The hardest question I was asked while responsible for more than sixty country websites had nothing to do with rankings. It was a version of a question every large organisation eventually asks itself. What do we actually say about this service, across everything we have published, in every market?

A crawl could tell me how many pages mentioned it by lunchtime. Nobody could tell me what those pages added up to, whether they agreed with each other, or whether the version a prospect in Germany read matched the version a prospect in Brazil read. The content existed. Understanding it was the problem.

That gap gets wider as organisations grow. Individual teams generally know what they publish. A product marketing team knows its product pages. A country team knows its local site. An agency knows the campaign it was briefed on. What almost nobody has is a reliable picture of the whole estate, because the estate was never built as a whole. It accumulated, through launches, acquisitions, rebrands, regional initiatives and the ordinary turnover of people who knew where things were. Add business units, languages and external partners, and the problems multiply. Content gets duplicated. Messaging drifts. Statements go out of date. Gaps stay invisible because nobody is looking at the level where a gap would show.

None of this looks serious from inside a single page or a single market, which is why it persists. It becomes visible only when someone needs to understand the content estate as one thing, and by then the estate usually runs to thousands of pages that no individual has read.

Knowing what you already have

The Content Marketing Institute's enterprise research for 2026, drawn from 296 marketers in organisations of a thousand or more people, put resource constraints at the top of the list of challenges at 42%, followed by measuring content effectiveness at 38% and cross-departmental collaboration at 33%. Only 22% said producing enough content was a problem. That ordering matches my experience exactly. Large organisations are good at publishing. What they struggle with is knowing what the publishing has produced.

The first symptom is that coverage becomes impossible to state with confidence. A global content manager can usually tell you that a service is mentioned on hundreds of pages. Whether the organisation explains that service properly anywhere, with the detail a buyer needs before they contact sales, is a different question, and the honest answer is often that nobody knows. Mentions are easy to count. Explanation is not. The second symptom is duplication that nobody intended. Three business units write their own overview of the same capability, each slightly different, each ranking against the others. A regional team rebuilds a page that already existed at group level because they could not find it, or found it and did not trust it. The third is fragmentation, where the information a customer needs exists but is spread across a product page, a support article, a PDF from 2019 and a blog post that has since been orphaned from the navigation.

There is a distinction here that matters more than it first appears. A gap in the existing website and an opportunity supported by customer demand are two different findings, and the audit tools that surface the first say nothing about the second. An organisation can have an obvious hole in its content and still be right to leave it, because nobody is asking. It can also have a subject covered on forty pages and still be failing the people who search for it. The question is never simply what is missing. It is what is missing that someone actually wants.

Localisation drifts into contradiction

International estates add a dimension that single-market content teams rarely have to think about. Every market has its own marketing team, often its own agency, and its own subject matter experts who know their customers better than head office does. Over time, each of them develops a slightly different account of what the business sells and why. Some of that difference is valuable. A service that is sold on compliance in one country may be sold on cost in another, because the regulation differs and so does the buyer. Local content should reflect that.

The trouble is that the difference does not stay within the useful range. A capability that group describes as available everywhere gets described as regional by a market that has not caught up. A pricing model changes centrally and three sites keep the old one for a year. A product is renamed and the old name survives in a dozen translations because nobody was tracking where it appeared. None of this is anyone's fault in a way that would show up in a performance review. It is the ordinary result of many competent people maintaining their own corner of a site that has no single owner.

I learned to distinguish two things that look alike from a distance. Localisation changes emphasis, examples, proof points and tone to suit the market. Contradiction changes the facts. The first is a sign that the market teams are doing their jobs. The second misrepresents what the business provides, and the customer who notices it does not care which team was responsible. Finding the second across 35 languages, when the people who can read each language are the same people who wrote the content, is a problem that no amount of editorial process on its own will solve.

Unsupported claims and false claims are different problems

Every enterprise website makes claims. Years of experience, number of customers, breadth of capability, speed of service, the benefits a client can expect. Most of these were true when they were written and many were supported by something at the time. The difficulty at scale is that the support and the claim live in different places and age at different rates. The case study that backed a benefit gets retired. The certification that justified a capability lapses or is renewed under a different name. The statistic in a headline came from a report that was accurate in 2021. The claim stays on the page because the page still performs and nobody has a reason to open it.

It is worth being careful about what this is and is not. A claim with no visible evidence behind it is not the same as a claim that is false, and treating the two alike creates the wrong kind of panic. Most unsupported claims on an enterprise site are probably still true. The problem is that the organisation can no longer demonstrate it, and in regulated sectors, or in front of a procurement team that reads carefully, that distinction matters. The useful question for a content leader is not whether the site is lying. It is how much of what the site says could be defended today if someone asked.

The same review surfaces the quieter problems. Explanations that assume knowledge the reader does not have. Pages that describe a service without ever saying who it is for, what it costs or how to start. Statements about the business that were true before the last acquisition. The Nielsen Norman Group's guidance on content audits is clear that the quality assessment should be done by people looking at each piece of content, and that is right for a site of a few hundred pages. Across a few thousand, in dozens of languages, that human review happens once, if at all, and is out of date before it is finished. The underlying challenge is maintaining confidence in what the organisation publishes, and confidence is hard to maintain in something you cannot read.

Finding problems is cheap and deciding what to fix is not

Any competent audit of a large site will produce hundreds of issues. Duplicates, thin pages, outdated statements, missing information, contradictions between markets, claims without support. Producing that list is the easy part, and I have watched teams mistake the length of the list for progress. The hard part is deciding which of those issues deserve money and attention, because the content team will not get the budget to fix all of them and should not ask for it.

A content problem is not automatically a commercial priority because it has been identified. An outdated page that nobody visits, on a service the business is winding down, can stay outdated. A contradiction between two markets on a product that drives a third of enquiries cannot. The judgement has to weigh the weakness of the content against the commercial importance of the subject, the evidence of customer demand, the resources available to act and the regional politics of who owns the page. Those inputs rarely sit in the same place, and the person making the decision usually has good data on one or two of them and instinct on the rest.

What content teams need, and mostly lack, is enough evidence to make a defensible decision about what should be improved, consolidated, updated, left alone or investigated further. Defensible is the operative word. A content director who can show why a hundred issues were deprioritised, with reasons that a commercial stakeholder would accept, is in a much stronger position than one who fixed everything they could reach and could not say why.

The tools answer different questions

The obvious objection is that enterprises are not short of tools. They have a content management system, often several. They have analytics, SEO platforms, a digital asset manager and a reporting layer that someone spent a year building. All of these are useful and I would not run a large site without them. They just answer different questions from the one I opened with.

A CMS manages content. It knows what exists, who edited it and when, and it can tell you that a page was published. It has no view on whether the page agrees with another page. Analytics explains performance. It can show that a page is losing traffic or converting badly, but it cannot say whether that is because the content is wrong, duplicated elsewhere or simply out of date. SEO tools find technical faults and search opportunities, and the better ones now include content quality features and some semantic analysis. The content audit platforms have moved on too, and some of them do real work on duplication and readability. I want to be fair to them. The gap is not that nobody has built anything. It is that establishing what thousands of pages actually say, how those statements relate to each other across markets and languages, and where the evidence behind them is missing, is a different kind of analysis from the ones these tools were designed for, and achieving it reliably across an entire organisation is still unusual.

The practical consequence is that the question of what the estate collectively communicates falls between systems. Each tool owner can answer their part. Nobody owns the whole, and so the whole goes unanswered until a rebrand, a legal query or a migration forces the issue, at which point the organisation discovers how little it knew about its own site.

Understanding existing content before producing more

The leadership problem underneath all of this is one of visibility. Content teams are asked to make decisions about an estate they can see only in pieces, and most of the pressure on them is to add to it. More articles, more landing pages, more dashboards reporting on the articles and landing pages, more editorial process to govern the production. Some of that is needed. A good deal of it is activity standing in for understanding, and the CMI finding that only 22% of enterprise marketers struggle to produce enough content tells you where the real constraint sits.

Better content management at enterprise scale usually means knowing what you have before deciding what you need. That sounds obvious and is rarely done, because the knowing is expensive and the producing is funded. The organisations that get this right treat their existing content as something to be understood, with the same seriousness they apply to their customer data or their product catalogue, and they give someone the job of understanding it across every market and business unit, with tools adequate to the size of the task.

Having worked with large international websites for a long time, I have become increasingly occupied by this problem. Content teams have more information and more tooling than at any point in my career, and understanding what their organisations already publish remains hard. It is a problem I recognise from the inside, and one I am now working on directly, by developing an enterprise content intelligence tool to help teams understand and evaluate what they already have.

References