There’s
a big trend in the industry today around control convergence. In a
nutshell, control convergence is an attempt to reduce the number of
controls in a system, while still keeping things effectively under
control. If you’ve been following me so far, you have a lot of tools
for modeling the different kinds of controls that you might find in
your company. So in this article, I wanted to walk you through an
example that will not only help cement in some of the concepts
discussed previously, but also give you a solid foundation for properly
supporting a control convergence effort, from a data systems standpoint.
Meet XYZ Healthcare Company
In our example the global finance manager at XYZ Healthcare Company
needs to control their accounts receivable process. Strategically, they
need to reduce the accounts receivable collection period from 95 days
to less than 45 days. This affects the collection process, in which
they can receive payments by either check or credit card.
The old process was to just send an invoice, and wait for the customer
to pay sending standard letters at the 30, 60, and 90 day late marks.
After some brainstorming, XYZ Healthcare revised its collection policy
to look something like this.
- Once the payment term has expired, collectors shall call
the customer on a regular basis until payment arrangements have been
made.
- For customers who wish to pay by credit card, payments shall be processed immediately over the phone.
- For
customers who wish to pay by check, sales reps shall make a customer
site visit within 2 days of the agreed payment arrangement, to collect
the check.
Organic Compliance Growth
Shortly thereafter, PCI compliance became a concern, so to control the
risk that sensitive credit card information will fall into the wrong
hands they added the following policy points:
- No credit card data will be recorded in the computer
systems by the collectors. The collector records all necessary payment
information on a special form, which is then passed to a credit card
terminal operator for processing.
- Once credit card
payments are confirmed in the credit card terminal, credit card forms
shall be passed to a supervisor, who shall shred the forms.
Then, the HIPPA ( privacy ) police came along, and noticed that to
validate the customer on the phone the collectors were using customer
files which contained sensitive data that could be mishandled. So, to
control the risk that personal customer data (i.e. social security
numbers) will fall into the wrong hands, the following policy point was
added:
- At the beginning of each day, the collector shall receive
a list of all the customers they will be collecting on, with sensitive
customer data for validation. At the end of each day, the collector
shall surrender the collection list to the supervisor, who shall shred
the collection sheet.
Finally, in the last SOX audit this process was tagged again for
control remediation. So, to control the risk that inexperienced
collectors will record inaccurate receipts, the following policy points
were added:
- The credit card receipts recorded by the collectors shall be reconciled to the amounts processed by the credit card processor.
- The
check receipts recorded by the collectors shall be reconciled to the
checks collected by the sales reps and further reconciled to the
amounts deposited by the supervisor.
And, to control the risk that fraud will occur with checks that are received, the following policy point was added:
- A separation of duties shall be enforced, such that the
person collecting the checks is not the same person depositing the
checks.
Control Convergence Team to the Rescue
At this point XYZ Healthcare has installed 6 different controls on the
same process to cover 4 different compliance concerns ( of course, this
is an simplified example – in the real world there could be hundreds of
controls around this process at this point). This is okay, but the
ongoing cost of compliance is directly related to the number of
controls that need to be tested, so they chartered a project team to
converge controls, and you’ve been recruited to support the effort from
a data systems standpoint.
Converging Controls: Step 1 – Build a Testing Harness
The team will be anxious at this point to jump straight into building a
solution. It’s a natural human tendency. You may even see some areas
where you could immediately improve their situation by leveraging
technology. You and your team must resist this urge, and first fortify
your process with a testing system.
Determine the metrics that are important to this process; both from a
strategic and compliance perspective. Here what I came up with:
- Collection Period –after all, this is the strategic metric that we’re trying to reduce
- Collection Forms Not Shredded – PCI control violation
- Customer Collection Lists Not Shredded – HIPPA control violation
- Failed Reconciliations – SOX control violation
- Separation of Duties Violations – SOX control violation
- Number of Controls – Your team is making progress if this number is going down
First and foremost, you need to build a data system that will collect
and store this data. Then, baseline your current process to see where
you are today. If you don’t baseline today, you won’t know if you’re
improving anything.
Converging Controls: Step 2 – Make Sure You Can Undo
You need to build a safety net before the innovation process happens.
If your “improvement” effort actually causes you to go backwards (
based on your collected metrics ), then you need to make sure you can
at least get back to where you were before. From a data systems point
of view, think seriously about such issues as a version control system,
and a configuration management system. Do a brainstorm – if you had to
get back to the old version, how would you do it? From a policy and
process point of view, help your business users by building a
repository where this can all be stored, with strong change data
capture architecture in place.
Converging Controls: Step 3 – Make an Improvement
Now, and only now, start thinking about ways to converge the controls. Here are the ideas I came up with:
- Create a combined credit card / personal information form for processing which can be shredded all at once.
- Automate
the reconciliation pieces. This not only reduces the number of controls
to test ( two reconciliations can be tested as one reconciliation
process ), but it increases reliability and reduces personnel workload.
- Install
a policy that only supervisors can deposit checks. Since supervisors
cannot be sales reps and vice versa, there is a built in separation of
duties control. Now create a daily supervisor assessment form which
records amounts deposited, and forms shredded. Between the automated
reconciliation and this daily practice, only one control test should be
needed that ensures that the assessment process is being executed
properly.
Converging Controls: Step 4 – Test the Improvement
Now before you declare victory, gather your metrics again using the
testing harness you created in step 1. We know for sure that control
points have been reduced from 6 to 3. That’s a pretty good start! But,
what happened to the other metrics? Did any control violations
increase? If so, you need to undo ( Step 2 ), and try again ( Step 3 ).
And what about the collection period? Did it actually increase as a
result of your “improvement?” If so, back to the drawing board!
Let’s hope however that your efforts were not in vain.
Even if everything else stayed the same, the fact that your controls
are cut it half is a reason to declare victory. And if you’re seeing
even lower control violations and a shorter collection period – pop
open the champagne, you did real good!