In Marco Iansiti and Karim Lakhani’s book, “Competing on AI,” they suggest that Artificial Intelligence (AI) is about reinventing “the core of the modern firm where processes are digitized and enabled by an AI factory that treats decision making as an industrial process.”
The aim of an “AI factory is to create a virtuous cycle between user engagement, data collection, algorithm design, and improvement.” The goal, they believe, is to create a “data supermarket.” But how do organizations go about creating a data advantage? This was the topic of a recent #CIOChat.
Where Does Data fit within the CIO Charter?
CIOs are clear that IT leaders have an important and sustaining role in creating a business data advantage.
Former CIO Isaac Sacolick says, “if data isn’t a top priority/mindset then that’s a problem! CIO must change the culture by defining what data-driven means at the company and lead self-service and data governance capabilities.” CIOs believe that IT organizations should own the following:
- Data health
- Data security
- Data quality
- Data utility
- Data accessibility
- Data governance
Former CIO Ken Leblanc believes achieving these ends requires “standards, orchestration, quality, rationalized repositories, integrations, underlying plumbing, and center of expertise.”
For CIO Carrie Shumaker, this means “ensuring there is a data governance process in place. It means, as well, ensuring data is made available, not locked down unnecessarily.”
Constellation Research Analyst Dion Hinchcliffe adds “the CIO top-level role in data is to be the steward and ombudsman. The line of business generally owns the data. The CIO role should include data security, data management, and data governance for the enterprise,” he says.
“CIOs should be an enabler to ensure data fully serves the organization,” he continues. “Data is the enterprise’s most valuable and irreplaceable business asset. This is why so many bad actors are focused globally on breaking into IT systems and taking it. The unfortunate truth is that most organizations greatly underuse what they have. Or they too tightly control and hoard data.”
With the Emergence of CDOs, Where Should CIOs Put Their Attention?
Nick Heudecker, Former VP of Data and Analytics at Gartner, claims “CIOs are at a strange crossroads with data today. They are both competing and collaborating with CDOs that may be taking their budget and staff, but they’re still ultimately responsible for data infrastructure.”
Some, says Microsoft Security Architect Wayne Anderson, believe, unfortunately, that the CDO is a yellow flag to the CIO to look at their role in the organization.” Without questions, CIOs need to make IT a better partner. To do this, they need to connect with the business. They need to make data transformative happen and need to become a data champion.
Former CIO Joanna Young argues for data to win there needs to be collaboration between technology functions. She says, however, “I have yet to see an organization with CIO, CTO, and CDO where the sum is greater than the parts, as roles and responsibilities get murky how can they be effective?”
With this said, IT typically does not own the corporation’s data. For this reason, Sacolick says, “a key responsibility of the CIO and/or CDO is defining owners and educating them on their responsibilities especially for data quality. With this, IT can support efforts with tools, monitoring, and automation.
What’s sad is that many organizations are just starting to realize this is part of the problem.” Sacolick argues, “too many organizations pin this tactical work on the CDO. Making them only a data fix it leader, is unfortunate. The most important role for the CDO is to generate business demand for using/consuming data and analytics and then demonstrating business impact.”
Sacolick argues that long term, however, “the CDO has to report to the CIO. There should be one chief of data. If the CIO is very data-oriented, they may hire an operational CDO. Financial services and other data-intensive businesses use CDOs for DataOps and may have separate Chief Data Scientists or Analytics Officers.”
Regardless of CDO reporting, CIO Joe Sabado encourages CIOs to “partner with CDO to ensure success of data culture adoption. And oh yeah, plumbing and security.”
Marco Vernocchi, EY Global Chief Data Officer, agrees with Sabado when he says “data is a team sport with business, data, and technology partnering in the game.”
Key Enabling Capabilities of Organizations with a Data Advantage?
LeBlanc believes in the digital age “business agility, data driven decision making, speed, visibility, and modeling all matter.” The first step in achieve these ends, claims Sabado, “is to know where your data exists. You can’t protect data/assets if we don’t know where they are.”
Likewise, Shumaker suggests that the foundation for a data advantage includes the following elements:
- Good enterprise reporting with a common report set.
- Self-service reporting tools.
- A data dictionary.
- Functioning data governance and identity management.
Sacolick asserts, meanwhile, “data is always changing. All the disciplines are iterative (hence, agile methodologies on data are important) and leaders must know when to pull which levers.” It is not surprising that Young asserts “there is a lot of ‘do not know what we don’t know’ to address data whereabouts. Organizations need to inventory their data.”
With the above said, Hinchcliffe suggests that “the best ways to create data advantage are as follows:
- Cultivate open master data.
- Cull data silos.
- Invest in data management.
- Provide self-service data access + analytics.
- Publish microservices-based business graph/API.
- Educate, evangelize.
- Story tell data successes.
Vernocchi says, “while each company is different, I see data quality as an end-to-end process rather than the outcome of a collection of functions. Data catalogs are a tool and data intelligence is the goal.”
What’s IT’s Role in Data
Without question, the CIO should be the shepherd of – or guide for – the data an organization process. This can involve asking the right questions of the owners to build an appropriate data catalog including to determining risk and value assumptions. Sacolick asserts, “IT’s role is ensuring the right data/quality data gets in the hands of the people analyzing data. What role do data intelligence and data catalog play in creating this advantage?”
Clearly, Young is right when she suggests “accessibility and usability is key and CIOs are either accountable or responsible. This often depends on technology or data savviness of long of business and other executives, as well as current data health.”
Sabado puts it simply by suggesting that IT should identify, protect, detect, respond, and recover data.”
In terms of the difficulty, Sacolick suggests that he has “done lot of work in construction and some manufacturing on data enablement. It is ‘data archeology’ work digging into their spreadsheets, where the data comes from, how formulas work, what manual clicks are performed, who does what and where the results are used. I have several tricks on how to identify data owners and slowly get people to assume responsibilities. But honestly, it isn’t easy.”
Hinchcliffe believes “a problem is that data quality often is only able to be addressed by IT. Business stakeholders are getting more capable all the time, but most of the time you can’t put them in top-level charge of the quality of data in IT systems and business applications.”
Given this, he says “IT systems must be the unblinking guardian of data quality. They should not let bad data in. It is important to use data management to clean it up plus normalize it. With this, it is critical to find ways to make it intelligent, open, reusable.”
How Does Multi-Cloud Impact Data Strategies?
Hinchcliffe suggests that “since a highly diverse multicloud is the future, data strategy and data management are profoundly impacted by it.” For this reason, he says “the CIO of tomorrow will expend tremendous energy and resources to steward data across their vast multicloud estates. The organizations that do this well will win.”
For Sacolick, “multicloud adds to the complexities, but the data landscape is full of greater challenges. It is important to have multicloud data integration platforms and governance on where you are starting what data, why, and how. Young, however, adds multicloud is just different infrastructure, the impacts are on tooling and differing APIs/services, but data is data in the end.”
Parting Words: Data Matters
Clearly, data matters. Sacolick suggests the problem is that laggards can stay with the status quo of the spreadsheet being emailed, access databases with no one supporting them, and CEOs making decisions on month-old data.” So where to start, Sacolick suggests it depends:
- No data / Data integration
- No questions / Citizen data science
- Too much data / Data Management
- Dirty-dark data / Data Governance
Regardless of where you are, it starts by the business realization that competitive advantage is with those that are great with data. Winners create virtuous cycle and a data supermarket. And they have as Satyen Sangani likes to say a focus on ‘data culture.’