Quality assurance for agencies: A scalable QA process

An agency quality assurance process is a set of checks at five stages (brief, self-review, peer review, final review and client review) with a checklist for each discipline and a log of every defect a client catches. Picture what happens without one: a homepage redesign ships on Friday at 5pm, and by Monday the client has emailed eleven errors, from typos in the hero to a broken pricing-page link and an SEO title that still says "Untitled Page". The team spends two days on emergency fixes, the account lead spends an hour apologizing, and the next quote gets negotiated down over "trust issues". That is what poor QA costs: not a single bug, but a compounding tax on margin, renewals and reputation. This guide breaks down QA stage-by-stage across an agency's delivery lifecycle, per-discipline checklists for design, copy, dev, and video, and a defect-tracking system that turns every catch into a process improvement.

The bottom line:

• Agency QA is not one step; it is a 5-stage pipeline from brief through delivery, with progressively tighter checks
• Per-discipline checklists matter more than universal ones; a copy QA list and a dev QA list share almost no items
• The later a defect is caught, the more it costs: a typo fixed at brief stage takes seconds, the same typo on a live site takes hours and an apology
• Most agency defects trace to four root causes: rushed handoffs, unclear briefs, missing checklists and an overloaded team
• A working defect-tracking system is the single highest-leverage investment in agency quality

The agencies that ship clean work consistently are not the ones with more talented people; they are the ones with a process that does not depend on individual heroics.

The real economics of agency quality defects

Software teams have long observed that defects cost more the later they are found. Here is how that plays out at an agency. The multipliers are illustrative, to show the shape of the curve.

The math is brutal: a single typo caught at brief stage might cost 30 seconds; caught after launch, that same typo can require 2 to 4 hours of emergency response, client apology, fix, redeploy, and post-mortem.

For example, if each client-caught defect costs a 15-person agency 10 to 15 hours of fixes, calls and re-review, then 4 to 6 of them a month is 40 to 90 hours of lost capacity, plus relationship damage nobody tracks.

The 5-stage agency QA pipeline

Effective agency QA is not a step before delivery. It is a pipeline with checks at five stages, each catching different kinds of defects.

Stage 1: Brief QA (the defect that costs nothing to fix)

Most quality problems originate at the brief, not at execution. A vague brief produces work that misses the mark, which produces revisions, which produces rushed re-work, which produces defects.

Brief QA checklist (every brief, before kickoff):

• Is the deliverable list specific (count, format, dimensions)?
• Is success defined in measurable terms?
• Is the target audience explicit?
• Are competitive and reference materials included?
• Is the approval chain explicit?
• Are out-of-scope items explicitly listed?
• Is the timeline realistic, with buffer for revisions?
• Is the responsible owner on each step named?

A 15-minute brief review by the strategist or AM before the team starts production prevents an average of 1 to 2 mid-project pivots, each of which can cost 6 to 12 hours.

Stage 2: Self-review (the creator)

The creator runs a discipline-specific self-review before passing work on. This is the cheapest catch in the pipeline and the most frequently skipped under deadline pressure.

Stage 3: Internal peer review

After self-review, work goes to a same-discipline peer (designer reviews designer, dev reviews dev). The peer is not imposing taste; they are running a checklist.

Stage 4: Account / final QA

Before anything goes to the client, an AM or senior reviewer runs a final pass from the client's perspective. They are not checking the same things the peer reviewer did; they are checking scope match, narrative flow and whether the package feels considered.

Stage 5: Pre-delivery / launch QA

For deliverables going live (websites, campaigns, video airings), a final environment-aware QA is non-negotiable. This is where you catch the broken redirect that worked on staging but not production.

Per-discipline QA checklists

The single biggest mistake agencies make is using a universal checklist. The checklists below are discipline-specific and represent the minimum viable list per discipline.

Design QA checklist

Brand and consistency:

• Colors match approved palette to hex (no eyeballing)
• Typography hierarchy follows style guide
• Logo clear-space and minimum-size rules respected
• Iconography and illustration style consistent
• Photography matches approved brand direction

Layout and craft:

• Alignment to 4 or 8 px grid throughout
• Spacing consistent (no random margins)
• Typography optical adjustments applied
• No widows or orphans in headlines
• Text contrast meets WCAG AA at minimum

Responsive:

• All specified breakpoints reviewed (typically 1440, 1024, 768, 375)
• Touch targets minimum 44 x 44 px per Apple HIG
• No content cut off, hidden, or unreadable at any breakpoint
• Critical CTAs visible above the fold at every breakpoint

File hygiene:

• Layers named and organized
• Components used (not detached copies)
• Final assets exported at correct DPI and format
• Naming convention followed in deliverable files

Copy QA checklist

Accuracy:

• All facts, stats, claims verified with source links
• Names, titles, dates, prices verified against source
• Legal disclaimers and required disclosures included
• Trademark and registered marks placed correctly

Voice and tone:

• Match the brand voice guide (cross-checked, not assumed)
• Industry terminology used correctly
• Audience reading level appropriate (Hemingway score or similar)
• No internal jargon that confuses the audience

Mechanics:

• Spelling and grammar passed through two checks (one human, one automated)
• Punctuation consistent (Oxford comma decision applied consistently)
• Capitalization rules followed (title case vs. sentence case decided once)
• No double spaces, smart quotes consistent

Format-specific:

• Headlines under defined character limit
• Meta descriptions 140 to 160 characters
• CTAs action-verb-led, not "Click here"
• Body copy paragraphs scannable, not wall-of-text

Web dev QA checklist

Functional:

• All nav links work and go to intended destination
• All forms submit successfully and trigger expected actions
• All interactive elements behave per spec
• 404 page renders correctly for bad URLs

Cross-browser and device:

• Latest Chrome, Safari, Firefox, Edge on desktop
• iOS Safari and Android Chrome on mobile
• One older browser version per stated browser support matrix

Performance:

• Largest Contentful Paint under 2.5 seconds
• Cumulative Layout Shift under 0.1
• First Input Delay under 100ms (or INP under 200ms)
• Lighthouse score above 85 for performance, accessibility, SEO

Accessibility:

• Automated scan via axe or WAVE passes with no critical issues
• Keyboard nav reaches every interactive element
• Screen reader test on critical flows
• Alt text on every image that conveys meaning

Security and integrity:

• SSL valid and HSTS configured
• No secrets in client-side code
• Form inputs sanitized
• Dependencies scanned for known vulnerabilities

Pre-launch only:

• Redirects mapped and tested
• Analytics firing on key events
• Production vs. staging environment confirmed
• Backup and rollback plan documented

Video and motion QA checklist

• Audio levels normalized (dialog around -12 LUFS, peaks not clipping)
• Caption file (SRT or VTT) accurate and synced
• Color graded against brand reference
• Frame rate and resolution match delivery spec
• Lower-thirds and supers spelled correctly
• Logo placements per brand book
• Final export tested in actual delivery platform (YouTube, broadcast, social)
• Versioned files named per convention with date and version

Defect tracking: The system that turns catches into improvements

A defect log is not bureaucracy. It is how you turn every QA catch into a process improvement, and how you spot patterns across projects.

Minimum defect log fields:

• Date caught
• Project / client
• Stage caught (brief, self, peer, final, client, post-launch)
• Discipline (design, copy, dev, video)
• Defect category (typo, alignment, broken link, scope mismatch, etc.)
• Severity (critical, major, minor, cosmetic)
• Time to fix (estimated)
• Root cause (rushed, brief unclear, missing check, etc.)

Run a monthly defect review where you look for patterns:

• Which categories appear most often?
• Which stages are leaking the most defects to client review?
• Which projects had outlier defect counts (and why)?
• Which root causes recur?

Run this for 90 days and you will usually find that a handful of defect categories account for most client-caught issues. Once named, they can be cut sharply with targeted checklist updates.

The four root causes of agency defects

Across most defect logs we have seen, four root causes dominate.

1. Rushed handoffs. Designer to developer, copy to design, draft to QA. Most handoff defects are caused by missing context, not missing talent. Solution: a mandatory handoff checklist that has to be completed before the task moves states.

2. Unclear or incomplete briefs. Briefs that miss success criteria, scope edges, or approval chains. Solution: Stage 1 brief QA.

3. Missing or unused checklists. Either the checklist does not exist, or it exists and people skip it under deadline pressure. Solution: bake checklists into the project management tool as required subtasks.

4. Sustained high utilization. This is the structural one. When a team runs above roughly 85% utilization for weeks at a time, QA is the first thing cut. The fix is staffing and capacity planning, not process.

For utilization specifically, see the agency capacity planning guide.

QA time budget by project type

QA must be scheduled into the project, not absorbed by the team's evenings.

These are starting points; adjust them once your defect log shows where problems actually come from. When QA gets cut to fit a deadline, the defect rate climbs and the rework debt comes due later, usually with a client apology attached.

Example: rolling out the pipeline in 8 weeks

This is a hypothetical example to show the sequence and the math. Say a 24-person digital agency averages 9 to 12 client-caught defects per major delivery and spends 60 to 90 hours a month on emergency fixes.

It rolls out the 5-stage pipeline over 8 weeks:

• Week 1 to 2: Built discipline-specific checklists, attached as required subtasks in their PM tool
• Week 3 to 4: Trained team on stages, started defect log
• Week 5 to 6: First monthly defect review shows most defects fall into two categories: broken links and brand color mismatches
• Week 7 to 8: Targeted checklist updates plus a 5-minute pre-handoff "color/link scan" specific to those two categories

If that cuts emergency-fix work by 65 hours a month, that is 780 hours a year; at a loaded cost of $75 an hour, about $58,000 of capacity recovered, before counting the effect on client trust.

Scaling QA by agency size

Under 10 people. QA is embedded in the producer's role, with a senior eye on the final step. Checklists matter most. A "buddy QA" pair model works.

10 to 30 people. A part-time or rotating QA lead role emerges. QA time is budgeted explicitly. Defect log is monthly.

30 to 50 people. Dedicated QA function or specialists per discipline. Automated tooling for dev (Lighthouse, axe, link checkers in CI). Quarterly QA retros.

50+ people. Full QA team, often with discipline-specific leads. QA metrics in performance dashboards. Cross-project quality standards governed by an ops or production director.

Common QA failure modes

Checklist theater. Boxes checked without actually checking. Solution: random spot-audits where a senior re-runs a checklist and compares findings.

QA as last-step bottleneck. If everything piles up at QA the week of delivery, you have a planning problem, not a QA problem.

QA without authority. A QA reviewer who cannot block delivery is decorative. Make the sign-off binding.

No defect log. Without a log, every catch is a one-off. With a log, every catch is data.

Blaming individuals. A defect-rich week is a process signal, not a person signal. The Deming "plan-do-check-act" mindset matters more than blame.

Building a quality culture that scales

Process alone does not produce quality. Culture has to value catches over output.

• Celebrate the catch, not just the ship. The designer who caught the wrong logo at peer review prevented a $4,000 reprint.
• Run blameless post-mortems on every client-caught defect. The question is "what in our process let this through," not "who did this."
• Make checklists living documents. Every new defect category adds a line. Stale checklists get pruned.
• Tie quality metrics to the project P&L. Show the team how rework hours destroy margin.

Quality is a leading indicator of retention: clients rarely leave over one mistake, but they do leave over repeated ones. Quality problems are the most common exit reason in consumer research too: in PwC's 2025 Customer Experience Survey, 52% of consumers said they had stopped buying from a brand because of a bad experience with its products or services.

Frequently asked questions

How much time should agencies budget for quality assurance?

Plan 12 to 20 percent of project hours for QA, scaled to risk: landing pages near the low end, multi-channel campaigns and web apps near the high end. Much lower and defects tend to reach clients; much higher usually signals process inefficiency rather than craft. Bake QA in as line items, not as the buffer that gets squeezed.

Who should do QA on agency deliverables?

Use separate reviewers from the creator at every stage past self-review: a same-discipline peer for craft, then an AM or senior for the client perspective. Creators miss their own errors far more often than a second reviewer does, which is why each layer catches different problems. Skip a layer and its misses go to the client.

What is the most common cause of agency quality issues?

Sustained high utilization combined with deadline-driven QA cuts. When the team is structurally over-committed, the checklist gets skipped, the handoff is rushed and the brief is glossed. The fix is capacity planning, not better checklists; checklists do not help when there is no time to run them.

Should QA be a separate role or part of the producer's job?

Under 20 people, embed QA in producer roles with a senior reviewer on final QA. Past 20, a part-time dedicated QA lead pays off. Past 40 to 50, expect at least one full-time QA role per discipline cluster. The agency hiring guide covers the sequencing for production roles.

How do you build a QA culture without slowing the team down?

Celebrate catches, run blameless defect reviews, and make checklists part of the workflow rather than something separate. Pair this with realistic timelines that include QA from the start. The culture follows the system: when leadership protects QA time even under deadline pressure, the team learns it is non-negotiable.

Want QA steps built into every project? In AgencyPro project management you can add review stages as tasks on each project, and versioned proofing and approvals (Pro plan and above) keep client review in one place. See plans or book a demo.