What Should You Actually Measure from Your Website?

A report should help you decide what to do.

A dashboard can be full of movement without answering a useful question. Sessions increased, engagement changed and a ranking moved, but is the website helping the business?

Start with what the site is meant to support. For a service business, that might be suitable project enquiries or quote requests. For a shop, it may be completed purchases. A publisher or organisation could have a different purpose. The important thing is to define useful activity before choosing the charts.

A sensible report connects outcomes, the actions leading towards them and the audience reaching the website. That gives each number a job rather than treating everything measurable as equally important.

Give the measures a clear order of importance.

Business outcomes

Look at the quality of enquiries, accepted quotes, sales or other meaningful results in the business’s own records. A website form submission begins a conversation; it does not establish that a new customer has been won. Connecting reporting to what happens after the enquiry keeps the discussion commercially useful.

Completed website actions

Measure successful contact or project submissions, quote requests and completed purchases where relevant. Google Analytics calls important actions key events. The label matters less than defining the action accurately and checking that it records when the intended outcome actually happens.

Supporting signals

Phone and email clicks indicate intent, but do not prove that a call connected or an email was sent. Service-page visits, relevant landing pages and traffic sources help explain the route towards a useful action. Keep these signals separate from completed enquiries rather than adding them together into one impressive total.

On SERO’s own website, our measurement design distinguishes successful project enquiries and SEO review requests from supporting clicks. That is the principle: measure the completion you care about, then use other activity to understand it.

A conversion rate needs a definition.

A conversion rate expresses how often a defined action happens within a defined group. The denominator matters. Sessions containing an enquiry divided by sessions is different from enquiries divided by users. Decide which question you are answering and keep the calculation consistent.

For an illustrative example, if 10 of 500 measured sessions contain a successful enquiry, that session enquiry rate is 2%. It does not mean 2% of every person who encountered the business became a customer. Nor is 2% a recommended benchmark.

Compare similar traffic and periods. A campaign aimed at people ready to buy may behave differently from an educational article reaching people early in their research. Combining them can hide useful differences. With small samples, review the underlying counts as well as the percentage before drawing conclusions.

The practical question is which sources and landing pages contribute to suitable actions. Our article on why more traffic is not always the answer explores what to do when the destination itself needs attention.

Analytics is a partial view of the customer journey.

Some visitors do not consent to analytics, some block tracking and others move between devices or contact the business later through another route. On SERO’s site, optional analytics is held back until consent. A genuine enquiry can therefore exist in the business’s records without a matching analytics event.

How much is observed depends on the implementation. Google’s consent documentation describes how consent settings affect measurement behaviour. Do not assume that every analytics product or consent setup sees the same activity.

Attribution assigns credit to a source under particular rules. It is not a complete account of everything that influenced a customer. Someone may hear a recommendation, watch a video and later search for the company name. The final recorded visit will not necessarily explain the whole decision.

Use reporting alongside enquiry records and conversations with customers. Investigate discrepancies without expecting the totals to match perfectly or bypassing people’s privacy choices to make them do so.

Finish with a decision, not another chart.

A useful review can be short: what suitable activity happened, where it came from, which pages helped and what needs attention next. Rankings and impressions can help diagnose visibility, but should not crowd out the actions the website exists to support.

Check measurement quality before responding to a sudden change. Duplicate events, a broken success signal or a changed consent setup can alter reports without the business changing in the same way. Keep definitions stable and annotate important website or campaign changes.

You should leave the report knowing whether to improve a page, investigate a source, fix measurement or continue an approach that is producing useful conversations. If the numbers cannot inform any of those decisions, the report probably needs fewer numbers and a clearer question.