This is a fairly standard type of request.
You will need a database table, where there is a row for each customer that gets inserted. This row can include the data purchased and their email, along with any other information you have on that that would be pertinent historically.
This same row can store the number of views, starting with 0.
The script that renders the video will lookup the correct row in the database using a url like this:
Where id will be a hash that is stored and indexed in each customer row. I would recommend an md5 or sha1 hash based on some random data. The insert script should take care of this, as well as storing the user and probably sending out the email.
When the script runs, it will verify the user, render out the player and stream the video. It's important that the video stream is also delivered programmatically via a script, so that people can not simply pull it off the server. In other words, it needs to be somewhere on the server that is not web accessible and should only be delivered as a source to the player if the access checks were added and passed. This 2nd delivery script should increment the number of views.
3 is a really small number, and I'd recommend you reconsider the delicate balance between pleasing your customers and being overly concerned about people who might want to take advantage. Is it worth upsetting one A quality customer who has a wonky browser and internet connection, just so you can stop a few people from watching a video associated with your product? I'd recommend being a lot more generous with views (10 at minimum) before you cut someone off, and just so you know, there will always be people who download your video.
I have to point this out, because these players are just code and the mechanics of delivering a video are complicated, and easily reverse engineered by people who know what they are doing.
I'd recommend using the longtail video player as it makes hooking up to the events system in Google analytics, which will allow you to track the percentage of video watched. See: http://www.longtailv...ogle-analytics/
Hope this helps you get started.