What is a good website performance score?

There is no single score that captures website performance. Speed tools give a technical score, which is worth keeping healthy, but a full evaluation also looks at whether visitors complete tasks, whether the right people convert, whether the site is accessible and found for relevant searches, and whether inquiries become clients.

Do Core Web Vitals affect search rankings?

Google has said that page experience signals, including Core Web Vitals, are part of its ranking systems, but relevance and content quality matter far more. Improving them is worthwhile mainly because a fast, stable site is better for visitors. Do not expect speed fixes alone to transform rankings.

How do I measure website usability without a big budget?

Ask a few people who match your audience to complete your key tasks on their phones while describing what they are thinking, and take notes. Combine that with analytics paths, internal search terms and questions front-line staff hear. These low-cost methods reveal most serious usability problems.

How often should we evaluate website performance?

Monitor technical health continuously or monthly, since broken forms and outages need fast attention. Review the full set of measures quarterly, which is often enough to see trends without reacting to noise. Seasonal businesses should compare against the same period last year rather than the previous quarter.

How to Evaluate Website Performance Beyond Page Speed

Why page speed became the default measure

Page speed is easy to measure, easy to compare and comes with a score out of a hundred. That makes it the number people reach for when someone asks whether the website is performing. Vendors quote it in proposals. Board members run the test on their phones.

Speed matters. A slow site loses visitors before they read anything, and search engines use loading and responsiveness measures as one of many signals. But a speed score tells you how quickly a page arrives, not whether it does anything useful once it gets there. Plenty of fast sites confuse visitors, bury the contact path and attract the wrong traffic.

A better evaluation looks at the site in layers, from the technical foundation up to the business result. Each layer needs only one or two measures. The value comes from reading them together.

Layer one: speed and technical health

Keep speed in the picture, but measure it properly. Google’s Core Web Vitals look at loading, responsiveness to input and visual stability, and they are reported from real visitors’ experience where enough data exists, which is more meaningful than a single lab test from a fast office connection. Look at mobile results first, since that is where most problems show up.

Technical health goes beyond speed:

Broken links and missing pages.
Forms that fail silently or send to the wrong place.
Pages that look broken on certain devices or browsers.
Security updates and certificates that are current.
Uptime, especially during your busy season.

When speed is poor, the cause is often not the site itself but what has been added to it over time: chat widgets, several analytics and advertising tags, embedded video, oversized images uploaded by editors. A periodic look at what third-party scripts load on each page is a quick way to find weight nobody remembers agreeing to.

A useful question: when was the last time someone submitted every form on the site and checked where it went? If nobody knows, start there. It is the cheapest high-value test in website evaluation.

Layer two: can visitors complete their tasks?

Usability is the question of whether people can do what they came to do without effort or confusion. It is the layer most organizations never measure, because it does not come with a dashboard by default.

Start by listing the three to five most important tasks on the site: find a service and understand it, find a location and its hours, request a quote, book an appointment, find a document. Then evaluate each one:

Watch people try. Ask a few people who match your audience, but have not seen the site, to complete each task on their phone while thinking out loud. You will learn more in an hour than from a month of analytics.
Check the paths in analytics. Where do people enter, where do they go next, and where do they leave? A high exit rate on a page that should lead to a form is a signal.
Read internal site search. If the site has search, the terms people type are a list of things they could not find in the navigation.
Ask the phones. Front-line staff know which questions callers say they could not answer on the website.

Session recording and heatmap tools can help, but use them with a clear question in mind and with privacy settings configured properly. Watching recordings without a hypothesis is an excellent way to lose an afternoon.

Layer three: are the right visitors taking action?

Conversion measures whether visitors take the actions that matter to the business: inquiries, bookings, calls, applications, purchases. The common mistake is to measure total conversions without asking whether they are the right ones.

Measure conversions by type. A quote request and a newsletter signup are not equal, and a total that adds them together hides what is happening.
Measure conversion rates by audience and source. Organic search visitors, paid visitors and referral visitors behave differently. A falling overall rate may simply mean you are getting more top-of-funnel traffic.
Measure lead quality. Ask sales which inquiries were real opportunities and record it in the CRM. A site that produces fewer, better inquiries may be performing better than one that produces many poor ones.

The principles behind conversion-focused design are covered in what makes a B2B website convert, and the testing approach in the conversion rate optimization playbook. If conversions suddenly fall, work through how to diagnose a drop in website conversions before changing anything.

Layer four: can everyone use it?

Accessibility belongs in any performance evaluation, because a site some people cannot use is underperforming for those people by definition. It also overlaps heavily with usability: clear headings, readable contrast, labelled forms and keyboard support help everyone.

Evaluate it with an automated scan for common errors, a keyboard-only attempt at your key tasks, and a periodic check with a screen reader. Track the number of known issues over time, not just a pass or fail. The practical approach is laid out in website accessibility planning for Canadian organizations.

Layer five: is it found for the right things?

Findability is whether the people you want can find you when they search, in search engines and increasingly in AI assistants. The useful measures are not total traffic but:

Impressions and clicks for the searches that match your services and locations, from Search Console.
Which pages receive organic visits, and whether they are the pages you want people to land on.
Whether the business appears accurately in map results and AI-generated answers for relevant questions.

A site that ranks well for informational searches unrelated to what you sell can look healthy in a traffic report while producing nothing. Look at the searches, not just the volume.

Layer six: business outcomes

This is the layer leadership actually cares about, and the one that requires connecting the website to systems beyond it. Questions to answer:

How many inquiries from the website became qualified opportunities, and then clients?
Which sources produced them?
What tasks did the website handle that would otherwise have taken staff time, such as bookings, payments or document requests?
Is the website helping close deals that started elsewhere, such as a referral who checked the site before calling?

Answering these needs source data passed into your CRM and a habit of recording outcomes there. It is rarely perfect, and it does not need to be. A reasonable answer reviewed regularly beats a perfect one never assembled.

Putting it together: a simple scorecard

Pick one or two measures from each layer, capture a baseline, and review them together each quarter. Reading them side by side is what makes the evaluation useful. A worked illustration, for a hypothetical multi-location physiotherapy clinic:

Technical: mobile Core Web Vitals passing on key templates; every booking form tested monthly.
Usability: new visitors can find their nearest clinic and book within a few taps, checked with a handful of people each quarter.
Conversion: online bookings by location and source.
Accessibility: count of known issues, trending down.
Findability: clicks for physiotherapy searches in each clinic’s city.
Outcome: online bookings as a share of new patients, and front-desk calls about booking.

In that example, a quarter where speed improved but online bookings fell in one city points somewhere specific (perhaps the booking widget for that location, or a change in local search results). A speed score alone would have reported good news. The executive view of this kind of scorecard is covered in what an executive marketing dashboard should show.

If someone reports on your website’s performance, internally or as a vendor, these questions keep the report honest:

Which of these numbers would change a decision we are about to make?
Is this measured on mobile, from real visitors, or from a single lab test?
Are conversions split by type, and does sales agree the leads were real?
What is the baseline, and is it from the same season?
What did we change last quarter, and did the measure it was meant to move actually move?

When we evaluate a site at the start of a website project, the speed test takes a few minutes. The rest of the layers take longer, and they are where the useful findings almost always are.