What to Ask Before You Audit

Before you even begin the audit process, you need to make sure you’re collecting the right kind of data for the right reasons. Set yourself up for success by answering three critical auditing questions.

I love website audits. I really do. They may involve long hours and longer spreadsheets, but fiddly details are my jam. To those of you who don’t share my completionist streak, I say the same thing that a Parisian man once said to my mother when she told him she didn’t speak French: “Oh, I am so sorry for you.”

Audits impose order on chaos. They give shape to the vague, impenetrable monolith of “the website.” They create space for information about information, and they compile it in such clearly structured ways that nerds like me can’t help but fall in love with the process. There are concrete metrics, regimented criteria, blanks to be filled. I like filling in blanks. I like winning at very small tasks.

All the data you can collect from website audits—from the tiniest atoms of content to the biggest strategic views—drives your strategy forward. Audits helps you define scopes, target vulnerabilities, and identify strengths. They give you the information you need to make sound project decisions.

But to get that information—before you even begin the audit process—you need to make sure you’re collecting the right kind of data for the right reasons. Set yourself up for success by answering three critical auditing questions.

1: Why are you auditing?

Start by making sure you have a good understanding of the parameters of your project. What are the circumstances that led to you saying, yes, the best thing I can do with my life right this minute is to scrutinize every page of my website? (Aside from the career decisions that you may suddenly find yourself questioning.) What larger project will this audit support, and how do you know an audit is needed?

Here are some common (though by no means exhaustive) circumstances that suggest an audit is in order:

  • Your website is going through a redesign.
  • Your website is migrating to a new CMS.
  • Your organization is rebranding.
  • Your organization is adding significant amounts of new content or merging with another website.

In other words: you need an audit anytime you need to know a lot about the website as it currently exists in order to make decisions about its future state. And unless you’ve been maintaining a content inventory on a rolling basis—that is, you already have an accurate database of your pages—building that inventory is your first step.

(Quick sidebar: “content audit” and “content inventory” are not interchangeable terms! An audit is a process; an inventory is a product. An audit is the action of reviewing a website; an inventory is the spreadsheet of data that results from the audit. It may sound fussy, but making a distinction between process and product—and building a shared vocabulary—helps your clients and colleagues to better understand what the work is, and how and why you do it.)

Understanding the why behind your project also means looking at the project timelines, the people involved, and how the audit results might impact later phases. In other words, you need to explore the business impetus behind the project: what questions are your project stakeholders trying to answer?

Knowing this gives you context to begin asking your own questions—those about the audit itself.

2: What are you auditing?

What’s being audited, and by what criteria? What does the audit need to do? What kind of data does it need to collect? What kind of audit is it?

Traditionally (for as much tradition as any web work can claim), we tend to talk about audits as either qualitative or quantitative—as if audits only tell us how good the content is, or how much content there is, with little shared between them.

But (drumroll), that’s a false dichotomy—all audits have a mix of qualitative and quantitative factors. (Not to mention, not all audits are focused on content!) What you’re trying to learn from the audit is a much more definitive guide for the kind of audit you’ll conduct.

Purpose is the watchword (which I talked about, very briefly, in Gather Content last December). What’s the purpose of your audit? Be as specific as possible in answering that question. The answer should never be as simple as “We want to know what content we have on the site”—what about that content do you need to know?

Do you need a page count? Do you need to triage some sections over others for rewriting? Do you need to identify dead-end paths to make the case for another project? Are you checking image metadata as part of an accessibility evaluation? Are you comparing current voice and tone to new brand guidelines?

These are all perfectly reasonable audits that could happen within the context of a redesign, migration, or rebranding project. And they would each have wildly different criteria and wildly different results—so purpose is what selects your lens.

The more specific you can get about what you’re measuring and why, the more useful your data-gathering process (and the data itself) will be.

3: How are you auditing?

If you know why you’re auditing (from a project perspective), and what you’re auditing (from a purpose perspective), the next logical question is: how? A manual audit? An automated audit? Both?

Automated audits (sometimes called crawls) are conducted by a robot—software that crawls a given domain and returns data about the pages found there. Manual audits are conducted by, well, you.

I like to run automated audits as soon as possible when I’m working on a site redesign—even as early as during the pitch process, since crawls are the best way to determine the size of a site, which can impact project scoping. Could the client tell me how many pages their site has? Yes, of course. Has the client ever, ever, ever, in the history of anything, ever provided me with an accurate number? I’ll leave that as an exercise for the reader. (P.S. Love you, clients.)

There are many tools you can use to run an automated audit, each with their own particular parameters. You’ll want to find one that fits your project needs, but here are a few starting points:

  • Content Analysis Tool (CAT): perfect if you’re new to automated audits, or your audit needs are fairly straightforward.
  • Flock: a fancy dashboard, better suited for in-house teams that need to monitor website data over time (rather than for a one-time inventory).
  • Screaming Frog + URL Profiler: two advanced tools that do little on their own (as far as my own needs go), but the combination builds a spreadsheet of beauty.

(The gymnastics of wrangling automated spreadsheet data could command another 1,500 words here—perhaps a future post? Stay tuned.)

At first blush, automated audits seem to have everything going for them. Heck yeah, let the robots do the work! Who wants to go through thousands of pages? Who has that kind of time? And automated crawls are not just faster, but more efficient and accurate, too—after all, audits conducted by humans suffer from the biases and errors of those humans.

Automated audits are pure, unadulterated, infallible data—but that data only tells half the story. Annoyingly, we require human skills of observation and inference to provide context. You’ll need to allow for some level of manual auditing to supplement the automation, to make meaning from the cold, hard science.

If the prospect of a manual audit is daunting, remember that it doesn’t mean reading every single word on every single page. Go back to question two—the purpose of your audit might be focused on imagery, or presentation, or structure. You might be more interested in examining file paths than readability. And, if you are auditing the writing itself, you’ll want to audit representative sections and pages, focusing on specific criteria rather than an encyclopedic reading.

Combine as needed

Qualitative and quantitative, automated and manual, chocolate and peanut butter: it’s a mixed-up world and I’m fine with that.

My best audit results come from carefully combined methodologies. On a typical project:

  • I run an automated audit, then slice and dice the data.
  • I manually review the site at a high level, comparing what I see there to the implications of the crawl data.
  • Finally, I conduct a detailed manual audit focused entirely on the site structure.

That’s three audits! But together, they form the basis for much of my later content and IA strategy work.

And there’s so much more to this conversation—we haven’t even gotten into actually conducting the audit itself! Let me know in the comments what you want to hear more about, and what’s been helpful to you in your site audits! What tools do you use? What questions do you ask yourself? What can we share with each other to improve site audits for everyone?