Data governance framework seen as bedrock of airline MDM strategy

Data governance emerges as a prerequisite for a sustainable, long-term master data management strategy.

Like many organisations managing huge stores of potentially valuable data, Air France-KLM believes it has much to gain from an effective Master data management (MDM) strategy, both in terms of streamlining business intelligence (BI) processes and delivering reliable business insights to operational teams. At the same time, the airline group knows it has much to lose by rushing headlong into a decision on MDM tools and deployment strategies without carefully laying the groundwork first.

That’s the view of Philippe Profit, who runs Air France-KLM’s business intelligence competency centre (BICC). For Profit and his team, he said, the time has come “to set a new bearing and to move in a new direction with BI,” not least because of the need to achieve a greater level of consistency between the data held separately by the two airlines that merged in 2004 to create the Air France-KLM group.

Each company – Air France and KLM – has its own data on flights, crews, engineering and maintenance, ground services and commercial back-office operations. And both airlines have their own distinct histories in managing master data, Profit said.

So before embarking on an MDM project to marry the two sets of master data, Profit believes it is essential that a rock-solid data governance programme be in place. “Without data governance, effective use of data is always going to be limited, and any kind of conflict resolution in regards to master data will be a dead-end effort,” he said.

Data governance a must for MDM

That view is echoed elsewhere. Rob Karel, an analyst with IT market research company Forrester Research, stated that “data governance roles, responsibilities, policies, processes and organisational alignment must be a prerequisite for any MDM architecture or technology decision.”

But while MDM maturity involves a critical dependency on data governance, the conversations that Karel and his team have with Forrester clients suggest that very few have achieved that level yet. In a November 2010 survey of 188 IT professionals with MDM experience, only 17% of those asked to rate their organisations on data governance maturity felt that a high level had been reached. Eighty-one percent, by contrast, said their organisations displayed “average” or “very low” maturity.

Like Air France-KLM, many organisations are interested in MDM: Forrester saw a 90% increase in MDM-related client inquiries between the start of 2008 and the end of last year, according to Karel – and that, he noted, doesn’t include the significant growth in complementary inquiry topics, including data governance.

Karel expects the slow evolution of MDM to continue over the course of this year. “But most likely three or more years will pass before the barriers to adoption for MDM start to crumble for real,” he said. “Organisational barriers like executive sponsorship, lack of data governance and lack of alignment with business priorities remain significant hurdles that won’t be overcome quickly.”

Karel’s message is clear: Only when the business as a whole takes responsibility and accountability for the master data that affects business processes and decision making will MDM truly become a strategic investment that can deliver significant business value.

For many organisations, said Karel, there is still much work to do in getting the technical competencies and infrastructure in place for effective data integration before MDM can even be considered. Or, as he put, it: “Can you move data from point A to point B with integrity and confidence? The ability for your organisation to manage both the scheduled batch and near-real-time transactional movement of data from one system to another system serves as a critical enabler for a complex MDM architecture.”

Data integration at British Airways

Improving data integration efficiency has become a major focus at another airline, British Airways, which over time will face new BI challenges arising from its acquisition last year of Iberia.

In a current project due to be completed later this year, BA is migrating the integration platform that it uses for shifting data from operational systems into its data warehouse, known as the BA Integrated Commercial Warehouse (ICW), from a single node to a multi-node grid architecture. The higher scalability and availability that this new architecture delivers will mean a big improvement in data warehousing service-level agreements (SLAs), according to Ian Johnston, IT operations manager at BA. It will allow the company to extend its data integration activities into areas that it doesn't currently support, such as crew rostering.

“Right now, our SLAs work on the basis that reports are delivered between 9 a.m. and 5 p.m., from Monday to Friday,” he said. “By moving to a multi-node architecture, we’re in a better position to deliver reports on a 24/7 basis, because we’ve now got failover options for data integration. In this way, the ICW becomes a 24/7 warehouse.”

Going forward, Karel said, other organisations will need to invest in foundational data integration and data quality competencies, while continuing to evangelise organisational support for a formalised data governance programme.

MDM horizon at Air France-KLM

At Air France-KLM, meanwhile, the groundwork for MDM continues. The BICC, a cross-functional organisational team, now comprises an executive steering committee, a data governance council and a team of data stewards operating in different parts of the business

According to Xavier Henderson, a senior consultant at Air France-KLM who specialises in MDM, the data governance council will be guided in its activities by the seven objectives of the  “Data Governance Council Charter,” published by Jill Dyché and Evan Levy of Baseline Consulting in their 2006 book, Customer Data Integration: Reaching a Single Version of the Truth.* Above all, the role of the council is defining and putting in place standards, definitions, business rules and policies to help ensure that Air France-KLM’s yet-to-be chosen MDM system will scale over time, Henderson said.

Data stewards, he added, are already in place across Air France-KLM’s many business domains “but as yet without the overarching data governance piece that will tie their work into a coordinated, companywide strategy.”

Once data governance policies and strategies are clearly defined – and only then – it will be time to move ahead with MDM. Henderson said he plans to start small by “focusing on noncritical systems in areas of the business where there is already some MDM achievements and skills.”

And every MDM proof of concept that Air France-KLM performs will be accompanied by a clear business case. The proof-of-concept exercises will be split into two phases: Phase one will focus on data relating to marketing and communications, air traffic and ground services operations, while phase two will focus on finance and human resources. “We’re very conscious that, once we start down this road, we can’t stop – hence the role of data stewards in our governance efforts,” Henderson said. These data stewards are responsible for the integrity of data generated by their part of the business, along with its associated metadata. For example, it's their job to ensure that each data element (such as a customer postcode) is clearly defined, used consistently and doesn't conflict with other data elements.

That’s in line with guidance from Forrester, expressed in a report of which Rob Karel is the principal author, Trends 2011: It's Time For The Business To Own Master Data  Management Strategies. “Organisations will need to consider MDM a multiyear, multiphase effort that can and should start small, targeting the critical few lines of business, applications or geographies that can realise the quickest benefits – with additional parts of the business deriving value from a standardised MDM solution over time.”

But the MDM tool of choice is far from decided. In fact, said Profit, it’s likely that several tools will be used, according to the needs of different business domains. “The way we’ll manage master data relating to engineering and maintenance, for example, will be very different from the way we’ll manage master data relating to customers,” he said.

For now, getting the data governance programme off the ground takes top priority. And getting it right is critical to Air France-KLM’s future business success, according to Profit. “Our data is valuable,” he said. “All businesses say that, but for us, it’s not just lip-service.”


* Dyché, Jill and Levy, Evan, Customer Data Integration: Reaching a Single Version of the Truth, John Wiley & Sons, 2006, Hoboken, New Jersey, p.158.


Read more on Data quality management and governance

Data Center
Data Management