How to Audit Big Websites

In this post, we’re going to talk about conducting a qualitative website content audit for large websites—the process of analyzing the quality and effectiveness of your site’s content.

Preparing for a content audit

Before you start any audit—but especially a big site audit—you need to take some time to think through a few things:

1. Define your goals. Think about what you want to accomplish when auditing a big site. You can’t get a complete picture of your million-page site, but you can evaluate the quality of your content overall or within specific sections. You can identify areas for improvement. Or prioritize content projects. Or find content successes that you can build on. It all depends on your business goals.
2. Confirm team size. For a big audit, you’ll probably want help (unless you’re looking for a one-way ticket to Insanityville). Identify who’s available to spend some good chunks of time auditing and find out how much time they have. For maximum efficiency, you’ll want to have people do substantial bursts of auditing—not reviewing a page here and there between other tasks. Ideally, aim to have one audit lead who’ll run things and enough auditors to tackle your audit sample in a reasonable amount of time. We usually estimate that an auditor can handle 5 to 7 pages an hour. So, for example, if you want to review 5,000 pages, you’ll need six people working full time to get the audit done in a month. Don’t have that many resources? Have more time? Adjust the amount of content or team size accordingly.
3. Think strategically about sampling. Since you can’t look at every page of a big site, you’ll need to choose a sample—a small segment of your content to review. Do you want to look at a specific subset of your content in detail? Or grab a representative sample across the site and extrapolate your findings? Either way, it’s generally best to look at a good cross-section of site “levels”—dig deep, don’t just look at the polished, high-profile top layers.
4. Choose your criteria. You’ll also need to decide what aspects of content quality you want to measure. Often we’ll do a mix of web best practices and qualities that are specific to an organization’s unique goals. Is the content readable? Usable? Does it express your brand as you’ve defined it? Is the content appropriate for your primary audience? Distill these ideas into a set of four or five concrete audit factors. Don’t go overboard: eight audit factors times 5,000 pages equals an overwhelmed audit team.

Get set …

Once you’ve answered the big “how, what, and why” questions, it’s time to prepare for the audit logistically.

1. Create a “criteria sheet.” With many auditors on the same project, you need to create a clear, consistent set of criteria they can all share. Set up a criteria reference sheet that lists your audit factors, the rating scale, and some concrete things to look for to arrive at ratings. Have your team keep a copy in sight at all times. It keeps people aligned on standards and helps stave off “audit drift,” where people’s interpretation of audit factors starts to blur over time. (No, it’s not just you—that happens to everybody.)
2. Divide up the work. At some point, you’ll need to decide exactly which pages or pieces of content each member of your team will review. You can do that up front by creating a content inventory that details every page to review, and distributing sections of the inventory to the team. Or you may want to just assign high-level sections and then have individual auditors select pages on the fly within those sections. The upfront method can be speedier, but you may get better results when auditors who are immersed in the content decide which pages are most useful to evaluate.
3. Split the spreadsheets. For most audits, you need to create a spreadsheet to record your results. For a big audit, it’ll be especially important to keep things organized. When you have several people working in one file, it’s all too easy to overwrite someone’s work, forget who’s doing what, or wind up with multiple document versions that are hard to consolidate. To avoid confusion, break off individual worksheets for your auditors.

Here’s an example of a content audit spreadsheet:

Audit!

And then the fun begins.

1. Set up a kickoff. Sit down with your team. Outline the expectations—the timeline, how many pages each person will handle, who’s doing what, and the goals. Share any useful background information. Hand out the criteria sheet and go over it in detail. Ask them to voice any questions. There will be questions, and you want to iron out the wrinkles as early as possible—not when you’re three weeks in.
2. Workshop with the team. This is critical for getting everyone aligned. Getting half a dozen people to be consistent with each other—and with themselves, over the course of a month or six weeks—takes ongoing effort. Start off with an early workshop or two where you have people bring in their ratings and compare notes. Or put a page up on the projector and let them have at it en masse. Encourage your auditors to work side-by-side for a while and talk about what they’re finding.
3. Pull it all together. Collect the worksheets and combine them into one master spreadsheet so you can crunch the numbers and discover trends. Ask your team for notes or summaries and meet with them to discuss the patterns they saw in the content.
4. Get the word out. To share your results with others in your organization, put together a report or presentation that covers the most important things you discovered. Identify the biggest content problems—and think about ways they can be addressed. (We usually try to prioritize them to some degree based on how severe a problem is and how easy or difficult it is to address it.) Show examples. And don’t forget to note the content’s strengths as well, so that you can build on them in your future content efforts.

Making improvements to a huge body of content can be a daunting process. Even assessing the current state of the content is no small task. But with good planning and a solid process, you can break it down into manageable pieces and get the solid data you need to get the ball rolling.