A few points:

• You appear to have associated the email key with the name value and vice versa when creating your defref anonymous hash. I have changed them around in the code below
• You can populate the %inputs hash without the need for the $fname and $fvalue intermediate variables by using map
• Always put use strict; and use warnings at the top of your scripts to enforce disciplinr and catch errors
• The Data::Dumper module is worth it's weight in doughnuts when you are not sure what your data structures look like. I use it below to show what we started with and what has ended up in the hashes

Here's the code

use strict;
use warnings;

use Data::Dumper;

my @inputFields = (
   q{Requester Email:fredb@big.com},
   q{Project Manager Name:Fred Bloggs},
   );

my %inputs =
   map { split m{:} }
   @inputFields;

my $defRef = {
   requester_email => $inputs{ q{Requester Email} },
   requester_name  => $inputs{ q{Project Manager Name} },
   };

print Data::Dumper->Dumpxs(
   [ \ @inputFields, \ %inputs, $defRef ],
   [ qw{*inputFields *inputs defRef} ],
   );
[download]

and here's the output

@inputFields = (
                 'Requester Email:fredb@big.com',
                 'Project Manager Name:Fred Bloggs'
               );
%inputs = (
            'Requester Email' => 'fredb@big.com',
            'Project Manager Name' => 'Fred Bloggs'
          );
$defRef = {
            'requester_name' => 'Fred Bloggs',
            'requester_email' => 'fredb@big.com'
          };
[download]

I hope this is helpful.

Cheers,

JohnGG