If you're here, your intranet is likely live and people are logging in, but announcements still get missed, and you still can't say whether the platform is helping or just sitting there.

Measuring intranet effectiveness turns that suspicion into a specific, fixable answer. It comes down to three questions, asked in order. Are the right people showing up, do they actually do anything once they're there, and can they find and act on what they need?

A single number can answer one of those and still hide the other two. A 70% login rate looks like proof until you realize it says nothing about whether those people found what they came for, or did anything once they were there.

That blind spot is easy to live with when disengagement is already the baseline: with only 20% of employees engaged worldwide, a platform can stay visibly active while the people in it progressively check out. That's why the three questions work in order, and why no single number settles the matter on its own.

In this guide, we help you understand which metrics offer what insights, how to calculate those metrics, and what those numbers actually suggest. We’ll start with the diagnoses, and then help you fix the problems.

First, decide what success looks like for your platform

Before you pull a single number, get specific about what you built the platform to do, because that goal decides which metrics matter.

Name the job in a single sentence, something like "Are people finding the answers they used to email us for?" Let that sentence tell you which layer below to weigh most. You don't need every metric here, just the handful that answer your question, tracked consistently enough to show a trend.

The metrics that tell you what's working (and what to do when they don't)

In this section, we discuss metrics for measuring intranet effectiveness. Pick the area you’re worried about the most and measure that first. We also explain employee sentiment metrics, which can explain the why behind other metrics.

Each metric is a rate or a ratio rather than a raw total, because a raw total only ever climbs and tells you nothing on its own. Treat any benchmark figure here as a reference point for your own trend and not just a universal pass mark.

Adoption: Are the right people actually showing up?

Adoption is where measurement usually starts and where it most often misleads. User and login counts are the first numbers any analytics view shows you, because they're the easiest to pull, so they're also the easiest to over-read. A count of 180 logins means one thing in a company of 200 and something very different in a company of 2,000.

The metrics below fix that by putting every count next to the number that gives it meaning, whether that's your total headcount, how often people return, or how a given team reaches the platform at all.

  • Active user rate (active users ÷ total headcount): This turns a raw login count into a share of your actual workforce. A company of 400 with 180 monthly active users sits at 45%, a very different story from 90%, even though both report "180 active users." Track it monthly. A rate that stays flat or falls while headcount climbs points to one of two gaps: new hires aren't getting in, or existing employees are dropping off faster than newcomers replace them. Onboarding is the usual culprit for the first, and it's often something as small as login details never making it into the welcome email. The fix is to build platform setup into day-one onboarding, so every new hire lands already inside it.

  • Stickiness (daily active users ÷ monthly active users): This separates a daily habit from a once-a-month drop-in. A 10% stickiness ratio means only 1 in 10 of your monthly visitors showed up on an average day, while 30% or higher signals the platform has earned a place in the daily routine. Low stickiness alongside a healthy monthly count usually means people only open the platform when someone emails them a direct link. The first move is to give them a standing reason to open it unprompted. That reason could be opening the intranet for the day's news, their own tasks, and a recognition feed on the landing page.
  • Mobile login share (mobile logins ÷ total logins): For a desk-based team, minimal mobile adoption isn’t a big deal. But for frontline and deskless staff, who make up roughly 70 to 80% of the global workforce and often have no desk and no corporate email, it's frequently the truest reach signal you have. A single company-wide "percent logged in" number hides more than it shows once a large share of your workforce is shift-based. It blends deskless employees, who only ever reach the platform through a phone, into the same average as desk staff who log in from a browser all day, so the blended figure can look healthy while frontline reach is thin, or look low while those teams are actually well covered. Track mobile login share for shift-based teams on its own, and measure reach where they actually work.

Read together, these tell you whether the platform reached your people. When they look healthy, the next question is whether anyone does anything once they arrive.

Engagement: Do people do anything once they're there?

Engagement metrics measure whether your teams actually use the intranet. If you only track logins, you can’t tell if your employees just check in and leave. With engagement metrics, you can measure action instead of just tracking attendance.

Here are some engagement metrics to track:

  • Content engagement rate (interactions ÷ views, where interactions = likes, comments, shares, and confirmed reads): A view proves a post was delivered. Interactions prove someone did something with it. A company announcement with 600 views and 12 interactions runs at a 2% engagement rate, which means most people saw it and moved on. When views climb while this rate stays flat, your posts are being delivered and ignored. The first move is in the posts themselves: ask for a response, target each message to the people it applies to, and stop sending company-wide updates most readers can skip.
  • Comment and reply rate (percent of posts that get at least one reply): This is the clearest sign the platform works as a two-way channel rather than a one-way notice board. Watch the trend rather than the launch week, since opening-week numbers are inflated by curiosity and don't reflect steady-state use. Measure the steady state a month or two in. A low reply rate usually means the content doesn't invite a reply, so move a few posts a week from announcements to genuine questions and watch whether the rate moves.
  • Contribution breadth (unique contributors ÷ active users): Volume hides who’s actually taking part. If the same five names drive most of your posts and comments, a busy-looking feed is really a small group talking among themselves. Break this down by team and location, then go to the teams that have stayed quiet and ask a specific team member on each to post the first update, instead of waiting for them to volunteer.
  • Recognition rate (recognitions given ÷ employees, per month): Voluntary peer recognition is a leading indicator of a connected culture, because it shows members are celebrating a colleague’s work without an explicit request. A rising recognition rate reflects voluntary, mandate-free participation, while a low and flat one often means recognition depends on managers who don't have time for it. To improve it, make giving recognition a two-click action anyone can do from the landing page, then seed the habit by having leaders recognize people publicly and by name, so the behavior has a visible model to copy.

Taken as a set, these separate a platform people contribute to from one they only consume. But contribution is worth little if people can't find what they came for.

Findability: Can people find and act on what they need?

This is the layer teams skip most often, and the one employees usually judge the platform on most. It also drives engagement metrics we discussed in the previous section because when one search comes back empty, the team might look elsewhere the next time.

To make sure your content is findable, track these metrics:

  • Zero-result rate (searches returning nothing ÷ total searches): This is the most actionable number in your analytics, because every zero-result search is an employee telling you what they expected to find and couldn't.

    Pull your top 20 search terms and their result counts. A 40% zero-result rate across those top queries tells you that the content behind those searches is missing, mislabeled, or buried. Create or retitle the pages for the highest-volume failed queries first, then watch the rate for those terms fall.
  • Search success rate (searches that end in a click or a completed task ÷ total searches): This catches the near-misses a zero-result count won't. A search can return results and still fail if none of them were right. Low success can co-exist with a low zero-result rate when the content exists but the title doesn’t immediately or completely clarify what’s inside the page. The first move is renaming pages to match the real query wording sitting in your search log.
  • Stale content rate (pages not updated in 12+ months ÷ total pages, weighted by traffic): A forgotten page nobody opens does little harm, but a heavily used one going stale can do serious damage. An employee who searches for the current expense reimbursement limit and acts on a figure from 2021 learns that search can't be trusted. Start with the pages that get the most traffic and haven't been touched in a year, and work down from there, refreshing what people actually rely on before anything else.
  • Confirmed reads versus views (confirmed acknowledgments ÷ target audience): A view only tells you a page loaded. A confirmed acknowledgment tells you a named person read it. For a routine post, a view is enough. When the message carries consequences, like a compliance update, a safety notice, or a policy change, you need to know the right people actually read it, so measure confirmed reads against the target audience instead of total views.
  • Bounce rate by page type: A quick exit means different things in different places. Someone who lands on a single reference answer, gets it, and leaves in ten seconds likely got exactly what they needed. The same quick exit from a navigation hub or a landing page signals people can't find their next step.

    Time on page cuts both ways too. A long session can mean deep engagement or someone who can't find what they need, so judge it against what they actually searched for. Read bounce alongside your search and session data, and act on it only where a fast exit points to a dead end.

Together, search and content metrics tell you whether the platform does its core job. What they can't tell you is how people feel about it, which is where sentiment data earns its place.

Sentiment: The "why" your usage data can't explain

Usage data shows what employees did, but it can't tell you why. And it tends to spot trouble late, after engagement has already dipped. Sentiment data catches the frustration earlier and names the specific problem no aggregate number can.

Here are some metrics to track sentiment:

  • Satisfaction score (rating the platform positively on a short, recurring survey): Direct sentiment is a core success signal for any intranet, alongside usage and engagement. The pattern to watch is healthy adoption paired with falling satisfaction, an early warning your usage charts won't show for another quarter. When the score drops, the open comments below usually tell you what broke.
  • eNPS read as a trend (percentage of promoters - percentage of detractors ): Track the trend eNPS is moving in rather than the absolute score. There's no traceable universal "good" eNPS for an internal platform, so a fixed target is a distraction. But a score sliding for two quarters is a clear signal, even while it's still positive.
  • Pulse checks and open comments: Short, frequent, one-question checks catch sentiment shifts faster than an annual survey will. The comments are often more useful than the score itself because they point to specific problems you can fix.“I can never find the current org chart” gives you a clear action, while a 62% satisfaction score doesn’t, so read them line by line.

Sentiment takes longer to gather than a login count, since a survey cycle has to run its course. That's the tradeoff for the one signal that arrives before your usage charts move and the only layer that explains the other three. When a number shifts and you can't see why, this is usually where the answer is.

Related: For the survey mechanics behind eNPS and pulse checks, see how to measure employee engagement and running effective engagement surveys.

How to read the metrics together

No single metric proves the platform is working. The diagnosis comes from reading them as a set. A handful of patterns cover most of what you'll actually find:

  • High adoption, low engagement: People log in but don't do anything once they arrive. Usually the platform is delivering updates they can read passively but giving them nothing that asks for a response or rewards a return visit, so it becomes somewhere they check in and leave. Check the content engagement and reply rates to confirm the problem, then give them reasons to come back on their own: their own tasks waiting on the landing page, a recognition feed worth checking, and posts that ask a question instead of just broadcasting.
  • High engagement, low search success: The regulars are active, but they often can't find things. This could be a content and findability problem hiding behind a healthy-looking activity feed. Start with your top failed searches and identify content that needs adjustments to make it more searchable.
  • Healthy usage, falling satisfaction: If the numbers look fine but the mood doesn't, something is frustrating people that their behavior hasn't caught up to yet. Take a look at the satisfaction score, and if that seems low or trending downward, go straight to the open comments to identify the issue.
  • Strong overall averages, weak segments: A 65% participation rate can hide the fact that a specific warehouse location’s participation sits at 12% while the head office’s 90% participation rate does the heavy lifting. The average hides the group you most need to reach, which is why every metric above is worth breaking down by team, role, and location.

Most of what you'll find maps to one of these four. The value is in knowing which pattern you're looking at, because the pattern is what points to the fix.

A falling satisfaction score tells you to open the comments; a strong average hiding a weak segment tells you to break the number down by location. Name the pattern, and you already know where to start.

Best practices that keep your measurement honest, then turn it into action

Measurement is step one. To make sure these figures actually help you improve intranet effectiveness, follow these best practices.

Benchmark against your own baseline first, an industry cohort second

The honest answer to "what's a good number?" is that your own trend line is more reliable than any universal figure. Illustrative ranges exist, like the intranet KPI reference points published by Digital.gov, but they're examples to hold against your own history rather than targets to hit.

Vendor claims that "60 to 70% adoption is healthy" rarely trace back to independent data, so treat them with caution. Measure this month against last quarter first, then compare to a peer cohort for outside context.

Segment every metric by team, role, and location

Organization-wide averages hide the teams that need the most help. So break each core metric down until "who hasn't engaged yet" becomes a specific, named list instead of a comfortable company-wide figure. Splitting a single 65% average by site can help you identify sites that actually need attention.

Pair the hard numbers with what employees tell you

A metric can tell you participation slipped without telling you whether the platform frustrated people or something outside it did. Pair your usage data with direct input from short pulse surveys, an eNPS-style recommendation question, and open comments. The numbers point to where to look, and the comments explain what to fix. A dip in participation could be a platform problem or a wider cultural one, and only the qualitative side tells you which.

Close the loop: Pair every finding with a decision, on a set cadence

Collecting a metric you never act on wastes the effort of pulling it. Tie each reading to a move, so a stale top-traffic page gets rewritten, a silent team gets a check-in, and a failed-search spike gets new content. Review operational numbers weekly so you catch problems fast, and review trends monthly or quarterly.

What matters is that every review ends in an action, which is what makes measurement part of a real internal communication strategy rather than a reporting obligation.

Make the measurement easy to run, and easy to act on

Most of the friction in measuring an intranet is assembling the data by hand. Manually collecting data can include exporting logins, matching them to headcount, and chasing down who read what. But the faster and easier that work is, the more often you'll actually do it. A platform like Jostle removes a few of those steps.

Its Engagement Metrics benchmark your participation against your own history and your industry cohort out of the box, so "compared to what?" is answered without a spreadsheet. And Sign-off on News records confirmed reads by name, so you always know who received the update instead of just assuming that everyone did. You can even export reports of sign-offs for compliance tracking.

These capabilities take manual work off your plate. And that leaves you with more time to act on insights you get from these metrics.

Turning the numbers into outcomes leadership cares about

Leadership rarely cares about participation rates or search success on their own. They care about what those numbers mean for the business, which could be less time spent hunting for information, faster onboarding, fewer repeated questions, stronger policy compliance, and better communication across teams.

The job is to connect each platform metric to one of those outcomes. A higher search success rate can mean employees spend less time asking colleagues for help. Faster policy acknowledgment can reduce compliance risk. And broader participation can show that information is reaching more locations and frontline employees.

Be careful not to claim that the intranet caused every improvement. Engagement and productivity depend on many factors. Present the platform as one contributor, then support the connection with specific evidence such as time saved, fewer support requests, faster acknowledgments, or improved onboarding completion.

Related: Leaders and employees often read these signals differently. For where that split comes from, see the engagement gap white paper.

Turn your intranet metrics into a story worth acting on

Measuring intranet effectiveness comes back to the same three questions, read together: Are people showing up, are they doing anything once they're there, and can they find and act on what they need?

Hold every number against your own baseline, break it down by team and location until you can see the group the average is hiding, pair it with what your people tell you, and end every review with a decision.

Do that, and measurement stops just filling a monthly report and starts telling you what to fix next.

Everything above assumes you can pull adoption, engagement, findability, and sentiment signals without stitching them together from separate tools. Jostle builds that measurement in, from participation rates to confirmed reads to the pulse checks that tell you why a number moved.

Book a demo to see what it looks like against your own numbers.