Systems Integration

Our Managing Director, Gerard, recently presented to the HR Technology conference at the Hilton in Sydney. Here is a subset of the information provided that you may find useful.

We, at Silverdrop, are often involved with integration projects between Workday, chris, Elmo, PageUp, Mumba Cloud, AD, IDAM, financial solutions and Successfactors.

Existing setup

In a car ride to the beach, what is more important, the road or the car?

It is extremely rare to find any organisations that have a single HRIS platform that does everything. As such, most organisations have multiple HR systems and require integration.

When buying systems, nearly all organisations prioritise the functionality within the HR applications themselves, with little thought about how it will integrate with either what the organisation has today or what it plans to have in 5 year’s time.

This is akin to buying a car without having a road to drive it on.

The integration between systems is vital to the effectiveness of the systems themselves in driving productivity and improving the work day for managers and staff.

Systems need to talk to each other well and you need to determine the roles of each system in the same manner you do for staff. Many systems perform multiple roles and often there is a large cross over where multiple systems in your organisation can perform the same task. An example of this is viewing leave balances, applying for leave, changing a staff members position and changing a staff members address.

In a typical organisation that we would deal with, they have a payroll system, a T&A system, a reporting tool, a recruitment/onboarding system, Employee Self Service, an LMS, email and active directory. In most of these the starting point has been to have the payroll system defined as the source of truth or authoritative source of information for most staff data.

At this point, the master system is the payroll one and the slaves that are fed with information are the other, smaller systems.

With multiple systems, we are seeing that:

• • Staff need to access multiple applications
  • Executives receive reports from each system- they look different
  • HR people don’t have the specialist tools that they need for their specialist roles
  • 5 systems
  • 5 reporting tools
  • 5 staff portals
  • Data is randomly accurate

The aim should be to provide HR with specialist tools to do their job, provide staff with a single portal to view and change data, have a single analytics tool that aligns terminology and fonts and data should be shared and accurate.

Organisations can achieve this now. We are seeing that the systems exist for this to occur and some organisations are close to having this operating across their business.

How do you get there?

The first thing that I would say is that “YOU CAN’T CHEAT”.

By this I mean that, to complete the project well, you will need to complete all the tasks and spend the right amount of time and focus on these. If not, the integration project and results will not be successful.

We have worked on a lot of these projects recently and this is what we are seeing:

The tasks required before go-live include:

• • Plan
  • Build in Test environment
  • Unit test
  • SIT
  • UAT
  • Go-live build and
  • Go-live

Involved with each of these stages is your Project team, who does most of the effort. At this stage, when things go wrong, the group is not exposed to the whole business. Setup can be corrected, and the impact is contained in terms of the business, reputation and cost.

Sometimes organisations become fixated on achievement of the project within a very tight timeline.

We have seen situations where organisations try to “cheat” by missing steps. This may ensure the project goes live on-time but it creates collateral damage.

In this scenario, the project has skipped or rushed the testing, or they have dropped key functionality, that was very beneficial to the organisation in the rush to meet a deadline that was selected without consideration to performance or resource requirements or business impact.

It is very important, before choosing a project deadline for HRIS integrations, to understand the requirements, the tasks and the resource availability prior to confirming a delivery target. Many organisations start with a deadline and work backwards.

The problems with moving project tasks beyond go-live include:

1. The exposure is far wider within the business- any errors found are seen by everyone with access to the new portal/s. This often includes the MD, executives, managers and staff.
2. The costs to fix are greater- the systems are now moving, they are no longer stable quarantined environments. As such, the data is changing and very hard to align.
3. The funds dry up- the business has achieved the main aim of the project and the “nice to have’s” are often put to the side for future years.
4. The project team and knowledge has disbursed- often at the end of a project that has vendors and contractors involved, these people leave and move on to new projects. You can’t easily “get the band back together” again.

Our recommendation is to include the planning, testing, go-live, post go live support and BAU maintenance all in the initial project plan.

Key challenges

The main challenges that recur in all integration projects include:

1. Messy data
2. Identifying the authoritative source of data – per component
3. Determining master to slave data flows

Each system you have in your mix, will normally have both a specialised data set and a shared data set. In the case of a payroll system versus an LMS system for example:

• The LMS – Specialised data includes Learning courses, schedules, skills, qualifications. Shared data with other systems may include the staff first name, last name, dob, address, gender, location, role, email and phone number.
• The Payroll system – Specialised data includes bank account details, salary, allowances, super, pay history. Shared data includes first name, last name, date joined, role, location, gender, email address and phone number.
• The Award Interpretation system – Specialised data includes overtime rates, hours of work, schedules. Shared data includes payroll number, first name, last name, date joined, role, location, gender, email and phone number.

We can see in the systems above that multiple items can be shared.

In dealing with “messy” data, the key is to decide what will be the authoritative source for that specific item. No single system will be the authoritative source of all staff data.

To clean up the data before migrating to the new system, you need to decide a fixed comparison date for data. As such, you need to take a copy into a test environment, on the same date, for all systems being compared. This then enables you to complete a direct comparison.

At a future point in time, another copy of the systems will be made, to compare data. At this stage, you may only need to deal with the variances involved.

Another likely result of this analysis is that the processes for data entry into some systems will change. The ability to manually change data in some systems needs to be turned off and the staff need to be disciplined in following an order of entry.

If this is not followed then the systems will become misaligned again.

Reconciling post implementation

After an integration project between systems such as an HRIS package and a payroll system, the data sets can easily move out of line again.

The longer this is left unchecked, the larger the issue you will have. Imagine if you have 3 records that differ each day, within a month, this is 90 records that are not matching.

To minimise the impact, we recommend:

1. Complete a post-implementation data alignment check approximately 1 month after go-live. This will provide you with a solid understanding of how well your systems are now working together.
2. Spend a lot of time “fixing the leak in the boat”. Most people spend their time “bailing out the water”. They try to fix individual issues with pieces of data. The best method is to find the root cause and correct that. This stops new problems arising and provides you with the time to correct the individual issues without more being created.
3. Setup some data checks at regular intervals. This will enable you to locate and correct alignment issues before they expand too far.

Summary

1. Start with the vision
2. Address the challenges – master/slave, authoritative source of components
3. Reconcile – mixed sources, stable dates
4. Integrating is messy; address the vision
5. Synergy – manage changes over time

That will help you improve your integration results!

This is generic advice only, each circumstance differs.

If you would like further information, feel free to contact Gerard – gerardb@silverdrop.com.au