Menu
Menu
CIO guidebook: Will investing in machine learning help your organisation become smarter?

CIO guidebook: Will investing in machine learning help your organisation become smarter?

The six factors to consider when deciding to invest in machine learning




The dynamics of competition are changing for organisations. Customers are better connected to the internet more than ever and expect so much more from corporations. Products are becoming more integrated with technologies such as cars. New “micro multinationals” are born global, and are entering marketplaces with minimal costs. Technology is more affordable and design sophistication has significantly improved. The combination of these factors is leading to an exponential rise in volume, variety, variation and velocity of data.


Organisations will need to respond to this new dynamic. Machine learning provides us with an opportunity to respond to these new challenges. Large technology giants such as Google, Microsoft, Facebook, Amazon and Baidu have invested up to US$8.5B in artificial intelligence (of which machine learning is a field of study) over the last few years, signalling a strong future in this field of technology.


In general, machine learning is the ability for computers to learn without explicitly being programmed to do so. The machine learns through applying algorithms to data. The more data, the more machine grows in knowledge.


So, is machine learning right for you?


Are you prepared to invest or increase your investment ?


Here are some questions you might like to answer to help you in deciding, whether to adopt/ accelerate machine learning to uplift your organisation’s smarts.


A. Is your future customer base from Generation C?

B. Is your future product set expected to increase in complexity?

C. Is your organisation’s resources underutilised or unable to keep up with demand ?

D. Is your competition local and global?

E. Is your organisation susceptible to IT driven competition?

F. Is your organisation’s data quality poor making insights hard to achieve?

The next wave of technology is occurring and organisations, whether they like it or not, are right in the middle of it. Unfortunately, organisations will not have the option to sit out this race. Technology is in the spotlight and current technology stacks will fast become an anchor to organisation performance.

Bernard Seeto


The following sections provide further responses on the above mentioned questions.


A. Is your future customer base from Generation C?


Generation C (for connected) are people born after “digital natives”. Unlike the “digital natives” where they have adopted technology, Generation C were born (after 1990) into the digital technology era. They will become the new generation of customers that most organisations will look to market their products and services. They are characterised by:


1. Generally having a smartphone within a metre proximity to themselves so that they can be readily “connected” to the internet;

2. Connected to internet 7 X 24 X 365 basis;

3. Rely on internet as a main source of “truth” on matters such as information to purchase a car;

4. Seek instant gratification and lack patience, such as abandoning online videos if they take more than 5 seconds to load up;

5. Willingness to trade-off data privacy for a better price on services / goods such as use of telematics on driving behaviours for a better insurance premium;

6. Comfortable generating / sharing data on themselves / activities using audio, text, video, photos, positional data.


There are globally 2B smartphone users and 3.5B internet users, and it is anticipated these numbers will grow into the future. As a result, there will be a significant amount of data created/ produced. The variety of media from text to video will provide many organisations with a rich context of customer behaviour. However, the ability to convert into insight in a timely manner will not be achievable through traditional approaches, such as, people reviewing data warehouse reports that are days / weeks/ months old.


Machine learning when combined with other technologies such as data lakes will be able to evaluate, translate and identify deep relationship patterns. Unlike traditional approaches that will require granular programming to process data, machine learning uses the data to provide the insights. Machine learning can be used to provide heavy lifting of relatively predictable aspects of customer behaviour to help organisations better target customer needs.


Machine learning will be able to help organisations to better understand Generation C customers behaviours through evaluating their data crumb trails over the internet.

B. Is your future product set expected to increase in complexity?


Product complexity is defined as the number of products in a portfolio and or the product design sophistication, that organisations manage / service.

The growth in number of products is driven from:

1. Diversity of customer demands;

2. Nature of distribution channel and the level of product optionality required / not required for channel. For example products are simplified for internet channels;

3. Uncertainty of product profitability so unwillingness to cull potentially poor performing products.

Product design sophistication will increase over time with the advances in technology. The improvements in processing power, technology miniaturisation and availability of wireless connectivity has meant products can / will have technology integrated into them - “smart products”. It is anticipated these smart products or internet of things will reach 21 B connected devices by 2020, so product complexity will continue.

According to HBR, smart products will be monitorable, controllable, optimisable and autonomous. The product smarts will have a heavy reliance on machine learning to process the external data being received to optimise performance, personalise to user’s specific needs or performance without human intervention. For example, Tesla/Google autonomous vehicle uses machine learning to enable cars to be driverless in commuter traffic situations.


The challenge for an organisation’s sales and support staff is to handle the breadth of products and provide depth of expertise for product features, consistently to customers. This is achievable with automation of processes and the augmentation of machine learning to prepare initial responses to customers, or provide recommendations to customers on products. Machine learning can thereby help to increase throughput and productivity of staff for an improved customer experience.

Machine learning can help manage product complexity through tracking and acting on behalf of an organisation without human intervention.

C. Is your organisation’s resources underutilised or unable to keep up with demand?

In most organisations there is the challenge of matching supply of skills/resources to the demand from customers. Especially in service industries where it is difficult to create an inventory of people with readily available skills to provide point of need support. Meeting customer demand expectations is necessarily balanced against pressure on cost baseline.

Machine learning can be applied to help predict demand, optimise resource allocation and also incorporate knowledge (intellectual property) from subject matter experts to provide better context, and share experience. Machine learning can be combined with other technologies such as workforce scheduling tools to provide more effective resourcing models.


In areas such as cybersecurity, where hackers are potentially continually attacking an organisation, when a security breach is closed and hackers look to create another breach. A never ending cat and mouse game for organisations which can distract from the core focus of the organisation. Machine learning can be deployed to review patterns of usage and trends to help augment security team activities/vigilance, act on their behalf to negate security breaches.


There is the temptation to believe humans will be are replaced with machines and solve organisation resourcing issues. This premise is acceptable for routine, predictable and rules based work parts of people’s activities. For unpredictable/creative activities, machine learning will be useful to provide assistance.


The following framework provides a basis for determining utilisation of machine learning. The framework highlights for low data complexity (highly structured) combined with low work complexity (routine and predictable) then automation with machine learning for activity is achievable – machine autonomy. For high data complexity and unpredictable/judgement based work, it is best to augment decisions with machine learning.

Machine learning will help organisations to better predict demand and optimise resource allocation to reduce under / over utilisation of staff.

D. Is your competition local and global?


Globalisation is increasing and is being fuelled by the digital economy. Most startup digital companies are born global and are “micro – multinationals”. Through the use of digital platforms such as eBay, Alibaba, Amazon, they can think global while acting locally.

According to McKinsey, there has been a 45 times growth in global data flows from 2005 – 2014 reinforcing the trend of companies doing business across borders.


These digital global born companies are able to enter markets and acquire customers with relatively minimal costs compared to incumbent organisations. Further, while local companies may have local knowledge, the increasing availability of open data will accelerate global company expertise in local market place. Some of the types of open data available are: government, social, economic and geographical data.


Machine learning can utilise open data to identify patterns providing further insights to organisations. When combined with a digital presence and cross referenced against a customer value model, such as Bain and Company customer value pyramid, there is an opportunity to create an initial value proposition to new customers. The economics for market entry has become more cost effective. A local example is local providers of digital content such as TVNZ versus global provider such as Netflix.

Machine learning when combined with digital presence and open data will enable organisations to enter and compete cost effectively locally and globally.

E. Is your organisation susceptible to IT-driven competition?


Over the last few years there have been a step change in the affordability and capability of technology. Computational processing costs and storage costs have rapidly decreased. While there have been significant improvements in design of algorithms and learning systems as well as underlying architecture and infrastructure. These drivers are helping to fuel advancements in machine learning. Technology organisations such as Google, Microsoft, Facebook, Amazon and Baidu have invested up to $8.5B USD in artificial intelligence (of which machine learning is a field of study), reinforcing continued progress / development.


One of the biggest factors in refining machine learning is data. Large data sets are required to help machines learn. The lead time to gather data useable for machine learning can be significant. Hence a requirement to start early on your organisation’s machine learning journey is critical. Unlike the internet technology wave where late followers were able to catch up and gazump early adopter organisations. Machine learning reliance on data means early adoption is important for organisations so as to not get left behind.


So when taking into consideration this technology trend and combining with:

1. Generation C customers creating vast amounts of data in all shapes and sizes;

2. Products being augmented with technology- internet of things;

3. Competitors (local and global) creating digital presences utilising open data to accelerate expertise.


The next wave of technology is occurring and organisations, whether they like it or not, are right in the middle of it. Unfortunately, organisations will not have the option to sit out this race. Technology is in the spotlight and current technology stacks will fast become an anchor to organisation performance.


The new technology stack provides seamless connectivity between smart products, customer generated data, organisation enterprise technology platforms in a secure environment. The following is a conceptual model of the new technology stack:

Machine learning provides organisations with the agility to respond to IT driven competition.

F. Is your organisation’s data quality poor making insights hard to achieve?


There is a positive relationship between quality decision making and an organisation’s earnings. When organisations are under pricing and or facing profit margin pressure, quality decision making becomes more critical. Decision making is underpinned by quality data. There are up to 20 dimensions when defining data quality such as accuracy, completeness, timeliness and believability. However, when using data the quality is only known in retrospect, which is generally too late to make a difference. For the purposes of this paper we will focus only on accuracy dimension when commenting on data quality.


In most organisations data quality is weak and is generally considered an afterthought versus forethought. Most initiatives will focus on product features and not necessarily the type / level of quality of data required to be captured. There is inertia in most organisations especially when organisations are making acceptable profit margins. Revenue generation and cost management are through subjective considerations versus data driven decision making. Unfortunately, when external factors change for the worst, this subjective approach provides no benchmark / reference to review performance against.


Machine learning will not overcome your data quality issues or decision making approach. However, when machine learning is combined with data and subject matter expertise, machine learning algorithms can be better tuned.

Applying an iterative approach, a benchmark can be developed /refined between data and algorithm. Data quality improvement / corrections can be built into algorithm to eventually provide better input to decision making.

Machine learning provides organisations with a faster approach to identifying data quality issues and potentially resolving.


No silver bullet


Machine learning is not a silver bullet for an organisation’s competitiveness. When combined with automation, customer and product data and subject matter expertise, machine learning provides a useful way to unlock latent capability of your organisation’s data assets.


Unlocking an organisation’s data value increases the “smarts” of your organisation.

For organisations where their future will rely on/interact with: Generation C customers, smart products and compete with micro – multinationals then machine learning will definitely help your organisation to become smarter.


If you have answered yes to the majority of questions in this paper, then please seriously consider investing in machine learning. If you have already started investing in machine learning then consider “doubling up” on your investment.

Why act now?


According to Gartner’s hype cycle, machine learning is at its peak. It is anticipated it will take two to five years before reaching the plateau of productivity. Most organisations in this time will be laying foundations for the new dynamics for competition. That is, establishing new technology stacks, developing machine learning algorithms and most importantly gathering the data (internally and externally, structured and unstructured forms). Effectively building the capability and expertise to be fit for future competition.


Some examples of results companies have achieved to date with machine learning:

Are you ready to start or accelerate your machine learning journey?

Bernard Seeto is an experienced financial services industry business technology leader. His roles have included head of solution delivery at IAG New Zealand and executive manager customer development at Suncorp. Reach him at bernard.seeto@gmail.com.

Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the CIO New Zealand newsletter!

Error: Please check your email address.

Tags leadershippredictive analyticsIAGCIO100DXchange managementmachine-to-machine (M2M) technologycareersdigital disruptionFinancial Servicesmachine learninganalyticsfintechdigitalGartnerartificial intelligencemachine to machineBernard Seeto

More about CustomerseBayFacebookGartnerGoogleIAGMGIMicrosoftNetflixSeekTechnologyTesla

Show Comments