Why Number of Sales Recorded in an Experiment is Smaller than Real Sales Number

That can be due to a multitude of factors as follows:

1. Your goal is not correctly setup and is missing some of the conversions
2. There are ways to buy (convert) without actually going through the experiment pages (example: you run the test on one product page but conversions done by users that buy other products and do not go to the tested product will not be recorded)
3. You run the experiment on an Audience which does not include all visitors. Visitors not matching the audience will not be included in the test; the conversions done by this visitors will not appear in the experiment report;
4. We count in the experiment report only one sale per visitor; if they purchase multiple times, we'll record only first time;
5. You allow users to pay via Paypal (or similar payment systems) which, after payment, would redirect users to your order complete page after XX seconds. If they close the window or navigate away before that amount of time passes, the conversion will not be recorded