Should I Measure Conversions by User or by Session?
Choose user or session conversion rate based on the question you need to answer. Compare repeat visits, GA4 metrics, reporting windows and experiment assignment.
Use session conversion rate when you want to know how often a visit contains your chosen outcome. Use user conversion rate when you want to know what share of distinct measured users completed it during a defined period. You can report both, but choose one primary measure for the decision you are making.
A user can return several times before buying or booking. Those repeat visits explain why the two rates can differ without either being wrong.
If Niwa answers a question during an early visit and the customer returns later to complete a website reservation, session and user reports describe different parts of that journey. Neither report alone proves which interaction caused the booking.
The session-based question
A session-based rate asks: "How often does an eligible session contain the conversion?"
The formula is converting sessions ÷ eligible sessions × 100. A session containing several instances of the action still counts once in the numerator of that rate.
This is useful when evaluating a landing page or the immediate ability to complete a visit. Our introduction to CRO explains how to choose the useful outcome before evaluating a website change. For example, a service page may receive 2,000 eligible sessions, of which 80 include a completed reservation. The booking session conversion rate is 4%.
It is not the same as reservations per session if a session can contain multiple reservations. Keep the label and calculation aligned.
The user-based question
A user-based rate asks: "What share of distinct measured users completed the conversion during this reporting period?"
The formula is converting users ÷ eligible measured users × 100. Each measured user appears once in the numerator, even if they purchase repeatedly during the period.
This can be useful for a journey involving several visits. In an illustrative example, 100 measured users make 160 sessions. Ten users convert in 12 sessions. The user rate is 10%; the session rate is 7.5%.
The difference reflects the counting units. You cannot conclude that the website is performing better simply by choosing the larger percentage.
What "user" actually means in analytics
A measured user is not necessarily a verified individual person. Browser changes, device changes, identification settings and measurement availability affect how activity is connected.
Google Analytics offers user and session key event rate metrics. When reproducing a rate manually, use the user population specified by that metric and report. Do not casually replace total users with active users or mix definitions from different tools.
Keep the event specific too. A rate for any key event may include other actions beyond the purchase or reservation you intend to evaluate.
The reporting window changes the answer
If someone visits on Friday and buys the following Tuesday, a Friday-only user report will not show that later purchase. A longer report may include it, but it also includes people first measured later in the period.
A calendar-period user rate is therefore not automatically a first-visit cohort conversion rate. A cohort analysis follows users who entered during a specified period and gives them a consistent opportunity to convert afterward.
Use a cohort when the question is whether newly acquired visitors eventually become customers. Define the follow-up window, such as 14 days, based on the actual decision cycle. Fourteen days is an example, not a recommended universal rule.
Keep experiment assignment consistent
Reporting choice and experiment assignment are separate decisions. If a test assigns variants persistently to users, repeated sessions from the same assigned user are related observations. Treating them as independent visitors can make the sample appear more informative than it is.
Use the experiment platform's analysis method and the outcome defined in the test plan. Do not choose the user or session rate after looking at the results simply because one produces a favourable answer.
For ordinary reporting, keep both rates visible when repeat visits matter. Choose the primary one in advance and explain why it answers the business question.
Applying this to a Niwa-assisted booking journey
Prepare accurate service descriptions and availability, then configure Niwa's booking workflow. Niwa uses your website's business context to answer questions, offer available times and complete reservations on the website or directly in a WhatsApp conversation.
For a reservation completed during a measured website visit, the session outcome tells you whether that visit contained the booking. A user or cohort view can describe a longer website journey. Bookings completed entirely in WhatsApp belong in a channel-specific report unless a reliable measurement integration links them to an eligible website user. Keep service questions, time selection and saved bookings separate.
Connect booking records to your measurement system and define identity and attribution rules before combining channels.
Contact Niwa with your WordPress site and the task you want customers to finish. Decide how you will count that outcome before using a rate to judge the customer assistance.
