Conversation sessions: the sessions that started a conversation with Olvio during the date range, each session counted once. Each session is identified by an anonymous session ID, which the widget assigns when a shopper sends their first message (reusing one from an earlier conversation when it can) and adds to their cart once Olvio replies. On Growth and Scale, chats opened from Olvio Advisor links count too.

Attributed order: an online store order placed during the date range by a session with a conversation with Olvio that started within 7 days before or after the order. The 7-day window is fixed and has no setting.

Orders during the period: every online store order Olvio recorded during the date range, whether or not the buyer saw the widget or chatted. Orders from other sales channels, such as point of sale or draft orders, aren't included.

Impression sessions: product-page loads where the Olvio widget appears during the date range, whether or not the shopper scrolls to it. Most are recorded without identifying the shopper, even for shoppers who chatted earlier, so each load usually counts separately: a shopper who opens three product pages counts three times. Rapid reloads of the same page are generally counted once.

Shoppers who chat choose to. They may already be closer to buying than the average visitor. The two rates also don't count the same thing: the Olvio rate counts sessions, while the baseline divides orders by product-page views. So a higher Olvio rate doesn't, on its own, show that shoppers who chat buy more often or that Olvio caused those sales.

The baseline mixes two groups. Your conversion rate divides all online store orders by product-page widget views. It rises when many orders come from shoppers who never saw the widget, and falls when shoppers view many product pages before buying.

Short date ranges swing. An order counts on the day it's placed, but the conversation behind it can have started up to 7 days earlier, outside the range. On Today or Yesterday, the Olvio rate can look very high, even above 100%, or very low when recent chats haven't turned into orders yet. Longer ranges give steadier numbers, as long as they stay within the period Olvio keeps your data (see What isn't counted below).

Compare with Previous period (the same number of days just before your range) or Previous year (the same dates one year earlier), or choose No comparison. A Previous year comparison looks at dates a year or more back, past the 90 days Olvio keeps widget views and orders, so for these two tiles the comparison shows 0% and usually no badge. The same can happen with Previous period when the comparison period reaches further back than Olvio keeps the data. On Starter, for example, the default comparison (the 30 days before the last 30 days) has almost no conversations left, so Conversion rate with Olvio can show no badge or a very large change.

Rates show two decimals, for example 3.25%.

The ▲ or ▼ badge is the relative change from the comparison period, not a difference in percentage points. A rate that goes from 2.00% to 3.00% shows ▲ 50.0%. The badge is hidden with No comparison, and when the comparison period's rate is 0 but the current one isn't.

Each tile's chart plots the daily rate and, when a comparison is on, a dashed line for the comparison period. On a one-day range such as Today, the chart shows dots instead of lines.

Days start and end in your Shopify store's time zone.

Time without a plan. Olvio records nothing before you choose a plan or after your subscription ends. Orders placed during your free trial are recorded.

Changes after an order is placed. Orders are counted when they're created. Later refunds, cancellations or edits don't change the count.

Deleted data. Olvio deletes older data automatically. Widget views and orders are deleted after 90 days on every plan, so both tiles show 0% for days more than 90 days back, even when you choose Last 365 days. An order also stops counting toward the Olvio rate once its conversation is deleted at the end of your plan's conversation history (30 days on Starter, 90 on Growth, 365 on Scale). Each rate is recalculated from the data still stored, so the rate for a long date range can be lower or higher than it really was.