Google Analytics shows you pageviews. Your hosting dashboard shows uptime. Your agency report shows a design that looks impressive. None of that tells you whether your website is functioning as institutional infrastructure: whether it is serving the stakeholders who matter most, meeting the compliance obligations your Board is accountable for, and performing at a level that supports rather than undermines the organisation’s credibility.

Performance benchmarks for nonprofit websites need to be framed around governance outcomes, not marketing metrics. This guide covers the benchmarks a Communications Director should track, what ‘good’ looks like for established nonprofits with £500K–£10M budgets, and how to present performance to a Board that cares about institutional risk rather than traffic numbers.

One rule runs through all of it. Every figure below is either traced to a named source with a date, or it is labelled as a working standard rather than a sector benchmark. An untraceable threshold is not a benchmark. It is an opinion with a decimal point in it, and it will not survive the first trustee who asks where the number came from.

Why most nonprofit performance metrics miss the point

The metrics that web agencies report (traffic growth, bounce rate, session duration, pages per visit) describe activity. They do not describe whether the website is doing its job. A nonprofit website that attracts plenty of visitors but has a broken donation flow, inaccessible programme pages and outdated governance documents is failing at its institutional purpose regardless of its traffic numbers.

The benchmarks that matter for nonprofits are these. Can your primary stakeholders find what they need? Can they complete the actions that serve the organisation’s mission? Does the site meet its compliance obligations? Is the content accurate and current? Those are governance benchmarks, not marketing benchmarks.

Page speed and Core Web Vitals

Google’s Core Web Vitals are the closest thing to a standardised performance measure for websites. They cover three things: how quickly the main content loads (Largest Contentful Paint, LCP), how responsive the page is to interaction (Interaction to Next Paint, INP), and how stable the layout is as the page loads (Cumulative Layout Shift, CLS).

Two points about how the thresholds work, both of which change what you should put in front of a Board. Each metric is assessed at the 75th percentile of real page views, segmented across mobile and desktop. So the benchmark is what three quarters of your actual visitors experience, not the score you saw the one time you tested from the office. And Interaction to Next Paint replaced First Input Delay as a Core Web Vital on 12 March 2024. If a report on your desk still shows FID, it is quoting a metric Google retired more than two years ago, which tells you something about how the report was produced.

LCP target: under 2.5 seconds. In the audits I run, this is the one that fails most often, and the causes are almost always the same three: uncompressed hero images, several tracking scripts loading at once, and a heavy animation library shipped on every page whether or not the page animates.

INP target: under 200 milliseconds. An INP failure is usually a third-party embed (a chatbot, a donation widget, a social media feed) occupying the main thread while the visitor is trying to press something.

CLS target: under 0.1. Layout shift happens when elements load after the page has appeared and push content around. On nonprofit sites the usual causes are images without defined dimensions, web fonts arriving late, and a cookie consent banner that shoves the page down as it appears.

Test on mobile first. M+R Benchmarks 2026, reporting on 2025 data, found that mobile users were 52 per cent of all visits to nonprofit websites in its sample, against 48 per cent on desktop. A pass achieved only on desktop is a pass on the smaller half of your audience.

Run PageSpeed Insights on your homepage, a programme page and your donation page. If any of the three fails on mobile, it is a priority fix, both for the people using the site and for search visibility.

What a trustee should ask when this moves: which page failed, how many visits does that page carry, and is the cause something we own or something a third-party supplier owns?

Accessibility benchmarks

WCAG 2.2 AA conformance is not optional for nonprofit websites. The Equality Act 2010 creates accessibility obligations for any UK organisation providing services to the public, digital services included. The European Accessibility Act, Directive (EU) 2019/882, adds a further duty where an organisation supplies an in-scope service (in practice, online sales) to consumers in an EU member state. Reaching an EU audience is not by itself a trigger.

Colour contrast: 4.5:1 for normal text, 3:1 for large text. This is not a matter of taste. WCAG 2.2 success criterion 1.4.3 sets both ratios and defines large-scale text as at least 18 point, or 14 point bold. Check a contrast tool against your actual brand colour values, not against how the page looks to you.

Automated scan baseline: zero critical violations. Run an automated scan on your homepage monthly and watch for zero critical violations and a declining trend on serious ones. What that scan cannot do is more important than what it can. Deque’s Automated Accessibility Coverage Report, published in 2021 from more than 2,000 audits across over 13,000 pages, found that 57.38 per cent of issues by volume were caught by its automated tests; Deque’s argument is that the widely repeated figure of 20 to 30 per cent undercounts coverage because it counts success criteria rather than actual issues. Take the more generous number and roughly two issues in five are still invisible to the scanner. A green automated report is a floor, not a verdict.

For scale, the WebAIM Million 2026 report found detectable WCAG 2 failures on 95.9 per cent of the top one million home pages, at an average of 56.1 errors per page, up from 51 the year before. Being typical is not a defence, and it will not read as one in a funder’s due diligence.

Keyboard navigation: complete. Every interactive element should be reachable and operable by keyboard alone. Press Tab through your homepage. If you cannot reach the navigation, the donation button or the contact form, you have a critical failure that no automated scan will have told you about.

Alt text coverage: 100 per cent of meaningful images. Every image that carries information needs a descriptive alt attribute. Decorative images take an empty one.

For the full accessibility framework, see WCAG AA Accessibility on Webflow. For the governance argument for accessibility investment, see WCAG Accessibility as a Nonprofit Governance Obligation.

What a trustee should ask when this moves: has anything been added to the site since the last scan, and who signed it off?

Content currency benchmarks

These are working standards rather than sector benchmarks. No published report sets them, and I am not going to dress them up as one. They are the thresholds I apply in audits because they are the points at which an institutional reader starts drawing conclusions.

Content accuracy is a credibility metric. Funders conducting due diligence check whether programme descriptions match current delivery, whether the annual report is current, and whether governance documents are visible and up to date.

Annual report: current financial year. If a funder visits and the most recent annual report is two years old, they will form a view about governance quality before they reach anyone at your organisation.

Programme descriptions: reviewed within the current quarter. Programmes evolve. New geographies, updated eligibility, changed delivery models. The website should reflect the current state, not the state it was in when the page was first written.

Leadership and governance: current. The team page should list current staff and leadership, and trustees should be named. For a registered charity the trustee list is already public on the Charity Commission register, so omitting it from your own site does not protect anything. It just reads as a site nobody maintains. Departed staff should come off within a week of leaving.

News or blog: activity within the current quarter. If the most recent post is six months old, it signals either that the organisation has stopped communicating or that the website is not being looked after. Both readings cost you.

Compliance documents: dated within the current year. Privacy policy, accessibility statement, safeguarding policy. Each should carry a review date inside the past 12 months.

Currency is a scheduling problem more than a writing problem, and it is best solved by putting the reviews in a calendar someone owns. See a nonprofit website maintenance schedule for how that is set up.

Conversion benchmarks

For nonprofits, conversion means one thing: did the visitor complete an action that serves the mission? The relevant actions are donations, contact form submissions, newsletter sign-ups, document downloads (particularly annual reports and governance documents), and volunteer or programme applications.

Donation page conversion. M+R Benchmarks 2026, reporting 2025 data, found that 11 per cent of desktop users who landed on a primary donation page completed a gift, against 8 per cent of mobile users, and that 1.6 per cent of all website visitors from every traffic source made a donation. Two caveats before you put those numbers in a Board pack. The sample is weighted towards large US organisations and the money is reported in US dollars, so it is a reference point rather than a standard a UK charity has failed to meet. And the desktop-to-mobile gap in that data is a useful diagnostic in itself: if your own mobile rate sits well below your desktop rate, the problem is the mobile donation experience rather than the appeal.

Everything else: your own trend is the benchmark. There is no named sector report publishing reliable conversion figures for general enquiry forms, programme application forms or newsletter sign-ups at charity scale, so do not accept one from anybody. What matters operationally is the direction of travel in any case. Record the rate each quarter, and treat a decline against flat traffic as a fault to investigate: the form, the mobile layout, or something that changed upstream.

Set these up as Key Events in GA4, or as the equivalent conversion events in whichever analytics platform you run. See Analytics and Consent on Webflow: GTM, GA4, and Consent Mode V2 for implementation guidance, and note that privacy-focused alternatives such as Plausible, Fathom or Cloudflare Web Analytics measure the same conversions without setting cookies, which simplifies PECR consent obligations.

What a trustee should ask when this moves: did the rate change, or did the traffic mix change? A drop caused by a paid campaign bringing in colder visitors is a different problem from a drop caused by a broken form.

Search visibility benchmarks

Google Search Console provides the data. The metrics that matter are these.

Branded search impressions: stable or growing. If impressions for your organisation’s name are declining, something has changed, either in search behaviour or in how Google evaluates your site. Investigate rather than assume.

Non-branded impressions for mission terms: growing. These are searches about what you do rather than who you are: your issue area, your service type, your geography. Growth here indicates the site is building topical authority.

Indexed pages: stable and matching expectation. An unexpected drop signals a technical problem. Growth faster than your publishing rate suggests duplicate URLs or parameter-based crawling issues.

You will be offered a click-through-rate benchmark at some point. Decline it. Rates vary so much by query type, with branded queries running far ahead of informational ones, that a single cross-sector average tells you nothing about whether your own titles and descriptions are working. Use your own twelve-month baseline instead, segmented into branded and non-branded, and treat a change against that baseline as the signal.

Two pieces of context from M+R Benchmarks 2026 are worth carrying into the discussion. Organic, non-paid search still accounted for 39 per cent of all visits in its 2025 sample, and that traffic declined over the course of the year. A falling organic number is now partly a sector condition rather than purely a failure of your own site, which is a distinction a Board is entitled to have drawn for it. How organisations are being found when search itself changes shape is covered in answer engine optimisation for nonprofits.

Uptime and technical health

Uptime: 99.9 per cent. Treat this as a contract term rather than a sector benchmark. It is the figure to require of a hosting provider and to check you are actually getting, not a number anyone has measured across the charity sector.

SSL certificate: valid and auto-renewing. An expired certificate breaks the entire site and shows a security warning to every visitor. Managed platforms handle renewal automatically, but verify that it is working rather than assuming.

Console errors: zero critical errors. Open browser developer tools on your homepage. JavaScript errors in the console mean something is broken, and third-party scripts are the most common cause.

Broken internal links: zero. Run a crawl quarterly. Screaming Frog’s SEO Spider is free up to 500 URLs per crawl. Every broken internal link is both a dead end for a visitor and a signal that the site is not being maintained.

Quarterly review framework

These benchmarks should be reviewed quarterly. The review should produce a one-page operational summary for the Communications Director and a traffic-light overview for the Board.

The quarterly review covers: Core Web Vitals on three key pages, accessibility scan results and trend, content currency on five key pages, conversion for donation and contact forms, search visibility trends from Search Console, and technical health (uptime, SSL, console errors, broken links).

The output should be a simple dashboard: green for benchmarks being met, amber for declining trends, red for benchmarks not being met. The Board version needs the traffic lights and one sentence per metric. The operational version needs the data, the trend and a prioritised action list for the next quarter.

One discipline is worth imposing on the Board version. Every red and every amber carries the source of its threshold in the same sentence. A trustee reading ‘LCP slower than the 2.5 second Google threshold’ can act on it. A trustee reading ‘page speed: amber’ can only ask who decided that, and the answer had better not be nobody in particular.

For related guidance, see a website maintenance schedule and optimising images for performance.

Further reading

What these benchmarks actually tell you

Any single metric in isolation is meaningless. What matters is the pattern across several dimensions over time. A site with excellent Core Web Vitals but outdated content is technically fast and institutionally negligent. A site with current content but broken accessibility is well maintained and legally exposed. A site with good conversion but declining search visibility is serving the audience it already has and failing to reach anyone new.

The benchmarks above cover the performance dimensions a nonprofit communications team needs in order to manage a well-governed website, and, just as usefully, they mark out where no dependable sector figure exists. Where a published benchmark exists, use it and cite it. Where it does not, measure your own trend and say plainly that is what you are doing. A Board can work with either. What it cannot work with is a number nobody can source.