This is exactly what web config transforms were created for. The link you provided in your post has a walkthrough of doing this specifically for connection strings.
To start with the transforms, right-click your web.config file in the project explorer and choose "Add Config Transforms". Assuming that you have ConfigA and ConfigB in your solution configuration, there will be two new files added, Web.ConfigA.config and Web.ConfigB.config.
If you open these new files, they'll be pretty empty except for a bunch of comments. They actually contain a connection string example in them that you can use though - it looks like this:
<connectionStrings>
<add name="MyDB"
connectionString="Data Source=ReleaseSQLServer;Initial Catalog=MyReleaseDB;Integrated Security=True"
xdt:Transform="SetAttributes" xdt:Locator="Match(name)"/>
</connectionStrings>
Uncomment this section, and change the "name" property to the name of the connection string in the base web.config file. Set the "connectionString" property to the actual value that you want to use for ConfigA. So, like this:
<connectionStrings>
<add name="myConnectionString"
connectionString="Data Source=ConfigASqlServer;Initial Catalog=ConfigADatabase;Integrated Security=True"
xdt:Transform="SetAttributes" xdt:Locator="Match(name)"/>
</connectionStrings>
Repeat the process for the Web.ConfigB.config file, with the desired connection string for ConfigB.
Now when you use the Publish command in visual studio, it will automatically transform the base web.config file, and set the "connectionString" attribute to whatever configuration you're in when you publish.