The pressure upon firms to transform and modernise continues to increase, with global spend on digital transformation estimated to more than double to $3.4tn by 2026, from $1.6tn in 2022.
Firms want to drive down costs, provide new and compelling customer journeys, and improve the speed and quality of supply chains, operations, and decision making.An ERP upgrade is often identified as a critical strategic initiative to meet some or all of these aims, but while the benefits to a robust, modern ERP are well known, the challenges in executing a successful ERP-driven transformation are often not.
Too often ERP programs fail – Gartner estimates up to 75% ERP projects fail or do not meet their intended objectives. While companies generally do not make them public knowledge, the list of companies with failed ERP programs is extensive, recently including a global grocery chain headquartered in Germany that cancelled its ERP program after it had cost them €500m.
The causes of failure are numerous, from lack of implementation controls to inexperienced programme leadership to poor selection of integration partners.
However, there is one major, common cause of program failure that requires much greater understanding – data migration.
Data is commonly recognised as an essential company asset by leadership teams, yet thereremains a lack of appreciation for the level of complexity and effort required to migrate data from legacy sources into a modern ERP, and the importance of having the right proven data migration toolsets, methods, operating models, and experience.
It is easy to forget that for many years data was not a priority, and the 'sins' from the past are brought into sharp focus at the point of data migration. For companies with long histories, the historic abuse of company data standards, data governance, data policies, and data cleanliness will become their number one challenge in their ERP transformation (or indeed any digital transformation).
So why is data migration so challenging? In our experience, we see five common challenges that cause companies to fail with the data migration:
- Data migration is assumed to be simple, and simply two
words: "Data" and "Migration"
If you are moving from any legacy technology – SAP or non-SAP – to S4 or to any other ERP target landscape, it is tempting to think data migration will be simple. It's just moving data from one table to another, right? ERP tools offer many templates to make the process seemingly simple.
Moving from multiple legacy systems to a new platform? A quick scan online lists many tools and system integrators who promise a quick and simple data migration. Reality proves otherwise. This upfront assumption sets off your transformation with unrealistic cost and timeline expectations, and typically an over-simplification of the organisational and technological approaches required to make your transformation a success.
- Shortage of expertise; underestimating
effort
Data Migrations inherently require a range of data experts:
Data experts within the company: Who in the company understands the local interplays between business processes, systems, and data, and can understand/ shape the detailed data requirements of the target system? When experts are found, they are usually needed for long durations. This is often a problem, as these same experts are typically highly sought after to run the business.
Data expertise within external partners/suppliers: Each company and business unit is unique and complex. Previous experience of data migration in Retail does not prepare system integrators for leading migration in the Utilities or Energy sectors, for example. Expertise in global transport logistics data objects is little use for data migration of finance data objects. Partners or suppliers are typically forced to 'find' expertise that does not exist.
Expertise with the chosen data technologies: Spend on global transformation is increasing year on year – job markets have not caught up. Partners or suppliers are forced to appoint junior, inexperienced resources and drive on-the-job training.
It is common for us to encounter ERP transformations with very few client and partner experts, which directly causes delays and budget overruns. We see these challenges for all tiers of system integrators.
- Black Box Activity
Data becomes extremely difficult very quickly. Even a simple migration can require hundreds of tables that cover many topics. Then there is the technology and acronyms – ETL, BODS, LSMW, MDG, Staging, Environments, Pipelines, X-Refs, Pre/Post-load validation and so on.
Data migration is a professional discipline, which many invest their entire lives in. The industry speak is often accompanied with a lack of pragmatism and communication from integrators. This often serves parties – it encourages a lack of scrutiny and governance. We see a lack of focus on 'practical technology' – making data technologies and approaches simple to understand and manage.
- Overselling technology and underestimating
costs
We have witnessed many system integrators (across all tiers) talking about their unique philosophies, approaches, technologies, and tools that support data migration. These can often sound and look great, but in practice – and despite the availability of many impressive technologies – we continue to see an extensive use of migration manually or by Excel, and off-the shelf processes require high levels of customisation to make them relevant for your business.
Similarly, we have witnessed integrators making an over-simplification of the challenges at hand, offering overly competitive rates from the outset. This is a recipe for major downstream challenges, as costs inevitably rise, and integrators/ suppliers look to cover costs and pass this on to the business.
Combine the preference for lower costs, a tendency for integrators to over-sell upfront, and a shared and unrealistic wish for data migration to be simple, with little expert scrutiny upfront of data migration approaches, technologies, and tools, and it is easy to predict the programme frustrations and challenges this leads to downstream.
- Hands off
Given the packed agenda of senior leaders in any business, it is easy to understand why their sponsorship of data-related activities is often de-prioritised. While they will understand the importance of data from an academic perspective, very few will engage in data willingly. This is a problem.
Successful data migration requires a huge investment of time and effort by many people from an organisation. A lack of senior sponsorship leads to data being consistently de-prioritised. Front-line staff will prioritise their day jobs and passively resist programme needs. This comes at a cost.
Unless data ownership exists, the quality of the data migration will be compromised. This directly increases the chances of program delays, and of system failure at the point of go-live. One client lost more than $6m per day over a nine-week period because their inventory effectively became invisible at the point of go-live.
Five key success factors – how to get Data Migration right
Outside of the boardroom, industry practitioners will admit that all transformations are demanding. They recognise that late, over-budget programmes generate high levels of stress for the boards, shareholders, and investors for extended periods of time, and that these symptoms are 'par for the course'. It doesn't need to be like this.
Our top five critical success factors are as follows:
- Migrate the minimum by reducing volumes – resist
the "I want it all" mindset
The point at which businesses recognise the value of legacy data is typically at the point where they are asked to let some of it go. Simply put, the more data, the more effort and risk to the data migration. The more data that is required, the more data quality issues will be present, the more expertise and effort will be required to fix the issues, and the greater the risks to timescales and costs. Recognising the importance of data as an asset to an organisation is important, but it is critical to be precise on 'what' data is the real asset – which is often significantly less than 'all business data'.
- Get a working data migration execution
model
Unless a firm is blessed, the fact that a business is initiating an ERP transformation typically means there is significant complexity to wrestle with. Therefore, the set-up of an effective data migration execution model is vital, adequately resourced with the right capacities and competencies. This operating model must plan to involve at a minimum:
(a) Global data experts and data managers.Global experts must be accountable for defining the company data strategy expected of the migration, across each factory, site, office, process, data object, and so on. The operating model must also plan for their active involvement throughout the migration to resolve any conflicts with global strategy and provide expert leadership.
(b) Local data experts. Local experts will play an essential role in testing the global strategy, identifying exceptions, cleansing and preparing the data, configuring local data (e.g. mappings), and ultimately validating the data. Depending on the business, high numbers of these experts may be required. Warning: These local experts may be hard to find or may not exist and those that do will likely be sought after for BAU activities and other streams within the programme. There are ways to mitigate these risks.
(c) Strong governance. The organisation of disparate and numerous individuals must be carefully orchestrated between business and system integrators. It is very easy in complex data migrations for a point of 'stagnation' to be achieved, where multiple interdependencies contrive to stop any progress. Governance should be ruthlessly transparent and goal orientated, with clear accountabilities.
- Select the right data migration
technologies.
Selecting the right tools and technologies for your data migration will directly and quickly mitigate risk. Effective use of technology will reduce manual inputs required from the programme team of experts and strengthen and accelerate Master Data Management. Transparency and control will be efficiently introduced over data transformations and mappings, flexibility and loading performance will be enabled, and the need for custom coding will be reduced. Processes such as cleansing and validation will be simplified too. Some system integrators will resist the use of market tools and technologies because they do not have the native skills in their organisation.
- Define efficient methodologies.
While a system integrator will typically have their 'off the shelf' processes and approaches, they are seldom fit for purpose for each unique business. Practical, efficient processes that leverage tools and technologies must be defined – with RACIs clearly outlined for your unique operating models – but it is essential that they are not over-engineered as data migrations become complicated quickly. Pay specific attention to escalation routes, data boards, and steering committees to focus on efficient decision makingat each level of the operating model, from the local data expert up to the programme leadership level.
- Focus on legacy data dependencies.
Many of the biggest challenges in data migrations stem from legacy data interdependencies. Where they are not clearly identified, the data migration quickly runs into trouble due to missing analysis and due diligence of dataand cross-references, which can cause extended periods of unplanned work. For example, migration of open projects, will be dependent upon multiple other data items (e.g., Sales Orders, Purchase Orders, BOMs, and WBSs). Unless dependencies are not clearly identified and tested on the data, the migration will quickly falter.
With such a high failure rate for digital transformations, it is right that boards and shareholders ask themselves the question: "how do we best protect our investment in ERPs?".
At AlixPartners, we have rich experience of working with large, complex global corporations in their ERP transformations in either leadership, assurance, or program rescue roles.
We understand the commercial importance of delivering successful transformation – and the operational and technical challenges of getting it right – making the difference between a flagship success and a drawn-out programme failure.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
