Start with the job your website needs to do
A performance target is only useful if it reflects something the website is meant to accomplish. Before opening an analytics dashboard, write down the actions that matter to your organization. A local service provider might want visitors to find its service area and submit an enquiry. A community group might want people to discover events and sign up. These are different jobs, so they need different measures.
Choose one primary action for each important type of page, then note the steps a visitor may take before reaching it. For an enquiry page, you might track visits to the page, clicks on the contact link and completed form submissions. Include a check on the visitor experience: whether people can find the information they need, especially on a phone or when using a keyboard. Not everything valuable will appear as a single number, so pair measurement with occasional review of the pages themselves.
Decide what each measure means
Write a short definition for every measure before collecting baseline figures. “Enquiries,” for example, could mean form submissions, email-link clicks or all contact attempts. Pick a definition that your tools can support and keep it consistent. If you count completed forms, confirm that the event fires only after a successful submission, not when someone merely opens the form.
Separate outcomes from supporting signals. Completed enquiries are an outcome; visits to the contact page are a supporting signal. A rise in page visits without a rise in enquiries may point to a problem worth investigating, but it does not prove what caused it. Keep the initial list small enough that someone can review it regularly. A few clearly defined measures are more useful than a crowded dashboard whose numbers nobody can explain.
Check the data before recording a baseline
Test the journey yourself on a desktop and a phone. Can you reach the relevant page from the navigation? Does the form work? Does your analytics tool register the intended action once, rather than twice? Check whether internal staff visits, test submissions or spam are likely to distort the figures. Record known limitations instead of quietly treating imperfect data as precise.
Make sure your approach to analytics respects visitors’ privacy and your organization’s applicable requirements. Collect only information you have a clear reason to use, and avoid putting personal details from forms into analytics reports. If consent settings mean some visits are not measured, document that limitation. The purpose of a baseline is to support better decisions, not to imply that a dashboard captures every visitor.
Capture a representative starting period
Choose a period long enough to include your usual pattern of activity, then record its dates and the measures you defined. For a small organization, looking across several weeks may be more informative than judging a single busy day. If your work varies by season or depends on a recurring event, note that context and compare similar periods where possible. A baseline is a documented starting point, not a permanent verdict on the website.
Add brief notes about anything that could affect interpretation: a newsletter send, a website change, a temporary outage or a campaign. Break down results only where it helps answer a question, such as whether mobile visitors reach the enquiry form or which pages lead people to it. Small groups of visits can fluctuate sharply, so avoid drawing strong conclusions from a narrow segment. Keep a simple record of the measure, its definition, the period, the result and any caveats.
Use the baseline to set targets you can act on
Once you know the starting point, choose a change your team can reasonably influence. If visitors reach a service page but rarely continue to the contact form, a useful goal might focus on improving that path. You could clarify the service description, make the next step easier to find and check that the form is usable. Then monitor both the supporting steps and the completed enquiries, rather than treating traffic growth as the only sign of success.
State each target with a measure, a review period and a planned action. For example, an organization could aim to improve completed enquiries over its baseline while testing clearer contact prompts, then review the result after the next comparable period. If the measure moves, look for other possible explanations before crediting the edit. If it does not, revisit the page and the tracking. A good baseline makes targets less arbitrary and gives your team a repeatable way to learn.
