Thread: Updated MaxMind API, Score vs. RiskScore

Greetings,

We are running 9.3 and just discovered MaxMind and wanted to use it to reduce some of our administrative costs in verifying authenticity of orders with separate shipping and billing addresses. So we initiated a trial with MaxMind and inserted our License Key into the appropriate appconfig, and waited for transactions.

As transactions started filtering in, I noticed that ASPDNSF is storing the deprecated "Score" and "Explanation" fields in the database which work on an outdated scoring system of 0-10, which MaxMind no longer uses as of version 1.3 of their API. Instead, they use the more accurate RiskScore field, which scores on 0-100.

Is this slated for an update at some point soon? I no doubt will have to update the code myself to get this to store properly in our database, because honestly, the system is unusable as is, but the feature is great, and we want to use MaxMind now that we've found it.

Thanks for any info

I wrote a custom report a while back that pulled the RiskScore - here's the SQL ;-

C#/VB.NET Code:

SELECT TOP 25 ordernumber 
              AS 'Order' 
              , 
              CONVERT(VARCHAR(8), orderdate, 3) 
              AS Date, 
              email, 
              lastipaddress, 
              billingzip 
              AS Billing, 
              shippingzip 
              AS Shipping, 
              ordertotal 
              AS 'Order Total', 
              avsresult 
              AS AVS, 
              last4, 
              maxmindfraudscore 
              AS FraudScore, 
              Substring(maxminddetails, ( 
              Charindex('<riskScore>', maxminddetails) 
              + 11 ), Charindex('<', Substring(maxminddetails, 
                                     ( 
                                               Charindex( 
                                               '<riskScore>', 
                                               maxminddetails 
                                               ) 
                                       + 11 ), 6)) - 1) AS RiskScore 
FROM   orders 
WHERE  ( ordertotal > 88 
         AND ( billingzip != shippingzip 
                OR avsresult = 'I' 
                OR avsresult = 'N' ) 
         AND maxmindfraudscore > 0 ) 
        OR maxmindfraudscore > 9 
ORDER  BY ordernumber DESC 


This is the bit that gets the risk score :-

C#/VB.NET Code:

Substring(maxminddetails, (  
              Charindex('<riskScore>', maxminddetails)  
              + 11 ), Charindex('<', Substring(maxminddetails,  
                                     (  
                                               Charindex(  
                                               '<riskScore>',  
                                               maxminddetails  
                                               )  
                                       + 11 ), 6)) - 1) AS  RiskScore 


TTFN

BFG

We just started using it a few weeks ago ourselves and it has been useful in pointing out some potential fraudulent orders. We ran it on some old orders that we received chargebacks on and those orders scored high as well, so it seems useful, but we have seen some high false-positives as well. I wish ASPDNSF would update it to show better support for the latest API, but you can still see the RiskScore and other details - just select the "Billing" tab and scroll through the MaxMind details.

I wanted to update this thread with the solution I've found. It was pretty simple to edit the gateway.cs source file to capture the RiskScore instead of the Score for the database, so I did that. This makes it easier for our customer service personnel who may not be familiar with reading the XML that displays in the Billing tab

Received the following:

Beginning January 1, 2014, the 'score' output, which is present in minFraud versions 1.2 and earlier, will no longer be supported by MaxMind. If you haven't already, we highly recommend transitioning to using the 'riskScore', which gives you access to much more accurate fraud scoring, both in terms of increasing the amount of fraud detected and in reducing the amount of good orders flagged as fraud. Please see this blog post for more information about the differences between 'score' and 'riskScore'.

Anyone have info on when ASPDNSF will be upgrading to v1.5 API? I've submitted a ticket, but rather than wait, I thought I'd post here to see if anyone has info to share. :-)