In the world of risk and fraud, innovation is happening fast. Ignoring new technology innovations is simply not an option for financial services organizations looking to scale and grow their business. This session will cover how organizations can find new revenue in the underwriting department leveraging their risk scorecard. Learn about the importance of data capture and analysis in the underwriting process, in order to increase the size of your customer base.
Gaming the Odds and Gaining Competitive Advantage with Automated Onboarding
1. Gaming the Odds and Gaining
Competitive Advantage with
Automated Onboarding
Presented by Recombo CEO, Mike Gardner
on Wednesday, March 11th 2015 at:
2. So you probably think that I’m going to tell you
how much faster you could underwrite, and how
much money you could save, if you automated
your underwriting?
Well I could…
But, my guess it that you
already know that’s something
you should be doing.
3. What you will learn today:
1. How your risk department can use basic
statistics concepts to unlock revenue
2. How a risk score is developed and how it is
often determined improperly
3. What you need to do get started for your
organization
4. Las Vegas is the best city in the world to teach us
that there is MONEY tucked inside of RISK.
5. There’s a lot that needs to be done to unlock
the value of merchants…
Pass/fail reviews only provide “buckets” for sorting merchants
6. A histogram of your merchants can provide a
a view of your merchant profile.
7. So if this a chart (histogram) of all the merchants
that submit applications to your firm, can you
see the money you’re leaving behind?
10. Do Your Own Math
Assume the following:
• Average monthly value of a merchant $2500
• Average merchant applications per month 100
• 80% of merchants get approved
The cost of false positives in your risk model:
5%: $30,000/annum
15%: $90,000/annum
25%: $150,000/annum
11. Consider the following:
How much time, if
any, is your
organization spending
reviewing the
merchants you
declined, rather than
looking for new
merchants?
12. Hang on though, Mike! I don’t want to be taking
on more risk for my organization!
13. Risk isn’t a point – it’s a range.
High Risk
Low Risk
Your Threshold of Risk
0 100
Your Risk Score
14. What you score, is as important as how you
score
Reputation
45pts
Delivery
Method68pts
Value
20Pts
Identity75pts
1. Identity
2. Validity
3. Legitimacy
4. Financial Acceptability
5. Risk Level
High
Risk
Low
Risk
Your Threshold of Risk
0 100
Your Risk Score
15. Remember that
underwriting is like a
court room; your
merchants are presumed
innocent, and your
underwriting team is a
prosecuting attorney
16. As the case stacks against merchants, the less
desirable they become as clients
Reputation
45pts
Delivery
Method68pts
Value
20Pts
Reputation
45pts
Delivery
Method 68 pts
Value
20 Pts
Identity75pts
Sorry,
too risky
High
Risk
Low
Risk
Your Threshold of Risk
0 100
Your Risk Score
17. But when we use lots and lots of data sources in
our underwriting, we increase the probability
that our data potentially “overlaps.”
This is a
“specification
error” and I can
almost guarantee
that if you’re
scoring, you’re
over specifying.
18. Imagine our fictitious merchant is the big yellow
circle below, and we want to use just three
variables to describe his risk:
• Reputation
• Chargebacks
• BBB rating
Rep
Chargebacks
BBB
All these little areas
of overlap are the
problem
19. If we COULD calculate that overlap area (which
we can’t), and we took out that double
counting, we’d find our merchant is actually
within our threshold of risk – but we turned him
down!
Reputation
45pts
Delivery
Method 68 pts
Value
20 Pts
Hey, I’m not
risky; I’m a
false positive.
20. So now, when you look at your merchants, you
know the money is sitting right on the edge – it’s
the “good” people you’ve said no too
21. Since our risk models are likely “over specified,”
our first, and possibly most valuable place to
look for opportunity, is with the people you’ve
turned away.
Rep
Chargebacks
BBB
22. So what do I you want you to do?
Score everyone – At minimum, you
should be automating your underwriting process so that
you’ve got good data.
23. What else do I
want you to do?
• Keep every score you’ve calculated. You need
this data to understand the distribution of your
applicants
• Regularly review the “marginal no applicants”.
These are opportunity
24. And finally…
Monitor your data sources to
make sure they aren’t “getting
smarter” at your expense. As
your data providers widen the
number of variables they are
considering, they are increasing
your probability of “overlap.”
You need to keep adjusting your
risk models and weightings
accordingly.
25. The ETA recommends that companies maintain
an “agile approach to underwriting…constantly
reflecting improvement”.
Remember this:
Constantly reflecting
improvement means it is
TURNING DOWN AS FEW
merchants as possible, NOT
checking every imaginable
item to reduce your risk to
zero.
26. To learn more about Rapid Customer
Onboarding, and how it impacts and shapes
underwriting in the payments market, come visit
us at www.recombo.com, here at MAC, or
upcoming at Transact15 in San Francisco.