Agency Case Study: Running 50 Site Audits in a Month

# Agency Case Study: Running 50 Site Audits in a Month

A three-person marketing agency we work with set a goal that sounded slightly insane on paper: deliver 50 website audits to small business clients in one calendar month. Not 50 quick scans. Fifty actual deliverables — written summaries, prioritized fix lists, and a 30-minute review call with each client.

They hit the number. They also kept their existing retainer work moving and didn't burn anyone out. This article walks through how they did it, what they cut, and the template you can copy if you want to run audits at volume without producing junk.

If you run an agency, freelance for several clients, or just want a saner way to handle audits for your own portfolio of sites, the pattern here works.

The starting point

Before the sprint, the agency was doing audits the way most small shops do them: each one took 6 to 10 hours of human time, screenshots were copied into a slide deck by hand, the recommendations section was rewritten from scratch every time, and clients waited two to three weeks for a report.

That math doesn't scale. At 8 hours per audit, 50 audits would have eaten 400 hours — more than the team had in a month even if they did nothing else.

Their problem wasn't talent or tools. It was that every audit was being treated as a custom artisan project. Most of the work was repeated work, dressed up to look bespoke.

The reframe

The first decision was the most important one. They stopped pretending each audit was unique.

In reality, about 80% of what a small business website needs to fix falls into a short list of categories:

• Slow load times, usually from oversized images or render-blocking scripts
• Missing or broken metadata (titles, descriptions, Open Graph tags)
• Pages with no internal links pointing to them
• Mobile layout breakage
• Forms or buttons that don't work or are hard to find
• Thin or duplicate content on key landing pages
• Missing structured data on articles and products

If you build your audit around those categories instead of around the specific website, you can move much faster without losing accuracy. The remaining 20% — the truly site-specific findings — still gets human attention. But the foundation is templated.

This isn't a shortcut at the client's expense. Templating the common stuff frees up brain time for the genuinely tricky issues that actually move the needle.

The five-stage workflow

The agency broke each audit into five stages and assigned a strict time budget to each. The whole pipeline targets about 90 minutes of human time per audit, down from 8+ hours.

Stage 1: Automated crawl (5 minutes, mostly waiting)

Each client URL gets dropped into an audit tool that handles the deterministic checks — speed, Core Web Vitals, broken links, missing tags, mobile usability, structured data validation. This is the part that doesn't need a human and shouldn't have one.

The agency used our site audit tool for this stage because the output was already organized into the categories they cared about. The point isn't which tool you pick — it's that you stop manually checking things a script can check.

Stage 2: Manual eyeball pass (15 minutes)

This is the part that still needs a person. The auditor opens the site on desktop and mobile and walks through it like a customer would:

• Can I find what this business does in three seconds?
• Is the main call-to-action obvious above the fold?
• Does the contact form actually work?
• Does the site feel trustworthy?
• Is the page about the customer or about the company?

These are judgment calls. No tool can reliably answer them. But 15 minutes of focused attention catches the issues that customers notice and crawlers miss.

Stage 3: Prioritization (10 minutes)

Every finding gets a tag: fix now, fix this quarter, or nice to have. Three buckets, no exceptions.

The "fix now" bucket gets a hard cap of five items. If you give a small business owner a list of 30 things to fix, they'll fix zero of them. If you give them five, they'll fix three. Volume of recommendations is inversely correlated with action.

Stage 4: Report assembly (25 minutes)

The report itself uses a fixed template with these sections:

1. What's working — three honest positives. Always start here.
2. What's costing you customers right now — the top five "fix now" items, each with a one-sentence explanation of the business impact.
3. What to fix this quarter — five to ten items with brief explanations.
4. What's optional — everything else, listed plainly.
5. What we'd do first if this were our site — one paragraph of opinion.

That last section is what clients consistently said they valued most. It's not in the automated output. It's a human telling them what they would actually do. Two to three sentences. No hedging.

Stage 5: Client review call (30 minutes)

A 30-minute call walking through the report. Not a sales pitch. Just answering questions, clarifying technical items, and confirming what the client plans to act on.

Total human time per audit: roughly 80–90 minutes. Total elapsed time from request to delivery: about three business days.

What they cut

Most of these are habits agencies cling to without thinking about why:

• Custom cover pages with the client's branding. Took 20 minutes per report. Nobody ever mentioned them.
• Screenshots of every issue. Replaced with a single annotated screenshot per priority issue, max. Clients couldn't process more.
• Long executive summaries. Replaced with a five-bullet summary at the top. Anything longer got skimmed.
• Recommendation paragraphs for every finding. One-sentence explanations for "fix now" items, bullet labels for everything else.
• Detailed competitor sections. Replaced with a single comparison line where relevant. Most small businesses don't need a competitive analysis bolted onto a tactical audit.

What they kept: the human judgment call, the prioritization, the honest "what's working" section, and the call.

A real walkthrough

Here's one of the 50 — a regional plumbing company with about 40 pages, ranking on page two for most of their service area terms.

Automated crawl found:

• Largest Contentful Paint of 4.8s on mobile, well above the 2.5s threshold defined in the Core Web Vitals guidance
• 14 images over 500KB
• Missing meta descriptions on 22 pages
• No structured data on service pages or blog articles
• Two broken internal links

Manual pass found:

• Phone number wasn't tap-to-call on mobile (huge for a plumbing business)
• Contact form had eight required fields when three would do
• The homepage hero said "Quality Service Since 1987" instead of saying what they do or where
• Service area was listed only in the footer

Prioritized to five "fix now" items:

1. Make the phone number tap-to-call sitewide
2. Rewrite the homepage hero to "Emergency Plumber in [City]" with a clear call-to-action
3. Compress the 14 oversized images and serve them in modern formats
4. Cut the contact form from 8 fields to 3
5. Add service area and city names to the H1 of each service page

"What we'd do first" paragraph:

> If this were our site, we'd do items 1, 2, and 4 this week. They take an afternoon and they directly affect whether someone in a flooded basement calls you. Everything else can wait.

The client did all five within two weeks. Their next month showed a measurable lift in calls from the website. Not because the audit was magic — because the audit was short enough to act on.

The structured data note

About 30 of the 50 sites had blog content or articles with no structured data at all. Google's article structured data documentation explains the markup, but the short version: if you publish articles and you're not telling search engines they're articles, you're leaving rich result eligibility on the table.

This was a near-universal "fix this quarter" recommendation. It's low-effort and the upside is real.

The content quality question

Six of the 50 sites had a deeper problem that no checklist catches: their content was thin, generic, or clearly written for search engines rather than for readers. Google's helpful content guidance is the right reference here, and it's worth reading if you write or review content for client sites.

For those six, the audit included a frank section: "Your technical issues are minor. Your content is the problem. Here's what we'd rewrite first." Three of the six pushed back. Three engaged and ended up commissioning content work. Sometimes the most useful audit finding is the one that says your stack isn't broken, your story is.

What changed for the agency

A few things were obvious after the sprint:

• Audits became a product, not a project. Fixed scope, fixed timeline, fixed price. Clients liked the predictability.
• The five-item "fix now" cap raised completion rates. Clients actually did the work, which made the agency look effective in a way that 30-item lists never did.
• The 30-minute call became the most valuable part. More than the document itself.
• Junior team members could run audits. Once the workflow was templated, training a new auditor took days, not months.

The mini-checklist if you want to copy this

• [ ] Pick one audit tool that handles the deterministic checks and stick with it
• [ ] Build a fixed report template with exactly five sections
• [ ] Cap "fix now" items at five
• [ ] Write a one-paragraph "what we'd do first" opinion for every audit
• [ ] Time-box each stage and respect the budget
• [ ] Cut anything clients don't read (custom branding, executive summaries longer than five bullets, screenshot-heavy appendices)
• [ ] End with a 30-minute call, not an email

For agencies running the workflow at volume, the pattern compounds — see how it fits into agency-specific workflows and where Core Web Vitals fixes tend to land in the priority stack.

What this means for you

If you run one website, the same logic applies in miniature. The reason most small business sites stay broken isn't that the owner doesn't care. It's that nobody ever gave them a short, prioritized list of what to fix first. Audit reports with 47 items get ignored. Reports with five items get acted on.

You don't need an agency to do this for yourself. You need a crawl, a 15-minute honest look, a prioritization step, and a willingness to ignore the items that don't matter this quarter.

Run a free audit on your own site

If you want to see what the deterministic part of this workflow looks like for your own site, run a free audit with FreeSiteAudit. You'll get the same crawl-level findings the agency above used as the foundation for every one of their 50 deliverables — load speed, Core Web Vitals, broken links, missing metadata, structured data gaps, and mobile usability issues. From there, do the 15-minute eyeball pass yourself, cap your fix list at five items, and start with the one you'd be embarrassed for a customer to notice.

That's the whole method. Five items, three buckets, one opinion. Repeat as needed.

Sources

• Google Search Central: Creating helpful, reliable, people-first content
• Google Search Central: Article structured data
• web.dev: Core Web Vitals