MRO System Implementation and Data Migration - What You Need to Know
Last weeks conference confirmed the current market situation, airlines and MROs not only want to simplify their system landscapes but also improve data standards for a future proof organization.
When managed properly simplifying the MRO IT landscape can drive significant costs savings. One critical part in this is will be the migration of data as the data that resides in the system is the airworthiness of a fleet. Think of do the cycles and flight hours per flight align with the accumulated of the total recorded flight hours and flight cycles? The data is relevant to keep the aircraft flying.
We often hear from operators that they blame the MRO/M&E system for not working properly although they just implemented a new one. That they need to have additional excel sheets to keep track of everything. The reason behind this is that during the initial data migration for the MRO/M&E software system implementation the data was just pushed from the old system to the new system. There was not enough time spend on looking at the actual data itself, if it’s correct what was recorded in the old system or looking for the actual source of truth. Thus, engineers do not trust the new system and revert back to the additional excel sheets.
To get it right from the start we have put together an overview of the most important points you need to know below.
Introduction
Data migration projects are often perceived as daunting, risky and in general difficult to find a proper strategy. This can lead to overestimating the project, but, more commonly, underestimation the requirements and needs of a data migration project.
A data migration project is like building your own house: where first you must get the foundation right, then you can raise the walls and, in the end, put the roof on top. If you do not get the foundation right, the walls will be instable, and the roof will collapse. Like building your own house, as a self-made handyman, you will always face different challenges during each part of the project, and it helps to know the best practices in the industry to prepare you for these.
The Foundation
So, let’s start with laying the foundation of our data migration: this is a balanced mix of senior management support, change management and overall data migration experience. At the start and throughout any data migration project, these three need to be in place otherwise the whole house starts crumbling down, leading likely into failure of the project.
Senior Management Support
Do you know and formally agreed who the project sponsor is within your company? This could be a single person, often one of the potholders within the airline, or a steering board. The project sponsor has final decision power and controls the budget allocation for the project. Furthermore, the sponsor has a key role in gaining and keeping support from senior management. Hence, ensure Senior Management is enthusiastic and supporting the project.
You need the senior management on your side to support the changes your project team is about to make in the company. Next to that, they have ultimate control over company resources that might be needed for the project.
Pre-agree on initial budget and available resources.
This requires proper project scoping and planning on forehand, to give the right estimates. Ensure this is included in the overall company planning and &department heads are informed and covered (they potentially loose someone from their team for a certain period).
Keep senior management involved in progression and decision making, it keeps the project on course with company strategy and goals.
Change Management
Data migration is not only moving data from A to B. In one way or another, it will invoke new business rules and, possibly, company processes. Most of the time a data migration is part of a bigger change within a company or the start thereof. People are naturally resistance to change, especially if it affects a well-established way of working that your employees are used to, it’s difficult to get people out of such a comfort zone and create a new one. Think about all those engineers.
Ensure you have inhouse change management experience, or otherwise bring someone externally into the project team that can:
Act as bridge between the project team and senior management, balancing the interest and concerns of both sides.
Celebrate project victories but also keep up the positivity when the times are tough.
Is able to get the support of the company as a whole for the changes that are about to happen (refer back to the Senior Management support).
Data Migration Experience
Often there are smarter ways to approach something then via Excel. People and/or companies with extensive data migration experience can help to push your project in the right direction. They can show how to make optimal usage of technology and resources, while avoiding common pit falls and mistakes.
Technical IT data experience is good, but a combination with actual industry knowledge and experience of your field is the real success factor.
Gather people that can tell good data from bad data, meaning, next to technically being able to extract and process data, they understand what they are looking at.
Challenge, validate and learn
Ensure you have persons in your team that, because of their experience, can challenge the project decision and working methods. It helps to keep things sharp and heading in the right direction. With challenging, validations of data quality and consistency are triggered, improving the understanding and ultimate result. Furthermore, it helps the learning process within the project team to ramp up effectiveness and avoid making the same mistakes. Which brings us to the last point:
You Do Not Have To Reinvent the Wheel (unless you really want to)
There have been many data migrations within aviation before, and whilst every project has its own particulars, 80 to 90% of the path will be the same. Save yourself time, money, costly delays, unpleasant surprises and frustrations within the project time by ensuring there are people that have done this before and already bring the wheel. You still have to pick the size, polish it and mount it, but at least not invent it.
The Pillars
Once you’ve laid down the foundation of the data migration project and start the actual work, there are three key aspects, or pillars, that need be controlled throughout the project and will ensure the end-goal (or the roof of the house) is top notch.
The key focus of a data migration is of course the data and its migration, nonetheless, the use of applicable technology will help in the ultimate goal of achieving go-live with satisfying data quality. This all should be continuously managed within the project team and towards the rest of the business, such that the project remains on track. Let’s dive deeper in each of these three pillars data, technology and management:
Data
Initially, the project is seemingly simple, moving data from A to B, but what data do we need to move? Defining the “dataset” you need in this case is actually a large chunk of the project and one of the more resource intensive activities, where CAMO key-users and the data migration team have to closely work together.
First is to define the scope, which starts with the question; what data do I need for our destination “B”?
A good start to answer this question should be found in the documentation or experience of the actual vendor of the new, or, ideally, experienced users of the new system that are part of your team.
The scope is required to perform an accurate data mapping, to find out what data you need to move from where and possible must clean or enrich. This requires deep knowledge of your data and the possible source(s). Really knowing your data is difficult, often everybody has a “feeling” of the problems that might be there, or even comes across data inconsistencies daily.
However, without a deep and factual analysis, nobody actually knows what you will be up against. Furthermore, nobody knows everything and thus this will be very much a team effort. Experts from all the different disciplines need to work together. People from maintenance programs, modifications, components, planning, logistics and whatever other discipline is relevant, need to work closely together with the data migration team to perform the data mapping.
The important questions to get answered during the data mapping sessions are:
Is there a single source of truth, or is data to be combined?
What type of sources do you accept as trusted, and thus what should be treated as untrustworthy and requires additional verification?
Identify any gaps in the data and means to close these gaps, this could imply that you will need to go back to the OEM source or dirty fingerprints of the aircraft to find the answer.
To emphasize, Data Mapping is NOT a primarily IT exercise, it requires full participation of multiple disciplines and in general, the CAMO is expected to be the lead of the projects:
- Subject Matter Experts of CAMO disciplines:
Maintenance Programme
Components
Modifications
Maintenance Planning
Base/Line Maintenance
- Subject Matter Experts for material purchasing, inventory and stores
- Possible source (system) experts or experienced users of the systems you are migrating from
- Data migration team/expert
Once the data mapping is established, it is time to focus on the actual improvement of your data, one of the reasons such a project takes place in the first place. While scrutinizing the data set, this is the chance to prepare it for the latest industry standards and best practices. Think Spec2000, ATA200 etc.
Additionally, the target system normally also enforces standards or a certain structure in the data. To understand these, it is paramount to immediately invest in system familiarity and ensure there are consultants available to help you in grasping the required data structure and its specific challenges.
Once the data structure is mapped, business rules need to be established for how to transform certain data in the new standards/structure. Furthermore, a process needs to be put in place on how to handle exceptions to the rules and how to incorporate potential new business rules. With the use of technology, as you will read later, this is much easier than in a manual data migration process.
- The last key factor, or rather factors, in handling your data are data verifications.
From our experience this is an often-overlooked aspect in both importance as well as resource requirements. So, let’s get some things out of the way, in order to prepare you for your project:
Without proper data verifications, your project will either fail or it will be a very and long post go-live period, riddled with issues and diminishing business support due to the perceived unreliability of the new system. One data verification, is no data verification, expect a continuous effort throughout the project on verifying data, exception reports and decisions thereof. The feedback from this is essential for the next data migration test iterations to further enhance the data quality.
Proper data verifications are not a “15 minute” thing, expect that, depending on the extent of a verification, multiple people require multiple days to complete these. Meaning, they will not be able to pick up other project activities during that day.
Communication: ensure feedback is communicated effectively. The best method is to have a oral debriefing with the data migration team, where the verifier can share his/her findings and the data migration team can immediately look into these on the spot and discuss possible causes and solutions with the verifier to confirm before further processing.
Ensure there is a structured verification plan and also an archive of all findings. Verifications need to be repeatable to test if the outcome if improving. Furthermore, having this in place early in the project also allows to involve certifying authorities earlier or at least ease the certification phase. In the end, authorities need to be satisfied that you can still guarantee the airworthiness of the aircraft with the new system and the process towards such a system safeguarded this.
Automated verifications are possible and external parties can help. However, it is also the means to approve data for use in the new system and thus directly certifying that aircraft are airworthy based on this data. Thus, in the end, your own business users will need to be satisfied with the data verification and thus the quality and sign off on it.
While above statements might sound harsh in some cases, these are actually in many projects the make or break aspect for a project to be a success.
Technology
If the data content itself seems daunting, don’t despair as technology is here to the rescue and make life better and easier. That is, when it is used in the correct way.
The to-go tool that often pops-up when thinking about data migrations, is Excel. While history proved this is certainly possible, the usage of it is still largely a manual involvement and very prone to errors. Furthermore, it only accepts limited input sources and also is very limited in its output formats, requiring further steps for final conversion and thus adding complexity to the process.
Luckily technology advanced which gave rise to much better alternatives completely designed for this exact purpose. These ETL software solutions, which is short for Extract Transform Load, are capable of visualizing complex data conversions and apply advanced logic to these while on the fly outputting intermediate results and exception reports anywhere in the flow.
The most noteworthy advantage is that the data transformation process is repeatable instantly (okay, you are dealing with GBs of data, this could take a few minutes) with the click of a button. It opens the door to an iterative process of continuous testing and improving, being able to easily tweak the rules set-up in the ETL scripts to get needed results.
No matter the tool in between, you still have to deal with many different types of source data such as:
Previous/to-be-phased-out M&E system(s)
User created databases or Excel files
OEM websites/documentation
Dirty Fingerprints
Manually created mapping matrices/files
Not all this data is efficient/easy to retrieve or read, some formats such as PDFs (certificates, OEM documentation, Form 1s and other dirty fingerprints) or OEM sources, normally require tedious manual ready and conversion. Luckily, even for these seemingly hopeless cases, technology comes around to lend a hand. The two most common we as EXSYN also are using regularly during data migration are:
- RPA (Robotic Process Automation)
Imagine a computer wide robot or macro that can mimic any human actions on a computer, that is an RPA bot. This technology can be used automate repetitive tasks, such as scraping OEM websites for documents and read/download these, or transposing information to a place which normally would need to be done manually.
- OCR (Optical Character Recognition)
This might be a familiar function, but the technology is advancing rapidly. With OCR the computer can identify text in images (or PDF scans for example) and convert this into data/text. OCR is very useful for ready Dirty Fingerprints, release certificates and other forms of PDF documentation that would otherwise need to be transposed manually.
Management
The third and last pillar supporting the road to a successful go-live of your data migration project, is continuous project management. Depending on the size of the project this could comprise of one or multiple people, however, there should be one main project manager, who controls the actual project planning.
He/she can then be supported by someone who will manage the different Engineering/CAMO side key-users and a lead contact for the project from the IT department.
The first and foremost goal of the project management is to give the project and the team a direction at all times. This is achieved by having a, continuously updated, project plan, with clear goals, objectives and achievements. Any data migration project should at least have a:
Project Plan
Data Mapping document
Containing the data mapping and agreed business rules applied to data transformation.
Data Migration issue list
A list containing any finding during data verifications, their extend (e.g. how many aircraft/components/records affected) and the tested solution (or a due date for the solution)
Data Verification plan and sign-off document
A document detailing the different methodologies of data verification applied, the results and if these were satisfactory, once the latter applies also whom is to sign-off on the results and of course the actual sign-off. This will act as prove to the CAA that the processes applied for the data migration are guaranteeing the continuous airworthiness of the aircraft.
If the project encompasses a larger team and thus numerous tasks for each team member, it is recommended to divide the management thereof in different “workstreams” or major disciplines, such as: data migration, data verification and data cleansing.
Each of these workstream will have a single “workstream coordinator” who is reporting to the overall project management. The job of this coordinator is to translate the high-level project plan, into a detail, task driven, plan for their particular workstream.
- Delegation of tasks
This brings us to the second point, dare to delegate responsibilities by empowering team members to make decisions on a certain scope or self-regulate to a certain level. The project team is trying to accomplish a common goal and were selected in general for their expertise and/or experience in different areas. It is therefore important that the project management can exert a certain amount of trust to its team and allows it to get the work done, without continuous intervention or complex decision-making hierarchy. This will only slow down the project and run the risk of diminishing quality.
- Definition of Done
To steer such an organisation, it is still important that in the end, there is one responsible project manager that can make any final decisions when needed. However, this person should use all tools to his/her disposal to let the team self-regulate the day-to-day activities. For this, it is very important that every milestone/goal/objective/task has a clear definition-of-done (or DoD). The definition of done should contain a clear outline of scope, when a certain task can be marked as completed.
100% perfection does not exist, thus be realistic about these if you would like to complete your project. Especially technical team members can strive towards perfection in data migrations, that are hard or impossible to reach in a reasonable amount of time. The DoD is a very effective tool to help everybody keep focused on the scope, priority and progression.
Such definitions will also help the project management to prioritise tasks better and deal with planning changes. Be prepared, these will certainly come, both positive and negative. A clear scope both on high level as well on more detailed level, will help the management in making the tough decisions but also to reprioritise and possible adjust the scope for go-live.
For each data migration project, there will always be bits a piece left, either some tweaks or some larger elements like uploading OEM documentation or bringing across Structural Damage chart, that did not fit the initial project timeline after all, but were deemed not essential to reach the overall goal. Do not despair if this happens, this is normal, however, you can prepare early on this by assessing each task and mark it as:
Must-have (or no-go items) - Without this there will be no go-live possible.
Should-have (or go-if) - Data that is definitely necessary for a smooth operation but can go-live with IF available to a certain extend/scope or format if times does not permit to fully finish.
Nice-to-have - Any data that is good to have available, but more from a convenience/excellence perspective and can be lived without or sourced elsewhere for the time being.
Lastly, which is the most important job of project management, is people management. This is also where the change management experience comes in, that you put into the base of your project. The project manager needs to constantly align the expectations of both the project team as well as the project sponsors. Don’t be overly negative, but also be realistic and set unrealistic expectations, as these will be perceived as failure when not reach which can heavily impact morale.
More often than not, the people working in your project are temporarily assigned to the project from elsewhere in the organisation and thus their reporting manager is likely not you but actually in their original department. This puts you, as a project manager, in a difficult position where you cannot fully manage people’s time allocation, holidays, priorities and contract conditions. It is therefore of the absolute essence that you gain the full support of each individual team member, get them on the “side of the project”, and gain their full trust, both in the good times as well during the more though times. So, celebrate victories both in the project as well as in people’s private lives, but also support people during rougher seas. Most importantly, be transparent, involve people in your decision making and story, and consider each other’s ideas and input.
The Roof
Finally, the finish line is just across the horizon, almost there! A successful go-live cannot be accomplished without thorough preparations. However, there is one step before: The User Acceptance Test (UAT). This is the ultimate simulation in the target system, a general rehearsal for the day to day operation after the go-live so to speak. It is also the point at which you ensure yourself and the business that you are ready for it and everything is in place as it should be an official sign-off moment after successful testing.
For the UAT at least prepare:
- UAT load, which should be close the 99% of the dataset you will load during the cut-over.
- Test protocol with test scenarios.
- Business processes and workplace instructions (WPIs) to follow.
- Communication plan and governance around issue capturing.
- Sign-off documentation.
- Sign-off matrix, normally the different department heads, or selected deputies of (the key-users).
The build up to the UAT will also serve as the first cut-over rehearsal, thus the moment towards a go-live where you will load the data for the final time and activate the new system. It is therefore important that you also have a first full version of your cut-over strategy ready.
Cut-over strategy? Yes, this is the script that details the cut-over process step-by-step in minute detail, with pre-recorded timings and in sequence of events. It should say when a certain person is expected to do what. For Example:
01-01-2019, at 20:00 à IT Person will make a back-up of the old system
- 01-01-2019, at 20:15 à IT Person reports to Project Manager that back-ups have been made
- 01-01-2019, at 20:30 à Data Migration team loads first file
Etc. etc.
It is important that this process is fully predictable and fleshed out, such that any issues during the cut-over itself can be prevented or immediately spotted it if differs from the strategy. During the cut-over you have generally have a narrow and limited time window. After all, the operation has been locked out of the old system which is frozen and is awaiting the transfer to the new system which you are currently loading up and activating. Thus, everything must be prepared, and everyone’s role and responsibility must be clear.
Arranging proper 24/7 support from all involved departments is absolutely necessary and these should be represented at all times during cut-over. Ensure everybody has the appropriate access rights, decision-power and tools available for their role, but also in case something goes wrong and need to be mitigated.
Talking about mitigations, the last important part of your cut-over strategy is the fallback or back-up plan. There should be several No-Go/Go decision moments during the cut-over, where the coordinators of the different workstreams working on the cut-over come together and discuss progress to ultimately decide if to continue or stop. If all goes well, this would of course be a Go, however, you also must be prepared for the worst case. The fallback scenario should describe what steps must be taken to revert and resume operations on the old system.
The cut-over is a lot to take in, is like a full blow musical where everybody needs to do the right things at the right time and in the right place. Rehearsals are key and in the week before cut-over should be done multiple times until everybody can dream it.
Now the go-live has been achieved, we’ve built ourselves a house………., apologies, successfully performed a data migration project. Sure, there are always improvements you can do in the future to your house. However, now you are confident it was built on a solid base, the walls won’t budge, and the roof will not blow off. Thus, it can serve the purpose of your organisation for many years to come.
How EXSYN can help
EXSYN's team of aircraft data and aviation experts utilize a proven data migration framework and methodology. It has been applied to millions of terabytes of master data and includes:
EXSYN’s NEXUS solution to reduce project costs and duration
EXSYN’s data warehouse to accelerate your migration
Data mapping and load workshops to develop the best strategy for your situation
EXSYN’s data migration dashboard to monitor the progress and quality of your data migration
Migration of both structured and unstructured data
ISO 27001 data security certified migration approach
Feel free to get in touch with us, we would be happy to discuss your situation.