Basic data statistics-new, active, retained

These three data basically reflect the operation of the product

New general reaction a product market expansion ability, is flowing water influx.

Active is the main source of income

Retention generally refers to the addition of new retained (also active), the number of new users of the first n days of active, after 1st retained, after 2nd retained, N later retained. It's a reflection of loyalty. Also reflect the number of new users is irrational increase (similar to the first day of the stock market irrational Chase high Buy, the next day the selling gap---fill)

The user has two representations: device and account.

Statistics are based on a latitude, which can be: Overall, channel, version, and so on.

So there are total additions, new channels, new versions, and so on, active and retained.

If a product has more than one sub-product, latitude and more complex, such as sub-product 1 Channel 1 new and so on.

In the case of table design, the temporary table only needs [first use date] and "Last Use Date" 2 fields and a latitude field to calculate the new, active, and retained at a certain latitude. where [last use date] in the temporary table is upsert (update if exist else insert). [First Use date]= today, is today's new; [last Use Date] = Today is today's day work; [first use date]=20150827 and [last use Date] =20 150828, so that 20150827 of the next day retention,20150827 of the next day retention rate =20150827 of the next day retention/20150827 of the new. Because [last use date] is constantly updated, be aware of the timing of active and retained computing. The temporary table is full, and the dimensions are indexed

Basic data statistics-new, active, retained