation layer that makes any of it possible.

Adobe TargetJourney OptimizerReal-Time CDPIdentity graphsXDM schemasAudience activationExperimentation design

Cookie Governance & Consent-Safe Data

The layer most teams bolt on last and regret first. Consent platform configuration, tag-to-consent mapping, cookie inventories that match the policy you actually published, and regional variation across GDPR, ePrivacy, CPRA and TCF. Then the harder half: designing measurement that stays accurate as consent rates fall, rather than simply going dark.

OneTrustTrustArcOsanoConsent Mode v2Cookieless identityGDPR / ePrivacyCPRAData usage labels

AI & Agent Readiness

The newest layer, and the one with the widest gap between the demo and the deployment. Adobe's agents reason over XDM schemas, identity stitching and data usage labels — so schema hygiene, governance labelling and consent enforcement stop being back-office concerns and become the determinant of whether an agent can be given authority at all. We assess readiness honestly, including when the answer is that the foundation has to come first.

AEP AI AssistantJourney AgentData Insight AgentSchema & governance readinessData usage labelsPrivacy-preserving modelling

Live problems

What organisations are calling us about right now.

"We have to move to Customer Journey Analytics and nobody here has done one."

The largest live programme in the Adobe base, and we have taken sites through it end to end. The work is rarely the migration itself — it is the identity strategy, the data-quality assumptions inherited from an implementation nobody has audited in years, and giving stakeholders a straight answer about which historical reporting genuinely comes with them and which does not.

"We bought AEP eighteen months ago and we still aren't getting value."

Extremely common, and rarely a licensing problem. It is usually schema design that fought the use cases, an identity graph built on identifiers the site was never actually collecting, or activation paths that were never finished. An architecture review finds which — and sometimes concludes the platform was the wrong answer for the use case.

"Our consent rate dropped and we can't tell what's real any more."

Falling consent doesn't just shrink the sample — it biases it, in ways that quietly break attribution, audience sizing and every model trained downstream. The fix is part remediation, part re-architecture: consent-safe collection, server-side routing, and honest modelling of what can and cannot be recovered.

"We want to turn on the AI agents. Are we ready?"

Agentic features are the fastest-moving part of the Adobe stack, and the readiness question is genuinely technical: are the schemas coherent, is identity resolving, are data usage labels applied, will consent be enforced when an agent activates an audience? We answer it as an assessment rather than a sales conversation.

"We're a partner and we're short of senior Adobe hands on a delivery."

We work white-label behind agencies and Adobe Solution Partners on CJA, AEP and consent workstreams. You keep the client relationship and the commercial; we bring certified senior delivery capacity, in your time zone, without the hiring cycle.

The stack

Platforms we work in every day.

Depth in the Adobe Experience Cloud, with genuine working knowledge of the alternatives — which is what makes a vendor-neutral recommendation worth anything.

Adobe

Tealium

Treasure Data

Google

Piwik PRO

OneTrust

TrustArc

Osano

ObservePoint

Medallia

Demandbase

Conductor

Meta Pixel

LinkedIn Insight

Google Floodlight

Power BI

Tableau

Databricks

Microsoft Azure

Dynatrace

Dynamic Signal

Product and company names and logos are trademarks of their respective owners, shown here to describe the platforms we work in. Hoi Polloi Tech is an independent practice and holds no resale relationship with any vendor.

Adobe Experience Cloud

Adobe Analytics

Customer Journey Analytics

Experience Platform

Real-Time CDP

Journey Optimizer

Adobe Target

Data Collection / Launch

Web SDK & Edge Network

Audience Manager

Experience Manager

CDP, Tagging & Activation

Tealium iQ

Tealium AudienceStream

Treasure Data CDP

Google Tag Manager

Server-side GTM

Google Analytics 4

Piwik PRO

Meta Pixel

LinkedIn Insight Tag

Google Floodlight

ObservePoint

Consent & Privacy

OneTrust

TrustArc

Osano

Consent Mode v2

IAB TCF

Data usage labels

Experience, Voice & Demand

Medallia

Demandbase

Conductor Searchlight

Dynamic Signal

Adobe Target

Experimentation design

Reporting & Engineering

Power BI

Tableau

Databricks

Dynatrace

Azure

SQL

JavaScript

Product names are the trademarks of their respective owners and are used here to describe the platforms we work in. Hoi Polloi Tech holds no resale relationship with any vendor.

Why us

The depth of a specialist. The candour of an independent.

Our name says it — hoi polloi, the many. Enterprise-grade measurement architecture has been priced and packaged for the few. We think that is a distribution problem, not a difficulty problem.

Senior practitioners do the work. The person in the pitch is the person in the build. No pyramid, no handover to a team you haven't met.
Vendor-neutral, and it costs us money. We hold no resale relationship with any platform, so we will tell you when Adobe is the wrong answer — which is not something a partner earning margin on the licence can say.
Certified to Master level. Adobe Certified Master — Adobe Analytics Architect, current. Independently verifiable, which is more than can be said for most agency credentials.
Fixed scope, fixed price, fixed date. Assessments are productised so you know exactly what you are buying before you commit. No open-ended discovery.
Knowledge transfer is the deliverable. If your team cannot run it after we leave, we have sold you a dependency rather than a capability.
Full UK and EU working-day coverage, with overnight turnaround for US teams.

Your AI agents are only as good as the data underneath them.