Stored procedure to insert into multiple tables in sql server 2012, using id col from one table to insert into the other 2

Question

Hi all,
Apologies if any of the following sounds at all silly but I am fairly new to this so here goes...
I have 3 tables that require data insertion at the same time. The first table is the customers table, I then want to take the automatically generated custid from that table and inser it into 2 other tables along with some other data
Here's what I have so far which does not work:
CREATE PROCEDURE CustomerDetails.bnc_insNewRegistration @CustId int,
@CompanyName varchar(100),
@FirstName varchar(50),
@LastName varchar(50),
@Email nvarchar(254),
@HouseStreet varchar(100),
@Town smallint,
@County tinyint,
@Postcode char(8),
@Password nvarchar(20)
AS
BEGIN
begin tran
insert into CustomerDetails.Customers
(CompanyName, FirstName, LastName, EmailAddress)
Values (@CompanyName, @FirstName, @LastName, @Email)
set @CustId = (select CustId from inserted)
insert into CustomerDetails.Address
(CustomerId, HouseNoAndStreet, Town, County, PostCode)
values (@CustId, @HouseStreet, @Town, @County, @Postcode)
insert into CustomerDetails.MembershipDetails
(CustomerId, UserName, Password)
values (@CustId, @Email, @Password)
commit tran
END
GO
If anyone could help with this I would very much appreciate it as I am currently building an online store, if there's no registration there's no customers.
So to whom ever is able to help, I thank you whole heartedly :)

Answer 1

I hope by now it is apparent that statements like "doesn't work" are not particularly helpful. The prior posts have already identified your first problem.  But there are others.  First, you have declared @CustID as an argument for your
procedure - but it is obvious that you do not expect a useful value to be supplied when the procedure is executed.  Perhaps it should be declared as an output argument so that the caller of the procedure can know the PK value of the newly inserted customer
- otherwise, replace it with a local variable since it serves no purpose as an input argument.
Next, you are storing email twice.  Duplication of data contradicts relational theory and will only cause future problems. 
Next, I get the sense that your "customer" can be a person or a company.  You may find that using the same table for both is not the best approach.  I hope you have constraints to prevent a company from having a first and last name (and
vice versa).
Next, your error checking is inadequate.  We can only hope that you have the appropriate constraints to prevent duplicates.  You should expect failures to occur, from basic data errors (duplicates, null values, inconsistent values) to system issues
(out of space).  I'll leave you with Erland's discussion for more detail:
erland - error handling.
Lastly, you should reconsider the datatypes you are using for the various bits of information.  Presumably town and county are foreign keys to related tables, which is why they are numeric.  Be careful you don't paint yourself into a corner with
such small datatypes.  One can also debate the wisdom of using a separate tables for Town and County (and perhaps the decision to limit yourself to a particular geographic area with a particular civic hierarchy). Password seems a little short to me. 
And if you are going to use nvarchar for some strings, you might as well use it for everything - especially names.  Also, everyone should be security conscious by now - passwords should be encrypted at the very least.
And one last comment - you really should allow 2 address lines. Yes, two separate ones and not just one much larger one.