What is a digital audit?

What you will learn

By the end, you will understand:

Distinguish an audit from a tool report, score or sales pitch
Define scope, criteria, evidence and limitations
Write findings that connect observation to business action

Visual explainer

See the idea clearly.

A digital presence passes through a review lens; tools, manual checks, user tasks and business records combine into findings with evidence, impact, confidence, owner and next action.

An audit answers a decision question

An audit is useful when it supports a decision: why enquiries are falling, whether a website is ready for advertising, where search visibility is blocked, or which improvements deserve funding first.

Without a question and scope, an audit becomes an uncontrolled collection of screenshots and opinions.

A score is not an audit

Tool report

Measures the checks that tool can observe under its configuration and moment in time.

Audit

Combines suitable tools, manual review, business context, user tasks and limitations to reach accountable findings.

The audit chain

01
Decision question
02
Scope and exclusions
03
Criteria and baseline
04
Evidence collection
05
Verification/triangulation
06
Finding
07
Business/user impact
08
Confidence and limitation
09
Priority
10
Owner and next action

Every finding needs anatomy

Part	Question
Observation	What was actually seen or measured?
Evidence	Where, when and how was it checked?
Criteria	What requirement, objective or good practice is relevant?
Impact	Who or what may be affected?
Confidence	How certain is the conclusion and what remains unknown?
Recommendation	What specific outcome should change?
Ownership	Who decides/implements, by when and how is completion verified?

Evidence has different strengths

A single automated warning is a lead to investigate. Repeated field data, reproduced user-task failure, verified configuration and downstream business records usually provide stronger evidence.

Triangulation means checking a claim through more than one suitable source. Agreement raises confidence; disagreement is itself a finding to explain.

Audit ethics

Permission and scope agreed
No unnecessary personal data
Production tests are non-destructive
Credentials never copied into report
Claims separate fact from inference
Competitor data used lawfully
No invented metrics
Conflicts disclosed
Limitations visible
Sensitive findings shared securely

Real-world example

Example: a low performance score becomes a verified finding

Example

A mobile Lighthouse run reports slow image loading. The reviewer repeats the lab test, checks PageSpeed field data, inspects image sizes and tests a real mobile device. The finding states which pages and users are affected, notes that field data covers a broader population, and recommends responsive image delivery with a measurable verification step.

Try this

Turn one complaint into an audit question

Start with “the website is bad” or another vague complaint. Rewrite it as a scoped question with user group, journey, device/market, time period, evidence sources and decision the audit should support.