A website conversion rate can tell you whether visitors are taking an intended next step, but the percentage is only useful when you know exactly what counts as a conversion and which visits are included. A quote request, phone call, appointment booking and online purchase are different outcomes; combining them can hide what is working. Before redesigning a page or moving a button, define the action that matters, calculate it consistently and check that the tracking records it once. Then look beyond the headline rate. Traffic source, device, landing page, service and the steps between arrival and completion can reveal where the journey becomes difficult. A low rate may reflect a mismatch between a searcher’s need and the page, an unclear offer, a slow or broken form, or simply too little data to draw a sound conclusion. The right response is not automatically a new layout. This guide explains what to measure, how to check the evidence and how to choose a sensible next step for your small business.
Content
Define the conversion
Start with the customer action
Before calculating a conversion rate, decide what counts as a conversion for your business. It should be a specific action that represents meaningful progress towards a customer or business outcome—not simply an interaction that is easy to track. For a service business, that might be a completed quote request or a booked consultation. For an online shop, it may be a completed order. Write down the exact action and the point at which it is considered complete.
Be precise about the distinction between starting and completing an action. A visitor clicking a phone number, opening a contact form or adding an item to a cart may show interest, but none necessarily means the customer has made an enquiry or purchase. You can track those smaller steps to help understand the journey, but label them as supporting events rather than treating them as equivalent to your main conversion. Otherwise, a page may appear to perform well while producing few outcomes that matter to the business.
Choose a denominator that fits
A conversion rate compares conversions with a defined group of website visits or visitors. A common session-based calculation is: completed conversions divided by sessions, multiplied by 100. If you use this approach, decide whether you are counting conversion events or sessions in which at least one conversion occurred. Counting events can produce a rate above 100 per cent if a session includes multiple orders or enquiries. Counting converting sessions answers a different question and avoids that interpretation. Neither measure is automatically right; the important thing is to name the measure and use it consistently.
You may instead calculate the share of individual visitors who convert. That can be useful when you want to understand how many people take the action, but it requires a consistent definition of a visitor and a way to recognise repeat visits. Session-based and visitor-based rates are not interchangeable: one person may visit several times before converting. Check which denominator your analytics report uses, and record it alongside the rate so comparisons remain meaningful.
Match the denominator to the question you are asking. If you are assessing how often visits to a landing page lead to an enquiry, sessions on that page may be relevant. If the question is how many people eventually enquire, unique visitors may be more appropriate, provided your measurement method supports that comparison. Avoid switching denominators between pages or reporting periods without making the change explicit.
Document the definition before comparing
Keep a short written definition of the conversion, its completion point, the denominator and the period being measured. For example, specify whether a form conversion means a successful submission rather than a button click, and whether you count one converting session or every submission. This makes it easier for anyone reviewing the figures to interpret them consistently.
Also decide whether all conversions are equally useful. A completed enquiry may be a valid conversion even if it later proves unsuitable, but you should not quietly redefine the event whenever results change. If you later distinguish qualified enquiries from all submissions, report that as a separate measure. A clear definition gives you a dependable starting point for reviewing performance; it does not, on its own, explain why a page is succeeding or struggling.
Check your baseline
Confirm the data before interpreting the rate
Before deciding that a page is performing well or poorly, check that the figures represent real customer actions. Confirm which analytics tool or reporting system supplies the data, who has access to it and whether tracking is active on the pages involved. If the website or analytics setup has changed recently, note when the change happened; figures from before and after may not be directly comparable.
Inspect the conversion event itself. It should record the chosen action when it is completed, not merely when someone opens a form, clicks a button or reaches the page containing it. Where practical, carry out a test submission and check that it appears once in the relevant report. Also check for duplicate event firing, such as an event being recorded both on submission and when a confirmation page loads. A thank-you page that can be revisited or refreshed may inflate counts if each visit is treated as a new conversion.
Compare analytics conversions with the records created by the process: for example, submitted enquiries in a form inbox, confirmed bookings in a booking system or completed orders in a sales record. These sources may not match exactly. A person might submit twice, a staff member might make a test submission, or privacy and consent settings may affect what analytics records. Look for a material unexplained gap and investigate it before treating either figure as definitive. Do not assume that a tracking report includes every visitor or action.
Choose a time period that gives enough activity to make a useful comparison, and write down the exact dates. Compare like with like where possible: account for weekdays, public holidays, seasonal demand and any known promotions or service interruptions. Avoid comparing a quiet period with a busy one and attributing the difference to a page change. If tracking was interrupted, a form was unavailable or a campaign sent an unusual burst of visitors, flag those dates rather than quietly folding them into the baseline.
Check the denominator as carefully as the conversion count. Confirm that the visitor or session measure is the same in every calculation and report. A rate based on sessions answers a different question from one based on users, and mixing those measures makes comparisons unreliable. Apply the same page scope and filtering rules each time. If internal visits are excluded in one report but included in another, or one report counts all site traffic while another counts only landing-page visitors, record the difference and correct it before comparing rates.
Finally, save a short baseline note with the dates, conversion definition, denominator, report source, relevant filters and any known tracking limitations. Include the raw counts as well as the calculated rate so that a later change can be checked against its inputs. If the number of actions is small, treat the rate as an early indication rather than firm evidence; a few additional conversions can shift it substantially. Resolve clear tracking faults first, then gather a consistent baseline before changing the page.
Compare visitor groups
Compare like with like
Once your tracking and baseline are reliable, break the results into useful visitor groups. An overall conversion rate can hide differences: one group may be completing the desired action while another struggles to reach it. Comparing groups helps you decide where to investigate, rather than changing a page that is already working for many visitors. Start with four practical dimensions: traffic source, device, landing page and service.
For traffic source, compare visitors arriving through channels recorded in your analytics, such as organic search, referral links or direct visits. Check that the source labels are meaningful and consistent; untracked links or changes to tracking can make a group appear smaller or less successful than it really is. Consider why people may have arrived from each source. Someone who searched for a specific service may have a different goal from someone who followed a general referral. A difference in conversion rate is a clue to investigate, not proof that the source caused the difference.
Device comparisons can reveal whether the experience differs on a phone, tablet or desktop. Review both the rate and the number of conversions, and check that your event tracking works consistently across devices. If one device group stands out, try the relevant page and key action on that device: look for text that is difficult to read, buttons that are hard to use, slow or confusing form steps, and errors. Do not assume the device itself is the problem; visitor intent, connection conditions and the way people use your site may also differ.
Compare landing pages to see whether particular entry points help visitors understand what to do next. Check that each page matches the likely reason for the visit, makes the next step clear and provides enough information to support it. If the pages target different services or audiences, avoid treating their conversion rates as a direct design contest. Group visitors by service as well, where your tracking allows it. A high-level service category can help show whether enquiries or bookings are concentrated in particular parts of your business, while keeping the comparison focused on similar actions and visitor intent.
Use differences to guide an investigation
For each comparison, keep the conversion definition and reporting period consistent. Note the number of visitors and completed actions alongside the rate: a striking percentage based on very few visits may change quickly and should not, by itself, drive a redesign. Avoid splitting the data into so many combinations that each group is too small to interpret. First look at one dimension at a time, then check a relevant combination if there is a clear reason to do so. For example, a device difference on one landing page may warrant a closer look, but only if there is enough data to make the pattern useful.
Hypothetical: If mobile visitors start a quote form but rarely finish it, inspect the form on a phone before changing the whole page. Check the form’s layout and steps, and test that its completion event is recorded correctly. Then compare the evidence with other devices and relevant service pages. This narrows the question to a specific part of the customer experience. Record what you found and what you plan to investigate next; the pattern identifies a place to look, not a guaranteed explanation or a reason to change multiple things at once.
Trace the customer journey
Find where visitors pause or leave
A conversion journey is the sequence of pages and actions between a visitor’s first meaningful visit and the outcome you want, such as an enquiry, booking or purchase. Map the main route for each important conversion rather than assuming everyone follows the same path. A person may arrive on a service page, read an FAQ, then return later to complete an enquiry form. Looking at the whole route helps distinguish a genuine obstacle from a step that is simply part of the decision process.
Start with the pages that lead into the conversion and the steps visitors must complete. For an enquiry, that might include a service page, contact page, form start, form submission and confirmation page. For a booking or checkout, list each relevant step, such as choosing a service, selecting a time, entering details and confirming. Use your analytics reports alongside the actual website. Check that important actions are recorded consistently, and test the route yourself on desktop and mobile. Note any page or action that is missing from the data; an incomplete journey can make a drop-off appear to happen earlier than it does.
Review exits at each stage, but do not treat every exit as a problem. A visitor may leave after finding a phone number, deciding the service is not suitable, or finishing what they came to do. An exit from a confirmation page is also different from an exit halfway through checkout. Interpret the page’s purpose and the next expected action before deciding that its exit rate signals friction. Where possible, compare the journey for visitors who complete the action with those who do not, using the same time period and comparable visitor groups.
Inspect the step where progress breaks down
If many visitors reach a form but few submit it, inspect the form itself before rewriting the page that leads to it. Check whether the fields are clear and necessary, whether required information is obvious, and whether errors explain how to fix a problem. Try submitting it with missing or incorrectly formatted details. On a phone, check that fields are easy to tap, the keyboard does not obscure the next step, and the submit button is visible. Also confirm that a successful submission produces a clear confirmation rather than leaving the visitor uncertain.
For bookings and checkout, review each transition separately. Look for unexpected requests, unclear fees or service details, confusing choices, broken links, slow-loading steps and errors that prevent progress. A sharp change in completion between two steps can point to an area worth investigating, but it does not prove what caused the change. Check whether tracking is working on both steps and whether the pattern persists across relevant devices or visitor sources. If numbers are small, treat the pattern as a prompt to inspect the journey, not as firm evidence.
Record the suspected friction, the evidence for it and what you observed when testing the route. For example, note the step where visitors commonly stop and whether the same difficulty occurs during a manual test. This gives you a specific issue to investigate next, rather than a vague instruction to make the page better. Keep the diagnosis separate from the proposed fix: more than one cause can produce the same drop-off, and the journey data is a guide to where to look, not a substitute for checking the experience.
Choose what to test
Turn a pattern into a testable change
Use what you found in the earlier checks to identify a specific problem worth addressing. A useful test starts with evidence: for example, a particular page has a consistently lower conversion rate than comparable pages, or visitors repeatedly leave at the same form step. A single day’s dip or one unusual visitor journey is not enough to establish a pattern. Check that the issue persists across a relevant period and is not explained by a tracking fault, a different visitor mix or a temporary change in demand.
Write down a simple hypothesis before editing the page: “If we make [specific change], then [measured customer action] may improve because [evidence-based reason].” This makes the proposed change and its rationale clear. For instance, if people are abandoning a form after reaching a field that asks for information they may not have to hand, you might test whether making that field optional helps more people complete the form. Treat this as a hypothesis, not a promise of improvement.
Keep the change focused. If you rewrite the headline, replace the imagery, shorten the form and move the call to action all at once, a different result will not show which change mattered. Where practical, change one meaningful element while keeping other page content and measurement consistent. Record what changed, when it went live, which page and audience it applied to, and any other factors that could affect results. This record helps you interpret later comparisons and avoid repeating tests without knowing what was tried.
Assess the result with care
Before starting, decide which metric will answer the question. Use the same conversion event and denominator as your baseline, and consider a relevant quality check as well. A higher number of form submissions may not be useful if the enquiries are incomplete or unsuitable for the business. Depending on the goal, review the next step in the customer journey, such as whether enquiries can be followed up or bookings are completed. Do not change the success measure after seeing the result simply because another metric looks more favourable.
If your site receives enough comparable traffic and you can reliably split visitors between versions, a controlled test can help distinguish the effect of a change from normal variation. Keep the test running through a representative period and allow time for visitors to complete the action. Avoid stopping as soon as one version appears ahead: small numbers and day-to-day fluctuations can produce misleading swings. Consider whether the period includes unusual events, such as a seasonal peak or a service interruption, before drawing conclusions.
Many small business websites do not have enough traffic for a dependable split test. In that case, make the change deliberately and compare results with a suitable earlier period, while noting that other factors may also explain any difference. Look for supporting evidence in the journey data and customer enquiries, not just a change in the overall rate. If the evidence remains mixed, keep the conclusion tentative, gather more observations or choose a clearer issue to investigate. A careful decision to wait is better than treating noise as proof that a page change worked.
Conclusion: Change with evidence
A conversion rate is only useful when you know which customer action it represents and which visits are included in the calculation. Before changing a page, check that analytics and conversion events are recording correctly, and compare like-for-like periods. Then look for meaningful differences between visitor groups and follow the steps people take before they complete—or leave—a form, booking or purchase.
Use those observations to identify one specific change that addresses a plausible obstacle, rather than redesigning a page based on a single fluctuation. For example, if a particular form step appears to lose visitors, inspect that step and make a focused adjustment. Keep a note of what you changed, when it went live and which measure should reflect the change. Allow enough comparable data to accumulate before judging the outcome, and account for other changes that may affect it. If the evidence is inconclusive, keep monitoring or refine the question instead of treating uncertainty as proof. This simple discipline makes page decisions easier to explain, review and improve over time.
Explore related information: Booking Website Design & Development; Web Design Sunshine Coast.


