255 delegates from local and global leading organizations attended the CDAO Africa 2019 event, which delivered as promised. The attendees could be categorized into three main types of people that attended the event.
1. The leaders who seems to have solved some of the key challenges and wanted to share their experiences with the community
2. The intermediaries who are still finding their feet in the data world but that have started with some form of Data / AI practice
3. The followers who are either behind, or not yet convinced of all the latest buzz words such as Data Science, Data Strategy, Chief Data Officer etc… (mainly due to the hype of Big Data and Block-chain not too long ago)
Data Governance vs. Data Management
During the pre-conference day, the emphasis was mainly on Data Governance and the role of a Data Strategy. However, consensus indicated that these camps are still divided, as no-one has pulled this off with a 100% success rate, the current leading practices are exactly just that – leading practices. Therefore, there is no silver bullet just yet and everyone is hoping to see some maturity in this space over the next few years as the leading organizations start to combine change management, culture and industry accepted frameworks to deliver on their data strategies.
Nevertheless, one thing that did stood out is that there still seems to be some confusion around DATA MANAGEMENT and DATA GOVERNANCE was indicated by the HR analogy.
In addition, the HR to Business relationship provided a good analogy where Data Governance is like HR. In this analogy, HR does not manage employees or manage their workloads. It remains the role of management to do this within business. Moreover, HR provides the frameworks, best practices and governance for managing employees effectively, and how to derive the most value out of them. Similarly, Data Governance does not own any data, nor does it manage any data. It’s all in the name, it governs the responsible and effective use of data within organizations. Data Management, on the other hand and as the name indicates, manages the data, the creation thereof, as well as the consumption of the data.
DAMA, as with most governance frameworks, was discussed with some mixed feelings - some typified DAMA extremists, whilst others just don’t seem to gel with the concepts. It appears that the Data Management vs Data Governance argument is mainly to blame for this.
The key message derived from this could be: Data Governance (governance and control) != Data Management (management of data)
In the light of this, some delegates left the session enlightened whilst others could be more confused…
Data Science vs AI
The conference was packed with informative sessions ranging from AI and Digitization, 4th Industrial Revolution, Importance of Data, Responsible AI, Effective use of BI, Emerging Data Science Teams, Agility and Adaptability to change, Data Strategy, Customer Experience and more.
Day 1 of the conference was packed with presentations by local and international speakers ranging from Facebook to WesBank.
Varied approaches came strong to the fore, however the notion of timing received a lot of attention. Besides getting buy-in from business, funding and support as well as various other factors, execution timing proved to be crucial. Execute too early and adoption will be a problem, execute too late and you miss the opportunity to create value - First Principle Thinking vs Iterative Thinking vs No Innovation.
Although data is extremely important, context proved to be even more important. Understanding which processes create, affect, consume and stimulate data it then creates a clear view. When paired with business understanding it creates invaluable context. Although science can provide us with insights never seen before we need to understand it in the context of the business. A good example presented at the event was the consumption of chocolates in countries that have a high level of Nobel prize winners, which if interpreted incorrectly, can lead to the assumption that eating chocolate makes you more likely to win a Nobel prize. The science might seem trivial but evaluating relationships using statistical correlation is where context plays a critical role.
AI was discussed in some detail, and although there are companies moving at a rapid pace to leverage this as a competitive advantage, there are still many hurdles keeping AI from truly coming into its own right within organizations. There is a myriad of case studies indicating what can happen when AI goes awry not helping the case for investment in this space.
And then, what would any event be without a presentation mentioning the 4th Industrial Revolution (4IR). The complete silence in the room was astounding as Rob McCargow covered the top 10 skills that will be required to thrive in 4IR.
A quick recap:
1. Complex Problem Solving
2. Critical Thinking
4. People Management
5. Coordinating with Others
6. Emotional Intelligence
7. Judgement and Decision Making
8. Service Orientation
10. Cognitive Flexibility
The name might just be the problem for most, as the debate has not reached a pivotal point just yet. Is the 4IR a legend or just pure fiction? Some strong words were used to describe feelings about this in presentations to follow and in offline discussions. One common reality that everyone could at least agree on is that something coming, we just don’t know what the extent of the impact is going to be.
You need to know that “complacency is not your friend”. Moreover, you must “continuously adapt” and seek “how to learn”
Expectations and reality are not the same thing, and any new endeavour will be challenging. Straight paths and quick solutions do not exist for the AI and Data Science problems we are currently facing, and it will get more challenging in the time to come.
The industry has created this expectation of a person that is so skilled that you will only find them at the end of rainbows or in fairy tales. The problem is that we get so caught up in trying to find this individual that we loose sight of the value that could have been created in the same amount of time with much less effort and individuals who comparatively are much less skilled.
The question then, how do we balance this equation?
CS (Computer Sciences) + M (Mathematician) + DE (Domain Expertise) + CT (Communication Technology)
Catchy phrases such as “Data Science is a team sport”, and “Know the rules to break the rules” seems trivial at first, but does hold a lot of water when considering all the information shared by the leaders in these areas over the last two days.
Talent management was the golden thread throughout all the presentations with questions such as?
· Where do we find these individuals?
· How do get funding to appoint these teams?
· How to we retain them?
· How do we fill the talent gap?
· What type of person am I looking for?
Ashley Kramer shed some light on the hard truth behind analytics in the real world through some impressive, yet scary stats.
· 51% of time is spent looking for data
· 50% of data science models never get deployed
· 28% of individuals share analytics via email
We had a look at how “The Stack” should look and some of the challenges that everyone faces on a daily basis – specifically the analytic choke points due to disparate processes.
Then we had some more scary stats:
· 6 Billing hours spent working in spreadsheets
· 26 hours per week working in spreadsheets
· 8 hours per week repeating the same data tasks
All of this leading to $60 Billion dollars being waster on analysts doing repetitive manual work in spreadsheets.
So how do we thrive in this new analytical world where we have all these challenges and opportunities?
· Commit with conviction
· Embrace the new
· Unleash your talent
Moreover, Culture and Change Management were identified as some of the non-negotiable cornerstones of implementing and embedding analytics within the organization and to ensure a higher adoption rate.
Nevertheless, there were more questions and although some them were answered; most were left unanswered as we all left contemplating our own existence.
Innovation vs. Maturity
Based on some of the statistics a typical example showed that 87% of organizations have low BI maturity (Gartner, 2018). This However, posed another question - Does maturity matter and which aspects of maturity should we be focused on?
Business is demanding value from data:
· FASTER insights
· TRUSTWORTHY insights
· DEEPER insights
· BROADER insights
· PREDICTIVE insights
· INNOVATIVE insights
· BETTER decision-making
· AUTOMATED decision-making
The delta force, when the force of Business is met with an equal or greater opposing force – IT. However, IT in itself faces numerous challenges around data. These are:
· DATA GOVERNANCE - DQ, MDM etc.
· DATA DEMOCRATIZATION
· ECOSYSTEM – expands – IoT, Cloud, Data Lakes, AI etc.
· BIG DATA – Velocity, Veracity, Volume, Variety
· AI/ML – embedded in business processes
· IT SKILLS SHORTAGES
· GROWING IT COSTS
The value of data throughout the data lifecycle was a key take away. The organization’s ability to execute on Data Science and AI, strongly depends on their ability to make data available at the right time, with the right level of accuracy in order to derive value. By means of incorporating some basic data management techniques you can effectively manage fallouts and ensure better overall delivery of Data Science projects. In addition, this introduces the concept of adaptive data quality where data is self-healed based on principles and algorithms built into the data processing. Understanding your data provides a good basis to explore and expand your deeper type of insight capabilities such as AI and Machine Learning.
· Business is demanding access to the right data, faster.
· IT is facing some serious challenges
· Streaming and Batch technology convergence is happening
· IoT, mobile transactions and fraud analysis are driving event streaming
· Streaming and event sourcing architectures together with AI/ML and adaptive data quality can provide intelligence at much earlier stages in the data lifecycle.
In some organizations the use of BI and illumination of data and trends seem to go much further as the capability matches the maturity and business requirement, as we saw from F1 and their journey towards delivering better insight into key business drivers. Indicators prove that it is not always about the latest and greatest, but that some organizations have solved massive business problems through the correct use of basic Business Intelligence and Dashboarding techniques.
The Data Strategy and the Role of a CDAO, or equivalent, was a hot topic this year’s event again. South Africa still have less than 40 official Chief Data Officers as the role still needs to mature and show value creation. Through the effective use of newer techniques in Data Management and Data Science the adoption will happen much faster.
The data strategies, and approaches to creating them, were also worlds apart. Although some were based on DAMA, others were based on homegrown internal frameworks.
The question remains, where do you start and what should the focal point be for you to be successful?
1. One size does not fit all
2. Your organization’s maturity will determine what needs to happen
3. Do not try and boil the ocean
4. Share with people and test your assumptions
5. Do not create a Data Strategy without involving both IT and Business
6. If all else fails, use POPIA and GDPR as a driver
In summary, we know a little more than what we did last year. No one has won the data battle just yet. Chief Data Officers are a step in the right direction, but they need support - one person will not save everyone. Technology should be the very last thing you consider, too many organizations are ending up with state-of-the-art solutions and derive no value from them. Do not throw BI out of the window just yet, as many organisations are solving real problems through applying some basic techniques.
And all these things still come down to one subject – DATA. Data needs to be managed, maintained, governed and be at the core of all activities within any business. This reminds me of a quote from the Heroes TV series, “Save the Cheerleader… save the World” can almost be seen as “Save your data… save your organization”.
Looking forward to next year and special thanks to all the speakers who shared their wisdom with us, you really added value to the Data Community.