Alex Harris and Optimizely

I didn’t start with Optimizely. I started with the problem it solves.

I started running A/B tests in 2001, before Optimizely existed, building the variants, tracking and reporting myself. More than a decade later, I help organizations adopt the platform, connect it to the rest of their stack, build programs around it, and now design agent workflows with Opal. The full story is on the next page.

Experimentation and personalization leader · Optimizely MVP since 2022

What the job actually looks like

Experience
- CMS
- Commerce
- Personalization
Learn
- Web Experimentation
- Feature Experimentation / Full Stack
- Analytics and behavioral evidence
Know the customer
- ODP
- Profiles
- Audiences
- Behavioral and customer data
Operate content
- CMP
- Planning and workflows
- Approvals
- Reusable content
Automate the work
- Opal / Agent Platform
- Specialized agents
- Virtual Teammates
- MCP and integrations

Business decisionWhat should change, for whom, why, and what should happen next?

Five connected systems, one decision at the center. Not a stack of separate SKUs.

10+Years hands-on with Optimizely
20+Onboarding engagementsDocumented, as a named resource in the co-delivery program
2019Co-delivery launch teamPartner onboarding for Optimizely’s new customers
2022Optimizely MVP since
+204.5%Partner clicks in a program I ledTeam result, first quarter, >99% confidence

Decision rules adopted
A large university’s enrollment leadership adopted the experiment decision rules I rewrote. The governance story
A campaign redirected
A customer-data check showed a growth list held no current members, so the campaign became win-back. The data story
AI cost brought down
Repeatable analysis moved off Opal after one day used 500 to 600 credits. The Opal story
Programs onboarded
Named on 20+ co-delivery onboardings and owner of the packaged Web onboarding offer. How it worked

The job

The job I usually end up doing isn’t a product name.

Customers rarely have an Optimizely-only problem. They have a business problem, an Optimizely subscription, and a stack of other tools that don’t always agree about who the customer is.

Customer outcomeThe decision the business needs to make better

Optimizely

Experimentation
Personalization
ODP
CMS
CMP
Opal

Evidence

Web analytics
Behavior analytics
Research

The testing tool almost always had an analytics tool next to it.

Customer data

CRM
Customer data platform
Warehouse
Identity

Activation

Email and lifecycle
Commerce
Sales channels

Operations

Content workflow
Ticketing and project tools
Governance

These showed up as programs matured, not on day one.

Where I sit: in the middle, connecting the pieces

UnderstandWhat is the business trying to change?
ConnectWhich signals and systems actually matter?
DecideWhat Optimizely should own, and what it shouldn’t
EnableGet the team running it on its own
MeasureProve it changed the decision, then pick the next one

Categories, not a checklist. No customer had every one of these, and the mix changed as programs grew.

The platform is only valuable when the customer knows what decision it is supposed to help them make.

That’s the thread through everything in this section. I’ve worked across the whole chain: the customer problem, the experience, experiments and personalization, the data underneath, the integration, the rollout, the training, the measurement, and now the agents.

That breadth is why I end up translating between marketing, product, engineering, data, sales and customer success. The pages that follow show what that looks like in practice.

Where to start

Pick the situation closest to yours.

The same platform experience answers different questions depending on where a team stands today. Open the one that matches.

I’m evaluating Optimizely

Start with the business problem, not the product demo. See how I scope what Optimizely should own before a line of config exists.

Problem-first scoping
Four real solution designs
Where Optimizely stops

See the approach

I have it, and need my team running it

The onboarding system that gets a team from a license to a program they can operate themselves, plus the artifacts they keep.

Seven-stage onboarding system
Templates left behind
Two programs at scale

See the playbook

I already run programs and want more value

What compounding value looks like beyond a single lift: trusted measurement, a real result, and the four kinds of scale.

A quarter of real results
A decision not to personalize
Value maturity model

See the outcomes

I want the AI and agent direction

How agent chains, human approval gates, MCP and Virtual Teammates turn AI capability into something governed and measurable.

Give the agent a job
Human authority
Virtual Teammates

See the systems

Every path leads back here through the credentials and platform depth below.

Inside

Five chapters, one question each.

Platform depth

Where I’ve gone deep, and where I haven’t.

Open any capability to see what I actually did with it. The labels are deliberately conservative: a demo counts as exposure, and one configured teammate counts as a prototype.

BUILT
ONBOARDED
SOLUTION DESIGNED
INTEGRATED
PROTOTYPED
ADVISED
EXPOSURE

Web ExperimentationONBOARDEDSOLUTION DESIGNEDINTEGRATED
Where most of my Optimizely years went: onboarding, program design, governance and delivery.
Experimentation Certified Strategist 2025
Feature Experimentation / Full StackONBOARDEDSOLUTION DESIGNED
Server-side and feature-flag programs across product teams, and the call on when to use it instead of Web.
PersonalizationONBOARDEDSOLUTION DESIGNED
Campaigns inside onboarding programs, and the discipline of deciding when a segment has earned its own experience.
Optimizely Data Platform (ODP)SOLUTION DESIGNEDINTEGRATED
Identity between a public site and a member portal, audience checks, consent and reachability.
CMSSOLUTION DESIGNEDADVISED
Client work alongside developers: QA, templates, forms, audience planning.
CMS (SaaS)BUILTPROTOTYPEDINTEGRATED
Built pages through the content API, verified through Graph, and an MCP integration over the REST API.
CMPPROTOTYPEDINTEGRATED
The human approval step inside agent workflows.
CMP Certified Administrator 2025
Content RecommendationsADVISED
Part of connected-platform designs; not a deep hands-on area.
Product RecommendationsEXPOSURE
Enablement exposure only.
CommerceADVISED
B2B commerce as part of a manufacturer’s data and account picture.
Opal / Agent PlatformBUILTPROTOTYPEDSOLUTION DESIGNEDINTEGRATED
Agents, tools, skills, workflows and approval gates, plus advice on where not to use it.
Agent Platform Certified Administrator, Tools Certified, Development Accredited (2025)
Virtual TeammatesPROTOTYPED
Configured one, connected to a read-only system.
1:1 / autonomous experiencesPROTOTYPEDEXPOSURE
Built rule-driven personal pages; learned Optimizely’s autonomous method as a partner. That product is theirs.

Credentials

The paper trail.

Credentials support the story. They aren’t the story, so here they are in one place.

Optimizely MVP since 2022Recognition, Strategy
Optimizely Experimentation Certified Strategist 2025Certification
Optimizely Agent Platform (Opal) Certified Administrator 2025Certification
Optimizely Agent Platform (Opal) Tools Certified 2025Certification
Optimizely Agent Platform (Opal) Development AccreditedAccreditation
Optimizely CMP Certified Administrator 2025Certification

I didn’t arrive at this from a certification. I arrived at it by solving the problem by hand, years before the tools did it for me.

Next chapterMy Optimizely Story

View my resume LinkedIn

Optimizely and related product names are trademarks of their owner. This section reflects my own professional experience with the platform. It is not affiliated with or endorsed by Optimizely, and client examples are described by industry, not by name.