With HR data, its all centered around an individual human. Create a way to identify everyone as a first order of business. This will save time for the data team, and for design and test folks too.
So the first step is to create an entity layer key.
Typically there are 4 types of “humans” in the system:
- Employees (current, former, part time, etc)
- Trusts and other entity-proxy records
Keying this will allow for unique identification of the entities for later data juggling – and – will also force (early in the process) clear definition of variance from this list of 4 profiles – allowing for more complete target environment configuration on plans and eligibility.
The second step is to clean some of the source data
Dust stuff off a bit in the source database. Take these basic steps and a lot of the new SaaS based HR systems will like you better:
- Complete all addresses; incomplete addressed, DOB, SSN.
- Identify where zip code based coverage will be tripped up by decentralized pool addresses.
- Identify election and effective date data hooks for all 4 type of humans.
- Surface any “relationship” data hooks and any required variance therin.
- Audit all free form text fields where a name is/can be entered and assess impacts.
- Newer systems understand “Domestic Partner” better than older systems, look at the relationship data footprint in your system and surface it as you will hit conversion issues specific to that “relationship” type.
Adopt a 1% rule and bind to an escalation plan
Make sure you’re not spending too much time on too few records. Spending 3+ days on 12 entity records must be done only after its decided that the downstream effect of manually handling those 12 records is a higher cost.
Hire or assign a notes minion
HR conversions are complex and involve many steps, including often many separate stakeholder groups. This poses a challenge for data teams during test because problems to one group are often invisible to another. If you take notes, and actively manage root cause analysis methods, you can save everyone a lot of headaches as the following three key issues spin in the universe concurrently:
- Ongoing business (changes, adds, terminates, etc)
- Plan and carrier conversions
- Approaching open election dates
I think Abe Lincoln is famous for saying “Give me six hours to chop down a tree and I will spend the first four sharpening the axe.”. Nowhere in our experience could that planning benefit a conversion team more than with HR data.