Data Migration

Many approaches exist for tailoring data conversion tasks to meet the individual needs of a company and a project, but a risk management approach can provide useful insights into both planning and resourcing conversion tasks.

Data conversion can be so infrequent it can skew the organizational risk management thinking and not get the attention it deserves. When traditional risk management approaches – i.e. risk avoidance, risk assumption, risk transference and risk mitigation – are contemplated, it seems at first glance risk transference may be the best approach for conversion. This is due to the high impact and medium likelihood of a failure and often the low internal IT knowledge base or experience in this area. Let’s examine this problem and determine if this thinking is accurate, recognizing that this issue is complex and approaches are unique to each carrier.

Manual Data Conversion

Using a policy data conversion example, here’s the math:• Policies in force = 120,000.• Working days/month = 20.

Assume an equal distribution of policies monthly. The calculations show 6,000 policies per month or 300 policies per day need to be processed. If one person can process 20 policies per day, then 15 people are required for one year to do the full conversion manually.

If one full-time employee costs $75,000 per year, then the cost of conversion is 15 x $75,000 = $1.125 million, or a little under $100,000 per month. This includes all of the detailed scrutiny per individual policy, updating and actions taken on all policies and exceptions and assumes an ‘on-renewal’ approach to conversion.

Benefits to this approach include:• re-underwritten files;• standardized information sets;• more predictable and visible timeline; and• no costs associated with a software development lifecycle.

Risks to this approach include: • finding, deploying and managing a group of knowledgeable staff (in this example, 15 people), either pulling them from normal day-to-day activity or hiring and training them;• increasing policy counts and/or adding historical transactions quickly becomes prohibitive (any difficulties doing these things with 15 staff become exponentially more complex with 150 or 250 staff); • manual data entry is error-prone; and• two systems are operated, maintained and used during the migration, decreasing productivity while at the same time increasing complexity, costs and risk.

Automation Through Third-party Software

Under this scenario, a vendor provides or uses data migration software to fully automate data conversion, implemented by a combination of both external and internal resources.

The cost of requirements, development and testing, using ‘expert’ resources billing at hourly rates, can easily add up. Using our previous example, let’s say the vendor supplies 15 people who bill $1,000 per day for 20 working days in a month. That’s a monthly bill of $300,000 — more than triple the in-house cost. Of course, a valid counter-argument is that 15 ‘expert’ resources are not required; only a sub-set of this number is needed.

No matter how many external resources you have, you still need in-house people who understand your existing data, format and technology (a.k.a. ‘the target’), as well as requirements for your new data format and technology (a.k.a. ‘the destination’).

A typical issue in conversion is that the data is populated correctly into the destination, but then it doesn’t react or transact in the same way as data that was populated through the sequential manual data entry process that originally created the data. The need for in-house resources to complete testing is absolute for a successful conversion. These are the internal ‘experts,’ and are best considered that way. Internal experts will still likely be the ones to identify and resolve exceptions to the automated process.

Benefits to this approach include: • auditable, accurate processes produce an accurate result;• mid-project changes can be managed and dealt with in-flight, and applied for all policies;• experience from “other carriers” can be applied in the form of vendor resources, and in proven software; and• internal IT resources can be deployed on other projects.

Risks to this approach include: • finding a proven product and methodology to integrate with internal processes and that can deliver anticipated results on time and on budget;• exceptions and unique situations must be dealt with manually;• costs can escalate without proper project management to ensure visibility; and• software vendor staff must understand insurance issues and work closely with internal team.

Hybrid Approach

The ‘hybrid’ approach is to automate the majority of the solution, but not all of it. This can be done through analyzing what data can be populated ‘easily,’ via automation, and then allowing the ‘difficult’ and/or high-risk data automation to be completed manually. This approach balances costs, risks and timeliness of the partly automated solution with the resource costs and risks of the manual approach. This hybrid approach accommodates the typical reality that there will always be exceptions to successful conversions. These exceptions can be accounted for in a plan, so the conversion plan is exhaustive for all cases.

The number of exceptions in the first ‘round’ of converted data is generally the largest, so the plan to handle them manually or add them to the automated solution for the remaining months is an available alternative. Analyzing policy distribution across the months can also result in a planned schedule, producing a productivity lift as the conversion proceeds. For example, a partially automated solution converting the first three months of data, representing 25% of the conversion, can be timed to begin so it represents the smallest percentage of policies in force.

Analysis of the exceptions, errors and issues can be used to modify the conversion process and address the largest number of exception cases and/or the highest-risk cases. By the time the largest volume of policies is scheduled for conversion, the majority of exception cases have been automated and/or addressed.

In addition to the benefits and risks of each approach, other benefits include: • the speed, repeatability and auditability of the automated approach is being combined with a lower-risk, more predictable manual approach; and• incremental results will motivate the team, and ensure accuracy.

Risks to this approach include: • it still uses internal experts and third-party resources, which can increase costs;• problems inherent in both manual and automated approaches are represented in the hybrid approach; and• if third-party resources are not knowledgeable about insurance issues, problems will persist.