Direct answer
What does it mean to treat a website as a product?
A website operates as a product when it helps a defined user complete a useful job, connects to business or product systems, produces evidence about behavior and outcomes, and has an accountable team that improves it after launch.
A communications site can still be excellent. It may explain an offer, publish a point of view, or support a campaign. Product management becomes necessary when the web experience carries more responsibility. It may expose changing inventory, guide a complex decision, move a user into an application, support a marketplace exchange, or connect several teams through shared content and components.
The practical distinction is ownership. A project ends when the site ships. A product keeps a problem, a user, a roadmap, measures, guardrails, and a team after launch.
Decision framework
The Web Product Test
The Web Product Test is Nenad Ivanovic's five-gate framework for deciding how much product management a website needs. It is practical decision guidance, not an industry maturity score.
- 01ValueStart with a useful job.
- 02ActionHelp the user make progress.
- 03SystemConnect the experience to the system.
- 04SignalTurn behavior into evidence.
- 05StewardshipKeep ownership after launch.
Value
Question: Does the experience help a defined user complete a useful job?
Evidence: User research, search behavior, task completion, service outcomes, or a clearly observed decision need.
Warning: The page exists because the organization wanted to publish it.
Action
Question: Can the user do something consequential, rather than only read a message?
Evidence: Search, compare, apply, book, configure, transact, submit, save, or move into a connected product flow.
Warning: Calls to action discard context in a generic contact form.
System
Question: Does the website connect to product data, workflows, services, or reusable design patterns?
Evidence: Shared data, APIs, content models, design systems, structured content, analytics definitions, and governed integrations.
Warning: The website presents stale information or rebuilds the same pattern for every initiative.
Signal
Question: Does behavior create evidence that changes a product, growth, or service decision?
Evidence: Defined events, task success, qualified progression, search demand, user feedback, and guardrail measures.
Warning: Reporting counts visits and releases without showing what changed for the user or business.
Stewardship
Question: Is a team accountable for quality and improvement after launch?
Evidence: A named owner, roadmap, decision rights, contribution rules, accessibility and performance budgets, search integrity, privacy, security, and incident ownership.
Warning: Ownership ends when the redesign is delivered.
4 or 5 gates
Manage the website as a product.
2 or 3 gates
Use a hybrid model and define the boundaries.
0 or 1 gate
Keep a communications model with clear quality ownership.
Criteria for a product-led web experience
| Criterion | What good looks like | Useful evidence |
|---|---|---|
| User and job | One primary user and job are clear for each important journey. | Research, search intent, service data, and support themes. |
| Useful action | The web experience helps the user make meaningful progress. | Task completion, qualified progression, and successful handoff. |
| Product continuity | The website and wider product behave like one system. | Shared terminology, identity, data, state, and interaction rules. |
| Connected data | Dynamic information stays current and trustworthy. | APIs, governed CMS models, sync health, and named ownership. |
| Design system | Teams reuse patterns without losing necessary context. | Adopted components, content models, standards, and contribution rules. |
| Learning system | Teams can connect behavior to decisions. | Events tied to hypotheses, outcomes, and guardrails. |
| Quality guardrails | Speed preserves accessibility, performance, search, privacy, and security. | Budgets, checks, audits, and review owners. |
| Ongoing stewardship | A team owns improvement after launch. | Roadmap, service levels, decision rights, and incident response. |
A design system supports this model when it reduces repeated decisions and protects coherence. A library of components is not enough. Product-led design systems need clear ownership, contribution rules, accessibility standards, content guidance, release practices, and evidence that teams can use them successfully.
The same applies to analytics. More events do not create more learning. A useful measurement plan begins with the decision a team needs to make, the user behavior that can inform it, and the guardrails that prevent a local improvement from harming trust or product quality.
Evidence in practice
Nursa's website became part of the healthcare marketplace
At Nursa, the website had to do more than describe a healthcare staffing platform. It needed to make changing marketplace information discoverable, connect website and application journeys, support search at scale, and let several teams improve the experience without routing every change through the core engineering organization.
Webflow's official Nursa customer story states that the team migrated more than 40,000 pages in 19 days. The product lesson is not the platform or speed alone. The work combined information architecture, CMS modeling, URL planning, search preservation, performance, reusable components, APIs, analytics, and operating ownership.
Verified scale
40,000+
Pages in the completed migration, confirmed by Webflow.
Verified duration
19 days
The complete migration duration, confirmed by Webflow.
In Webflow's official enterprise technology stack session, Nenad describes a staged decision process. The team first tested whether it could build a converting, search-ready experience. It expanded the system as new questions appeared around conversion, shared data, app attribution, security, performance, and CMS scale.
That sequence matters. A working pilot produced evidence before broad commitment. Low-value legacy content was removed instead of copied. Internal developers, Webflow's enterprise team, and implementation partners had distinct responsibilities. Website and app data moved toward a more continuous user experience.
Read the focused evidence behind the migration and operating model.
See the Nursa website case studyProduct leadership connects the web experience to the business
A product-led website crosses organizational boundaries. Product leaders define the user and outcome. Designers shape comprehension, action, and system coherence. Engineers own technical integrity and integration. Growth and content teams develop discovery and learning. Security, privacy, accessibility, operations, and brand add constraints that must be visible before release.
Nenad is Vice President of Product at Nursa. His verified work spans product leadership, growth systems, UX, design systems, information architecture, analytics, and enterprise web platforms. Webflow selected him for the 2023 Chicago opening keynote and invited him back in 2024 to present Nursa's enterprise technology stack with Nursa CEO Curtis Anderson and Webflow's Peter Hsieh. The speaking archive records that public evidence.
His publications and media archive separates authored work from official customer stories and independent coverage. His portfolio shows the broader product and design record, while case studies document end-to-end product work. The platform product management pillar explains how repeated proof becomes reusable, governed capability.
Seven ways websites stop behaving like products
Failure 01
A campaign calendar becomes the roadmap
Publishing volume replaces a clear user problem, product decision, and measurable outcome.
Failure 02
A redesign changes appearance, not the job
The interface changes while task success, system constraints, and accumulated search value remain unexamined.
Failure 03
The design system becomes a gallery
Components exist without ownership, contribution rules, accessibility guidance, release practices, or adoption.
Failure 04
The website and product tell different truths
Data, terminology, availability, or workflow state drift across surfaces and weaken trust.
Failure 05
Activity becomes evidence
Page count, releases, tests, and visits rise without a link to successful user or business outcomes.
Failure 06
Autonomy expands without guardrails
More people can publish, but nobody protects performance, accessibility, analytics, search integrity, privacy, or security.
Failure 07
Ownership ends at launch
The system ships without a roadmap, maintenance budget, quality owner, or accountable steward.
Decision in one line
The website product decision
If the website performs valuable work, shares systems with the product, changes decisions through evidence, and needs ongoing stewardship, manage it as a product.
When only some of those conditions are true, use a hybrid model and define the boundaries. The goal is not to make every website more complex. It is to match the operating model to the responsibility the experience already carries.
Continue through the evidence