Does the right audience encounter the right surface?
MACHINE EXPERIENCE / JOURNEY-SPECIFIC
The website is not the unit. The journey is.
People do not experience a brand on one page. They move through search, AI answers, links, screens, choices, and actions. MX follows that path, finds where meaning or momentum breaks, and turns the finding into a measurable improvement.
Can people and machines identify the organization, offer, and relevance?
Does the journey provide what a real choice requires?
Can the intended next step be completed clearly across context and device?
Can a change be rerun against the same journey and a meaningful outcome?
WHAT AN MX AUDIT INCLUDES
One journey. A bounded record. A change you can rerun.
The brief fixes the audience, task, public boundary, and intended outcome before measurement begins. Exact surfaces and volume are confirmed after human review; submitting a brief creates no automatic scope, price, or commitment.
Freeze the decision.
Define the audience, need, starting condition, intended choice, and intended action so every observation answers the same question.
Map the visible system.
Retain the relevant pages, templates, indexing conditions, and public handoffs inside an explicit observation boundary.
Follow the real path.
Replay the named, public, read-only link path in both viewports and retain the screens, states, and outcome encountered.
Keep environments distinct.
When agreed in scope, observe named search or AI environments separately. One engine, prompt, or result is never generalized into universal behavior.
Change what matters.
Translate consequential friction into a small set of evidence-linked changes, with unknowns and evidentiary limits kept visible.
Define the proof before the fix.
State what should change, which journey condition should respond, and what the same bounded rerun must observe afterward.
A consequential journey crosses surfaces and needs a baseline.
Use MX when search, AI, pages, devices, or handoffs jointly shape whether the right person can understand, decide, and act.
A guaranteed ranking, universal score, or automated certification.
MX is not a generic traffic dashboard, exhaustive compliance audit, or substitute for evidence the available sample cannot support.
THE UNIT OF ANALYSIS
Follow the decision all the way through.
Observation stays separate from inference.
MX records what a person or named machine environment actually encountered. It does not turn the presence of a tag into a claim about ranking, interpretation, or outcome.
The handoffs matter as much as the pages.
The audit follows one important need across entry points, representations, devices, and actions. A technically healthy page can still fail the journey around it.
Each priority ends in a condition that can change.
The recommendation identifies what to alter, which part of the journey should respond, and what a rerun would need to observe before the work is considered useful.
THE AUDIT RECORD
What leaves the room.
A useful audit is not a pile of checks. It is a shared record of the journey, the friction worth changing, and the evidence needed to know whether the change helped.
Priority journey
The audience, need, starting conditions, surfaces, intended decision, and intended action.
Observed path
What people and relevant machine systems encountered, with unknowns kept visible.
Improvement priorities
The few changes most likely to remove consequential friction—not an undifferentiated issue list.
Rerun conditions
The same journey, repeated after change, with the behavior or outcome to compare defined in advance.
BASELINE → CHANGE → RERUN
Improvement should be observable.
An MX intervention is deliberately bounded: establish the journey as it works now, change one meaningful condition, rerun the same journey, and compare what happened. When traffic, instrumentation, or sample size cannot support a causal claim, the result stays directional. The method should not manufacture certainty.
START AN MX BRIEF →METHOD NOTES
What MX does—and does not do.
What is MX?
MX is 99xlab’s machine experience practice. It examines how a consequential customer journey works when people, search systems, AI answers, websites, and devices all shape what gets found, understood, chosen, and done.
How is this different from an SEO or conversion audit?
SEO, accessibility, analytics, content, and conversion may all be relevant, but none is the unit of analysis. MX follows one priority journey across those boundaries and asks whether the experience helps the right visitor make the intended decision and complete the intended action.
Why is there no single MX score?
A single number hides which journey, audience, surface, or outcome failed. MX keeps observations separate, states what was and was not tested, and prioritizes changes against the decision and outcome that matter.
What does an MX Audit produce?
A defined journey and baseline, observed friction and unknowns, a prioritized improvement plan, and clear conditions for rerunning the same journey after a change. The exact record depends on the brief rather than a fixed dashboard template.
Does the audit observe AI and search systems directly?
When the agreed journey includes them, named environments can be observed and recorded separately. MX does not treat one engine, one prompt, or one result as proof of universal machine behavior.
Does MX certify that every published claim is true?
No. MX evaluates the outside-in experience created by public surfaces. It does not certify the underlying operational truth of an organization, guarantee a ranking, or promise a business outcome.
Who reviews the work?
A person reviews the journey, observations, interpretation, priorities, and retest plan. Automated checks can support the work; they do not issue the conclusion.
What does it cost?
There is no public price. The audit is scoped around the organization, priority journey, surfaces, and measurement needs after a person reviews the brief.
HUMAN-REVIEWED / JOURNEY-SPECIFIC / NO AUTOMATED COMMITMENT