012 min
Why a content inventory and audit can overrule page views
A content inventory and audit asks a question that page views cannot answer, and a UK government team showed why in 2019. In April of that year, the Race Disparity Unit deleted one of the most visited pages on its Ethnicity facts and figures website. The page was called "Ethnicity in the UK". It had been in the site's top five by page views since the site launched in October 2017. In early 2019 the page earned about 6,000 unique views every month, more than 5% of the whole site's views.
By the usual measure, the page was a success. Then a content designer reviewed it page by page and found three problems. The page tried to meet about half a dozen different user needs. Some of its content was repeated elsewhere on the site. And some of it did not meet the rules for what belongs on GOV.UK. The team did not want to delete a page that popular on one person's opinion, so it opened the analytics.
What the page views missed
Page views count arrivals. They do not show whether people found what they came for. The analytics showed that 50% of visitors opened the first of the five accordions and only 14% opened the fourth. An accordion is a block that hides a section of text until the reader opens it, so many visitors never reached the lower sections.
Any team that ranks content by traffic alone can make the same mistake. A page with many visitors may be a page that many of them leave without finding an answer.
What the search terms showed
Search data supported the same conclusion. Google Search Console is Google's free report on how people reach a site through search. It showed that around half of all search clicks to the website came from people looking for one of two things: the ethnic breakdown of the UK, or a list of ethnic groups. The team concluded that the page was good at attracting searches but made visitors search through it for what they had come to read.
So the team deleted the page and sent its visitors to the UK population page. It rewrote that page's title and summary so it would still match searches with words like ethnicity, population and uk. It also rewrote the page on ethnic groups. Two pages now did the work of three.
What the team did has two halves, and they are different jobs.
022 min
What a content inventory records
A content inventory is the first half, and it answers a plain question: what is there? Nielsen Norman Group (NN/g), a usability research firm, describes it as a full list of the digital content a team holds, at page or file level, with facts about each piece. The inventory counts and records. It does not judge. Teams usually keep it in a spreadsheet, because rows and columns suit a list of facts and everyone already knows the tool.
| Column | What goes in it | What it shows later |
|---|---|---|
| Name | The real title of the piece, not only the page title | Lets a reader find the item in a discussion |
| Address (URL) | Where the piece lives | Shows how the site is organised |
| Owner or source | Who wrote it, who maintains it, or where it is fed from | Gives each decision a person to act on it |
| Subject and format | The topic, and a type such as article, video or PDF | Shows where one topic has too many pieces |
| Dates | When it was created and last changed | Points at content that may be out of date |
| Metadata | Page title, meta description, alt text | Shows what search engines and screen readers get |
A crawler or a content management system can export most of these columns, so the inventory is the part of the job that tools do well. Anyone can check a column like this without an opinion. The judgment comes next.
A common mix-up: the inventory is not the qualitative half
Some guides call the inventory "qualitative" because people do the work. The usual split runs the other way. The inventory is a quantitative log of facts. Quantitative means it deals in counts and recorded facts. The audit is the qualitative half, which means it deals in judgments about quality, because someone has to read each item and decide how good it is.
The mix-up has a cost. A team that treats listing as judging can mark a page fine just because the page exists.
032 min
What a content audit adds to the inventory
A content audit is the second half, and it answers the question a list cannot: is any of it good? An inventory row for the Ethnicity page would have shown a title, an address and a date. Nothing in those columns shows that the page tried to serve half a dozen needs.
A content audit adds a judgment for each item, using criteria the team agrees before it starts. NN/g says the criteria usually come in two groups.
- Writing and format practice. Plain language, clear headings and link text, short chunks with white space, enough contrast between text and background, no text embedded in an image, and no video that plays by itself.
- The organisation's own standards. Who the audience is and what task they are doing, the rules for tone and metadata, and what the content is supposed to achieve, measured with numbers such as clicks, views and bounce rate.
Each item gets a quality rating of high, mediocre or low, and then a decision: keep, update or remove. An item that needs an update gets a note on what to fix and a named owner. NN/g adds that tools can speed up the inventory, but people should do the auditing, because the judgment is the work.
ROT is only the start
Many teams begin the audit with ROT, a filter for content that is redundant, outdated or trivial. In a 2009 UXmatters column, Colleen Jones pointed out that a ROT pass finds stale and unimportant content but misses content that is mediocre, inappropriate, inconsistent or off brand. She added more heuristics, which are rules of thumb for judging, to catch those. Her list drew on Fred Leise's checks, which include completeness, consistency and information scent, meaning how well a link's label predicts what the reader will find behind it.
042 min
How to run a content inventory and audit
A content inventory and audit takes a few decisions to run, and the order matters.
Choose a scope.
It can be a whole site or app, one section, or one user journey across several touchpoints, as a user journey map would show. A full audit covers everything, a partial audit covers one section, and sampling checks a few pieces. If there is more content than the team can handle, start with the pages people open most or the ones that support top tasks.
Name an owner and tell people early.
One person or a small group owns both files. Authors, designers, stakeholders and researchers should hear what is happening and agree on the criteria before the first row is filled in.
Pick one familiar tool.
A spreadsheet works for most teams. Use a crawler or a content management export for the inventory columns.
Set a time limit.
NN/g suggests doing what you can in a fixed period, such as 6 weeks, so the team learns where its content succeeds and fails while interest is still high.
Fill in rows and judge.
Where the team cannot tell whether a piece meets user needs, it writes that in the sheet and asks the author.
Review decisions with the people involved.
Removals deserve the most discussion.
Start from the question, not the data
The GOV.UK team learned to decide what it wanted to find out before it opened analytics. A plain list of pages with their views can be enough for an audit. Finding gaps may need a different report, such as what visitors typed into the site's own search box. Choose the data after the problem is clear.
What analytics cannot tell you
Analytics said that something was wrong with the Ethnicity page. They did not say what to do. In a reply under his post, Spencer Ball, the unit's content designer, explained that the team ran no user research, because it judged the analytics compelling enough. The team kept watching the numbers afterwards. When the fix is unclear, a user researcher can test candidate fixes with real people before a page is deleted.
052 min
What one retailer removed after a content audit
A content audit often ends in deletions, so the clearest test of an audit is what happened after one. Home Science Tools is an online store that sells educational science products. Its blog, the Learning Center, was hosted on its own subdomain. In early August 2018, an agency finished a content audit of that blog and handed over a list of pages to prune.
The retailer's team removed around 200 blog pages, about a tenth of the total, beginning with the weakest. These were pages with little or no organic traffic, page views, conversions or backlinks. The agency reported that over the following months organic sessions rose 104%, transactions rose 102% and revenue from strategic content rose 64%.
Treat those numbers with caution. They come from the agency that sold the audit, from one site compared with its own past, and the agency calls the pruning one part of a larger growth effort. The safer reading is modest: removing pages with almost no traffic did not stop the site growing.
Low performing is not the same as low value
A traffic rule would have deleted pages the blog needed. Tag pages are the example. They hold little content of their own but still help readers move around the blog. The agency's advice for pages like these was to keep them for readers and add a "noindex" tag, which tells search engines not to list them. It listed FAQ pages, About pages and author pages as pages to keep indexed even when their numbers were weak.
So the question before a deletion is not only how often a page is used. It is also what readers use it for.
062 min
How a content audit shapes a rebuild
A content audit does not only react to numbers on a live site. It can also guide a rebuild, as two projects show.
Merging several sites
Jones used an audit on a project that migrated and merged several different websites. She chose a mix of heuristics from the lists she knew, analysed part of one site, and found that a couple of her picks did not apply, so she dropped them. The lesson is that criteria belong to the project. A team that copies a template of criteria will spend its time scoring things that do not matter there.
Overlapping guides
In 2016, a content designer from the Department for Business, Energy and Industrial Strategy joined the Government Digital Service for a few days to review the guidance for people starting a business. Feedback and data showed that the content was over-complex and fragmented, with several overlapping guides serving the same need and no single place to start. Callers to the Business Support Helpline who had visited GOV.UK said most often that they had not found what they needed.
The overlap was visible only by comparing what each guide was for, not what it was called. In a few days the team drafted two pieces of content, improved the main browse page and the search results, and listed further actions. An audit that sorts pages by the need they serve finds duplicates that a title search misses.
071 min
What to do with a page an audit removes
A content audit's remove decision is not finished when the page is deleted, because visitors, links and search results still point at it. A careful removal has three parts, and the Race Disparity Unit did two of them.
- Check the old page for anything worth keeping and move it into a page that stays. Texas A&M's web guidance gives this step.
- Redirect the old address to the page that replaces it, so a visitor with an old link still lands somewhere useful. The unit sent visitors to its UK population page.
- Rewrite the destination's title and summary so it keeps matching the searches the old page attracted. The unit did this for the same page.
A deletion that skips these steps can lose facts and search visitors together.
081 min
Content inventory and audit vs. nearby concepts
Several research and design methods look at the same product from another side. The axis that separates them is what each one judges and who does the judging.
| Method | What it examines | Who judges | Output |
|---|---|---|---|
| Not in the library yetContent inventory and audit | |||
| ResearchHeuristic evaluation | |||
| DesignInformation architecture | |||
| ResearchTree testing |
The inventory and audit supply input to the others. The audit decides which content goes into the new structure, and tree testing checks whether people can find it there. A heuristic evaluation looks at how the interface behaves, while the audit looks at each piece of content one at a time.
091 min
Keeping a content inventory and audit current
A content inventory and audit start to go out of date on the day a team begins to use them. NN/g notes that new content appears and existing pieces change as soon as decisions start, so the inventory needs updating almost at once. A content management system can keep the inventory columns current. The audit columns need a person.
That is why ownership matters. NN/g warns that people add the job to their other duties and it never works, and recommends sharing the load with the same group that was involved from the start. Some teams run a rolling audit, a recheck at an agreed interval such as weekly, monthly or quarterly, instead of one big project every few years.
The decision to make before starting is who owns the audit after the first pass is done.
?5 questions
Questions people ask
Do I need both a content inventory and a content audit?
How often should a team run a content audit?
What does ROT mean in a content audit?
Should I delete every page with low traffic?
What tools do I need for a content inventory?
§9 sources
Sources for content inventory and audit
Kaley, A. (2020, September 27). Content inventory and auditing 101. Nielsen Norman Group.
Jones, C. (2009, August 3). Content analysis: a practical approach. UXmatters.
Ball, S. (2019, June 26). Guest post: why we deleted one of our most popular pages. Inside GOV.UK.
Street, J. (2016, September 5). Improving guidance for new businesses. BEIS Digital.
Show all 9 sourcesShow fewer sources
Inflow. Content pruning case study: deleting pages to boost SEO. A case study by the agency that audited Home Science Tools, so the results are self-reported.
Texas A&M University. Content inventory and content audit. Web guidance from the Aggie UX University.
Wikipedia. Content inventory.
Wikipedia. Content audit.
Land, P. L. Content audits and inventories: a handbook for content analysis, 2nd edition. XML Press. A publisher page describing the book.
