A DESCRIPTION OF RECENT PROBLEMS ASSOCIATED WITH PayPal AND WHAT WE'RE DOING ABOUT THEM
Some of you have recently experienced delays in receiving the
RBR and A3P reports you have ordered online on our website.
Normally, reports are emailed to customers within 2 minutes of
receipt. The way we accomplish this is to have a computer
program running around the clock that checks a dedicated email
folder every 60 seconds for new incoming orders. Upon
recognition of such an email, the program quickly determines
the customer's email address, the report(s) ordered and validates
the transaction-ID provided by PayPal in the emails.
Recent technical problems experienced on PayPal's end have
caused these emails to go out without identifying the items
ordered. They basically just said "you received $5.00 from
customer X with transaction-ID of 0ZS11832Q5186544T".
No mention that customer X wanted the RBR for Wednesday's Aqueduct card.
In the 6 years we've been dealing with PayPal, problems
like this have been rare and overall we are very satisfied with the service they provide.
When this occured during normal business hours, we noticed it
immediately and manually emailed the RBR to the customer. But
in the late night and early morning hours when our system was
running unattended, it was unable to fill the orders that were
missing the name of the item ordered.
We immediately contacted PayPal and brought the issue to their
attention. While waiting for them to come up with a solution,
we started working on our own solution. In short, we created a"backup"
procedure that would be able to fill orders without
having to rely upon the notification emails that are normally
sent to us by PayPal seconds after verifying a transaction. For
those of you with PayPal accounts, you are familiar with the
various features you can use. You basically go to their website,
enter your user name and password, and you're in and able to
view your recent transactions and perform tasks like depositing
or withdrawing funds or creating reports.
As users of our software know, recent versions of our programs
interface with the internet in many ways. 9.0 allows users with
access to quickly download our A3P Past Performance Files. A+
Basic also has this feature for the "compact version" of A3P files.
We control access with a "key record" that identifies users and the
date their access runs through. For our purpose, this is a very
simple chore as no fancy security is imposed. But users do need our
software in order to use our files. In a more secure environment
like PayPal where sensitive data is stored, they need to prevent people
from accessing user information. They do this with passwords.
So basically, our task was to write a procedure that simulated
every keystroke necessary to get to the PayPal web page that
contained all recent transactions. This was a little tricky,
but yesterday we ran our first successful test and simulated
orders were processed with our backup system. Here's how our
system works. At various prescribed times, the program starts
automatically, opens a web browser window, "types" in the URL
for PayPal's website and "presses" Enter. It then "types" in
our user name and password and "presses" Enter. Those actions
take us to a summary page containing recent transactions. But
none of the information on that page identifies the 3 pieces of
data necessary to fill orders. What is important on that page
is the date of each transaction and a link that takes us to
another page identifying the details of that particular
transaction. By examining the date, we can easily determine
which transactions are recent. Those are the only ones we are
concerned with. After converting the HTML of the summary page
to plain text, our program is able to determine the URL of each
pertinent transaction. At that point, the program opens a URL
sub-window by "pressing" Ctrl-O and "types" in the first URL. At
that point, the window containing the details of the transaction
opens. After converting to plain text, our program can easily
pick out the buyer's email address, the transaction-ID and the
report(s) ordered. At this point, we have everything we need to
fill an order.
Going forward, we plan on running both systems simultaneously
and need to determine the intervals at which the backup system needs to check.
Our usual program checks every 60 seconds and runs 24x7. Initially, the backup
system will check 3 or 4 times each hour during times when the normal system runs in unattended mode.
What this means to you is that until PayPal fixes the bug, you will receive your "after hours" orders within 15 or 20
minutes instead of the normal 1 to 2 minutes. During "normal"
business hours when our system is running in attended mode, your
orders will be processed immediately either through our
primary system or manually if PayPal issues persist.
Since we store records of each transaction, both systems will "know"
which orders have already been filled so duplication of orders should not occur.
So when PayPal implements a fix to
their problem, orders should be filled in the "normal" manner.
Until then, the backup program will be capable of filling orders
using the procedure described above.
We learned long ago that pointing fingers and placing blame
never solves problems. We also learned that fixing problems
associated with external entities by coding "workarounds" is
the best (and fastest) way to provide service to our valued
customers. We've done it in the past when third party vendors
changed Past Performance Files without notice, when website
procedures changed without notice and when access to files was
simply withdrawn from the public.
We've been in this business since 1991 and our customers know
that when problems arise, we take ownership of them regardless
of who is at fault. And when reports are emailed late, we
refund the purchase price in full immediately. Customer
satisfaction is JOB ONE with us.