Why Your Staff Can't Find Anything on Your Intranet (And What It's Actually Costing You)

I had a conversation with a government IT manager a few weeks ago that I haven't been able to stop thinking about.

She told me her team had done a small experiment. They asked ten people across different departments to find a specific policy document on their SharePoint intranet. A document they all knew existed. The sort of thing you'd expect to take thirty seconds.

Average time to find it: eleven minutes.

Three of the ten gave up entirely and just emailed a colleague to ask for the link. (I've started calling this the "Ask Sarah" method. Every organisation has a Sarah. You probably already know who yours is.)

She wasn't surprised by the result. That's what got me. She wasn't even a little bit surprised. She just said, "Yeah, that's about right. Nobody trusts search."

And here's the thing. This isn't a small organisation. It's a government agency with over 2,000 staff. Every single one of them hits this wall multiple times a day. That's not a minor inconvenience. That's a structural productivity problem hiding in plain sight, and most organisations have just decided to live with it like a dodgy knee.

The Problem Everyone Feels But Nobody Measures

We've been building SharePoint intranets at Webcoda for over twenty years now. And in almost every initial conversation with a new client, the same complaint comes up within the first ten minutes: "Our staff can't find anything."

I'm not exaggerating when I say it's the first ten minutes. I've genuinely considered starting meetings with "So, how bad is search?" just to skip the small talk.

It's so universal it's almost become background noise. Everyone knows intranet search is bad. Everyone has workarounds. Nobody actually sits down and works out what it's costing.

But the research is pretty clear on this.

McKinsey Global Institute found that knowledge workers spend roughly 19% of their workweek searching for and gathering information (McKinsey, "The Social Economy," 2012). That's nearly eight hours a week. Per person. A full working day, every week, spent wandering around a digital building looking for the light switch.

IDC's research puts it even higher, with knowledge workers spending up to 2.5 hours a day searching for information, and failing to find what they need about half the time (IDC, "The Knowledge Quotient," 2014).

Those studies are over a decade old. But if anything, the problem has gotten worse. More tools, more platforms, more SharePoint sites, more Teams channels, more everything. The volume of content inside most organisations has exploded, and the ability to find the right thing at the right time hasn't kept up. Microsoft's own Work Trend Index has consistently documented this (Microsoft Worklab, 2024).

This is the dark matter of productivity loss. It doesn't show up in project reports. Nobody puts "couldn't find the procurement template so spent 20 minutes looking" in their timesheet. But it's happening, every day, across every department, in every organisation with a poorly structured intranet.

Why SharePoint Search Is Terrible (In Most Organisations)

I want to be clear about something: this isn't a SharePoint problem. SharePoint's search technology is genuinely capable. Microsoft has poured billions into the Microsoft Graph, the Semantic Index, and the search infrastructure behind M365. The engine is fine.

The problem is what organisations have done (or haven't done) with the content it's trying to search.

Here's what we typically find when we look under the hood.

No metadata strategy. This is the big one. Most SharePoint environments are basically giant filing cabinets with no labels on the drawers. Documents get uploaded with whatever filename someone chose that day. No content types. No managed metadata. No taxonomy. Search is trying to find a needle in a haystack where someone hasn't even confirmed the needles exist.

Flat or chaotic folder structures. Somewhere around 2015, the "we'll just put everything in folders" approach seemed reasonable. By 2026, you've got department sites with fifteen levels of nested folders, naming conventions that made sense to one person who left three years ago, and dozens of folders called "Archive" or "Old" or "DO NOT DELETE."

I'll be honest. We built a few of those early SharePoint sites ourselves back in the day. The thinking was "just give people a place to store documents and they'll organise them naturally." They didn't. Nobody does. We learned that the hard way.

Then there's the content sprawl. Over the years, teams create sites, subsites, team sites, communication sites, project sites. Some of those projects ended five years ago. The sites are still there. The documents are still there. Search doesn't know they're outdated. So when someone searches for "travel policy," they get four results from four different sites, three from 2019 and one that's current. Working out which is which feels like a game show nobody signed up for.

Broken permissions genuinely trip people up. SharePoint search respects permissions (which is correct and important). But in most organisations, permissions are a tangled mess of inherited and broken-inheritance settings applied over years of ad hoc changes. Your manager finds the document on the first try. You search the same words and get nothing. That's not a bug. That's a permissions architecture nobody has audited in years. Pretty scary when you think about how many decisions are being made based on whatever version of reality each person can see.

And then there's the taxonomy question. SharePoint has a genuinely good term store and managed metadata service. Maybe 10% of the organisations we work with actually use it. The rest rely on whatever folder names and document titles people came up with. A document about "annual leave entitlements" and one about "holiday leave policy" and one about "personal leave provisions" might all cover the same topic, but search treats them as completely unrelated. I don't get it. The tool is right there. It's been right there for years.

But the most expensive problem? Staff have learned not to trust search. After enough bad experiences, people just stop trying. They email Sarah. They ping someone on Teams. They save documents to their desktop "just in case." They bookmark specific pages. They build their own personal workaround systems because the official one has let them down too many times.

That last point is completely invisible to the IT team, which is exactly why it costs the most.

Let's Talk About What This Actually Costs

Right, I know what you're thinking. "Here he goes with the back-of-the-napkin maths." Fair. But let's do it anyway, because these numbers are pretty confronting even with conservative assumptions.

Take a mid-sized government agency or enterprise with 500 knowledge workers.

If we're conservative and say each person wastes just 30 minutes a day on bad search and information retrieval (well below the McKinsey and IDC figures), that's:

• 250 hours per day across the organisation
• 1,250 hours per week
• About 62,500 hours per year

At an average fully loaded cost of $65 per hour (which is conservative for government and enterprise roles in Australia), that's roughly $4 million per year in lost productivity.

Four million. On people wandering around their own intranet looking for things that are already there. That's pretty insane.

And that's the conservative estimate. If you use the full McKinsey figure of 19%, you're looking at something closer to $12 million. At which point it's not a "search problem" anymore. It's one of the biggest invisible line items in your entire operating budget.

Now, not all of that is solvable. People will always need to search for things. But if you could cut that wasted time by even 30% through better information architecture, metadata, and governance? That's over a million dollars a year in recovered productivity for a 500-person organisation.

Nobody writes that number on a whiteboard in IT steering committee meetings. But maybe they should.

What "Good" Actually Looks Like

After twenty years of building intranets, I can tell you that the difference between a broken search experience and a good one isn't magic. It's just structure. It's not glamorous. It won't make the CEO's innovation presentation. But it works.

A well-set-up SharePoint environment has a few things in common.

Managed metadata and a term store that people actually use. You've defined the terms your organisation uses (policy types, departments, document categories) and applied them consistently. When someone uploads a document, they tag it. When someone searches, those tags surface the right result. Boring work. Dramatic results.

Content types for different kinds of documents. A policy, a procedure, a form, a meeting agenda. These are different things and should be treated differently in search. Content types let you define what metadata each document type needs. It's the difference between a library with a catalogue and a library where they've thrown all the books into a room. (My kids' bedroom, basically.)

Proper navigation that reflects how people actually think. Not how the org chart looks. Not how the IT team thinks about it. How an actual human being who needs to find the travel claim form thinks about it. That's a real conversation, not a quick admin decision.

Site governance with actual teeth. Someone owns each site. There's a review cycle. Stale content gets archived or deleted. New sites get created through a process, not just whenever someone decides they need one.

Retention policies that clean up after themselves. Documents don't live forever (even though SharePoint will happily let them). Retention policies handle the lifecycle of content so your search results aren't polluted with documents from 2017 that nobody needs.

None of this is technically difficult. SharePoint has all of these features built in. The hard part is the governance, the planning, and the organisational commitment to actually use them. It's a people and process problem. Always has been.

What About Copilot? Doesn't AI Fix This?

I hear this question a lot, especially from organisations that are excited about Microsoft 365 Copilot (and fair enough, it's genuinely impressive technology).

Here's the honest truth about Copilot and intranet search: AI makes good content better and bad content worse.

Copilot uses Microsoft's Semantic Index to search your content. It understands meaning, not just keywords. So instead of searching for the exact phrase "travel expense policy," you can ask "how do I claim a flight for a work trip?" and Copilot will find the right document even if it doesn't contain those exact words. That's a real improvement, and it works brilliantly when your content is well-structured and up to date.

But here's the catch. If you've got four versions of the travel policy across three different sites, two of them from 2020, Copilot will still find them. It might even surface the outdated one because the old document has better metadata (or simply more text that matches the query). And it'll present that answer with full confidence, complete with a citation, looking for all the world like it's the correct and current answer.

We saw exactly this happen with a client earlier this year. We wrote about it in our article on how the Semantic Index actually works. Copilot found an outdated parental leave policy on a forgotten subsite and nearly sent the wrong entitlement information to a new employee. That's not a hypothetical.

AI on top of a mess just gives you faster, more confident access to the mess. It's like putting a turbo on a car with no steering wheel. You'll get somewhere quickly. Just not where you wanted to go.

So if you're thinking about deploying Copilot (or you already have), the foundation matters more than ever. Your metadata strategy, your content governance, your information architecture. These aren't "nice to have" cleanup tasks anymore. They're prerequisites for getting value from your AI investment.

The organisations getting the best results from Copilot aren't the ones with the biggest budgets. They're the ones that did the boring IA and governance work first. Same as F1, really. Red Bull didn't win four constructors' titles by having the fastest car on day one. They won because their engineering foundation let them extract more performance from every upgrade. Same principle. Minus the podium champagne.

Related Article10 min read

Your Messy SharePoint Intranet Is Now an AI Problem

Microsoft Copilot is now searching your SharePoint content to answer questions. If that content is stale, duplicated, or buried on a forgotten...

Read full article

Related Article11 min read

You're Probably Paying for SharePoint AI You've Never Switched On

Most Australian organisations with M365 licences have AI capabilities sitting dormant in SharePoint. Here's what's available, what it costs, and how...

Read full article

The Workarounds That Are Costing You More Than You Think

Before we get to solutions, it's worth calling out the workarounds that have quietly become standard practice in most organisations. Because each of them has a hidden cost, and honestly, some of them are painfully funny when you see them from the outside.

The "just email someone" approach. Can't find a document? Email someone who might have it. That person spends their time hunting it down, attaching it, and sending it back. Two people's time consumed for one document retrieval. And the requester now has a copy on their local machine that will never be updated when the source changes. Congratulations, you've built a beautiful system for distributing outdated information at scale.

The desktop hoarder. You know this person. You might be this person. (I won't judge.) They save everything locally. Their desktop has 400 files in a grid that makes sense only to them, with names like "Budget FINAL v3 USE THIS ONE.xlsx". Their desktop is their intranet. Those local copies diverge from the source almost immediately, version control goes out the window, and when that person eventually leaves, their lovingly curated collection of "the good versions" disappears with them.

The tribal knowledge network. "Oh, you need the procurement template? Ask Sarah. She knows where everything is." Every organisation has a Sarah. She's efficient, she's helpful, and she's become a human search engine without anyone explicitly giving her the job. She spends a genuine chunk of her day answering questions that a functioning intranet would handle in seconds. Her actual job? She barely gets to it anymore. And when Sarah's on leave, the entire organisation grinds to a halt on anything involving document retrieval. That's a pretty scary single point of failure for something as basic as "where's the form."

The bookmarks folder. Some people build elaborate browser bookmark collections to navigate around broken search. Three levels of folders, colour-coded, maintained with the dedication of a stamp collector. These break every time a site gets restructured. And they can't be shared easily with new starters, so every new employee builds their own map of the wilderness from scratch.

Every one of these workarounds feels sensible to the individual. And every one of them represents a failure of the system that costs the organisation time, accuracy, and institutional knowledge.

Where to Start (Without Boiling the Ocean)

I'm not going to pretend that fixing an enterprise SharePoint environment is a weekend project. It's not. But it doesn't have to be a two-year programme of work either.

The organisations that make the most progress start with a focused audit. Not a full content migration. Not a redesign. Just a clear-eyed look at where the biggest problems are.

At Webcoda, we typically start with what we call a SharePoint health check. Takes about half a day. We look at your site structure, your search configuration, your metadata (or lack of it), your content types, and your permissions architecture. We identify the top five or six things causing the most pain and put together a practical plan to address them.

Some of the quick wins we see over and over again:

• Setting up a basic managed metadata taxonomy for your most important content (policies, procedures, forms)
• Consolidating duplicate sites that are confusing search results
• Configuring search result sources so people get results from the right places first
• Archiving or removing obviously stale content that's polluting search
• Setting up a simple site governance model so the problems don't come back

These aren't massive projects. Most of them can be done in weeks, not months. And the impact on day-to-day usability is immediate.

The bigger structural work (content types, migration, full IA redesign) can come later, once you've proven the value of the quick wins and built organisational support for the effort.

While you're thinking about your digital presence, here's a quick way to check how your public website stacks up for AI discoverability:

The Conversation Nobody's Having

Here's what I think is really going on. Most organisations don't fix their intranet search because they've normalised the pain. "Our intranet is terrible" has become a punchline, not a problem to solve. It's just accepted as the way things are, like bad coffee in the kitchen or the printer on level 3 that jams every Tuesday.

But when you actually put the numbers on it, when you work out that your 500 staff are collectively burning 60,000+ hours a year just trying to find things that already exist in your own systems, the joke stops being funny.

And with Copilot and AI entering the picture, the cost of inaction just went up. You're not just losing productivity to bad search. You're undermining the single biggest AI investment Microsoft has ever made by feeding it unstructured, ungoverned content. That's an AI readiness problem now, and it's going to separate the organisations that get real value from their M365 investment from the ones that keep wondering why Copilot "doesn't work."

If you've been putting off the intranet cleanup, this is probably the year to stop kicking the can down the road. Not because I'm trying to sell you something (though we do this work, and we're pretty good at it after twenty years). But because the gap between well-governed SharePoint environments and ungoverned ones is about to get a lot wider.

The ones who've done the work will get genuine value from Copilot and semantic search. The ones who haven't will keep emailing Sarah.