Topic 2: Information and Decision
Making
Information systems have had 3 strong impacts on business and the general
environment.
1. Automating manual processes
2. Gathering information
3. Transforming information
Automated accounting done by credit cards and tills to make the accounting
task more efficient
When you automate things you get a huge amount of information.
How can we use that information?
Information is power; to possess information that you can make a decision
on gives you an advantage. Most business decisions are made on incomplete
information. If you have more information than a competitor you have an
advantage.
PaddyPower is a great example, they know exactly what your patterns are
and how to give you bets to make the maximum amount of money on you.
If you’re making them a lot of money they may even offer you special
promotions to keep you invested. This information is generated
automatically, not specifically, simply by automating gambling online.
The process ends with transformation.
The information gathered by automation completely transforms the way
that a company works.
Many good examples in the slides on how automating, gathering
information and transformation has taken place
Not every technology will transformate, but every technology promises to
transformate. Evaluating these technologies and deciding where to make
your investment is crucial.
From Moneyball Clip:
“They are asking the wrong questions”
Leading to asking the wrong questions, leads to using the data in the wrong
way and thus misguided decisions.
If somebody presents something like “Ireland’s best towns for work-life
balance” an essential question to ask if what data points are being used and
what model is being used with this data?
- Did the model weigh one data point over the other? House prices for
example are given weight over Cafés? The example in class did not
If the data is bad, and if the model isn’t good either, then you shouldn’t be
happy with the recommendations of the model.
Data is everywhere, and data is not the same thing as facts. Facts are much
harder to come by than data. While data seems to promise objectivity, it
actually requires analysis which is constant and subjective evaluation.
Models
Modelling : conceptualising a problem and abstracting it into a quantitative
and / or qualitative form (ie. using symbols/variables)
- Abstraction - making assumptions for simplification
- Tradeoff (cost/benefit) - more or less abstraction
- Modeling - both an art and a science -> no model is 100% accurate,
and making the trade-offs on what is and isn’t included in the model
is both an art and a science.
User Interface
User Interfaces employ real-time analytics to give the users real time
insights into various aspects of how the business is performing.
Capabilities of Dashboards
- Drill Down -It gives you the ability to go from a macro figure to a
micro figure. It localises problems and hence makes it easier to deal
with them.
The key development in user interfaces is the mobilisation of the UI. You
can now run a whole business from mobile through the data available on it.
Jeff Bezos, a man who arguable transformed the whole retail market using
information systems to do so, is saying here that we are still at Day 1. Why?
Because information systems help us get the knowledge we need to make
better decisions, see problems more clearly, and find ways forward that
were obscured before. Leveraging information systems, new ways of
business are possible (like Amazon), and thus the information system is
transformative for the whole industry.
However, simply offering and investing in a new technology does not make
you immune from the question of how you’re going to utilise and
commercialise the new technology. You must always state the case for its
relevancy, and its realistic potential to automate, inform and transform
business.
What is information?
A byproduct of computer systems is information. It is generated regardless
of your use of it. This information can be used for decision making.
Can a combination of Data & Models & User Interface create Information?
- If it doesn’t help the decision you’re making, it is just data, but if it is,
it’s information.
Can an information system contain knowledge ie. an answer to a how
question? There is a case to be made for yes, as the systems have best
practices built into them, and these best practices can help explain how
something is done, how things happen. Wisdom still resides in human
thought, as the why question is one that would be very hard to find an
answer to in a system itself.
Big Data has three main important qualities with respect to the actual data:
- Volume - the sheer amount of data that is available (magnitude)
- Variety - the many different kinds of data it offers; not just text and
numbers, pictures, videos, sensor data, audio etc. are also useful
sources of data you can algorithmically analyse.
- Velocity - the speed at which this data can be gathered and analysed
Organisational Knowledge
Companies produce a lot of information in reports and such. Information
may exist, but you cannot retrieve it, or are not aware of it.
However with the rise of generative AI may begin to solve this issue.
Characteristics of Good Quality Information - ACCURTE
Accurate
Complete
Cost-beneficial
User-targeted
Relevant
Timely
Easy to use
In management you need to make a good decision with incomplete
information. So the goal must be to create a system to give you better
information than your competitors, as this will lead to better decisions.
Information doesn’t have to be completely accurate or complete to be
useful. The trick is is to make them as accurate as possible, while not
sacrificing the time-sensitive nature of decision-making.
When does data create a competitive advantage and when does it not?
Even when customer data does confer a competitive advantage, it gives rise
to network effects only infrequently. And that advantage may not last.
- Data does not always give competitive advantage.
To understand the edge that data-enabled learning can provide, companies
should answer seven questions, which examine the value of the data;
whether its marginal value drops quickly; how fast it becomes obsolete;
whether it’s proprietary; whether the improvements from it can be easily
imitated; whether they enhance the product for current users, other users,
or both; and how fast insights can be incorporated into products.
Useful data enables good decisions.
Models and Assumptions
It is very important to recognise the fact that every model is based on
assumptions. Recognising and weighing up these assumptions is key, as this
can let us see what the model does and does not consider.
To understand how the model works you must figure out the assumptions
the model makes. If you have a model you trust or want to trust, you must
figure out how it works. Only then can you make the decision of whether or
not to use the model
Why did Chat GPT give a good answer for the Supply and Demand model,
and a poor answer for Porter’s Five Forces? Figure out how Chat GPT
works
Decision-Making - Intelligence Phase
- Scan the environment, either intermittently or continuously
- Identify the problem or opportunity situations
- Problem is the difference between what people desire (or expect) and
what is actually occurring.
Symptom vs. Problem - you can take action on a problem, but with a
symptom you don’t know what the root problem actually is yet.
- Timely identification of opportunities is as important as identification
of problems
If you want to be desirable for an employer, if you want to make a
difference, you must focus on the intelligence phase.
Information management is understanding how information technology
fits into business, let data scientists focus on the technology itself.
Decision Making & Information
Systems
Where should the Information System leave off and the decision maker
begin? <- fundamental IS management question
- How much agency do we give to computer, IS systems to make
decisions?
However combining the power of information systems with the intelligence
of a human can be the best way to do things.
Orgnasiations that learn from AI have 3 essential characteristics
1. Facilitate systematic and continuous learning between humans and
machines
2. They develop multiple ways for humans and machines to interact
3. They change to learn and learn to change
- Machine learning, in simple terms, is learning and changing based on
updated data
The more you give control to the information system, by definition, the
more assumptions you build into the system. This means you get a little
further from reality, if those assumptions are not always correct, especially
over time.
Can or Cannot?
Since we cannot make a prediction based solely on past events, how can we
use data to make decisions that affect our future, if we only have past data?
The oversight of the person is incredibly important :
The algorithm falls out of use due to previous, non-applicable assumptions
Example : Covid causing huge shocks to the system
Predictive analytics is very good at predicting the future, if the future is like
the past.
Regardless of what algorithmic model
you make, you must assume in the
future it will break and be unfit for
purpose. No model or system is future-
proof.
Prescriptive Analytics
The decision-maker is removed, the system makes a prescriptive decision
based on descriptive analytics. The business decision is inside the
machine/system.
The decisions the system have made are unknown to you. You don’t know
why you’re charged the insurance premium you’re charged. The airline
ticket you’re charged. The assumptions used for evaluation of a system are
hidden.
What happens if a system begins to misdiagnose, especially when you don’t
know what assumptions have begun to malfunction, because of the black
box.
Making ethically capable algorithms, aka mastering algorithm management,
may give you competitive advantages.
But how do we know that systems work as intended?
Who might they be unintentionally harming?
Making algorithms more ethically coherent can minimise unintentional
harm.
Algorithms are not neutral, they have built in assumptions. We have to
learn to evaluate and manage them to make successful algorithms.
The above example is a very clear example of what happens when it goes
wrong, what about more subtle ways things can go wrong, and its not as
clear?
Prescriptive analytics are very good in the short term. Narrow domain, not
broad domain. AI is a poor name, as it implies broad intelligence in a broad
domain, where it is realistically a broad intelligence in a narrow domain.
Banking has become a data science, assumption and model driven world.
Generative AI
What is General AI and Generative AI?
What are the differences?
Normal AI is the stuff like the fraud banking system mentioned above.
Generative AI is AI that produces text, images, video and audio, processing
language and mimicking it.
AI will take certain jobs, but it may create new jobs. AI will augment certain
tasks to enable people to do them better.
The goal of company use of AI is to fine tune AI to access company specific
data. It is used to draw answers from relevant user manuals and documents
to give decision-makers the information needed to make decisions.
Basically, making technology accessible to non-technologists, and AI is the
tool that is designated to allow this (planned anyway).
Personal Information Management
How do you personally use information systems?
How do you retrieve information that you’ve already gathered when you
need it?
Your ability to learn, your effectiveness, is heavily influenced by the way
you build a personal information system. This is about finding, keeping and
organising information.
Misinformation refers to inaccurate information ;
- Can be characterised by fake or misleading information that spreads
unintentionally
Disinformation refers to deliberately deceptive or misleading information
- Disinformation is generally associated with sophisticated efforts by
organisations such as government, military and political groups
You are in control of where you focus your information, and this is the
choice made when constructing a personal information system.
You are the editor now, you are in charge of sniffing out misinformation
and disinformation. It is all too easy to just gravitate towards the things you
already believe and live in an echo chamber. By putting a personal
information system in place you can get control over this.
We have a requirement to be editors, and it’s becoming more and more
difficult to do our jobs.
You can no longer trust at face value the reviews you see on the internet.
You need to take actions to become an editor of these, with actions such as :
- Who is the author?
- How current is the information
- Truth? From what point of view?
- Unbiased ? Strive for informative rather than persuasive sources
- Privelege? Who is missing from this conversation?
Find a number of sources you consider broadly true.
You’re going to keep going back to these sources to stay up to date.
When you find it, how will you keep it ?
How do you keep the electronic information that you now source.
If we operate on the premise of an information overload, we rely very
strongly on the technologies that store this information for us ; we are
simply not capable in the same way machines are capable, of storage. And
nevermind memory, but recall.
Keeping information is a serious topic. Backing up information is a good
idea, and checking whether or not the backup works is important if you
want to rely on it.
However keeping information is actually quite easy. How do you organise it
so you can easily find it when you need it again?
You can’t formulate these insightful problems if you can’t find or recall the
information that would illustrate it to you. This is why personal information
management systems are so important. What you don’t know, you can’t
think about.
How do you communicate what you know?
- When you make a digital footprint, no matter what your personal
desire, it will remain.
Information about You
What are the pros and cons of technology keeping and using information
about you?
How do we build a profit of a person based on how they access and scroll
through the internet.
A huge amount of personal information about you is collected and by using
these apps, it can build a very detailed profile of who you are as a person.
This raises security concerns.
- How are they storing this data?
- How are they sharing it?
- How are they using it?
Facebook “jumped the Chinese wall” and stopped from being an app
working in isolation and began to communicate with other apps to gather
even more information on you.
Can you use this kind of information collection for your benefit?
This suggests that you can.
Gathering data about yourself can help you see not are you in the average
heart rate, not average blood pressure range, but how does your heart rate
now compare with your historical heart rate. Keeping your own medical
history through this information gathering can have huge positive
implications.
It also allows for personalised medicine, and medical devices / treatments
that work for individuals, not averages.
Who owns the data and who has access to it?
Big Brother is watching, in this case, and the question is whether you have a
right to say no?
The company you sign up for such as FitBit, is not necessarily the company
that will have access to your data. The company you sign up for may no
longer exist, but the data it collected will, and someone will access it.
The whole idea of anonymous data, becomes questionable.
Mobile and Wireless Business
How does your device know where you are?
Don’t think that mobile means mobile phone. Mobile means mobile, it is
not stuck in one place, and this includes many devices and technologies.
Mobile Computing : Devices
- Mobile phone (including smartphones)
- Tablets
- Wearables
The smartphone revolution is under-hyped, more people have access to
phones than access to running water. We’ve never had anything like this
since the beginning of the planet.
What you’re carrying in your mobile phone is an incredibly powerful
mobile computing device, and that makes many things possible.
Mobile phones are much more than just one device, however. It is
something that has taken over many businesses and technologies.
It is a technology that is not hampered by a lack of landline connection, as
its capabilities are mobile.
Internet of Things (IoT)
Also called “ubiquitous computing” that is embedded in the objects around
us
- Smart homes, smart appliances.
Forget about computers and phones, everything that connects to the
internet is part of a connected web, an Internet of Things.
Anything that you can put a chip into, and connect to the internet, is a smart
device.
The question is if this is a bad technology, or if its development was bad
timing? Will there be a time that this technology will be widely adopted. #
You won’t need to look at a screen anymore. You put on a pair of glasses
and you see a display in the corner of your eye. Mobile is not just a mobile
phone, and the sense that we will continue to depend on them forever is
perhaps a little naive.
Jewelery with a small receiver and panic button that sends your location to
your contact in case of an emergency.
When you think of mobile, again, don’t confine it to a phone.
There will be no keypad, there will be no mouse. The attention of your
brainwaves is what the device will do. You’re becoming augmented with
technology that is implanted.
Where are we moving with mobile technology?
This connection of devices creates a smart web of devices that has the data
to run things better, which leads to the idea of smart cities.
There is no such thing as a mobile phone network, it’s a mobile device
network. The forecast shows that there will be more devices with sim cards
that are not phones than phones in the future.
The bus communicates via sim card where it is, and various data points to
other machines.
The idea is to get a lot of data, figure out what is good data, and make better
decisions with the data collected by the internet of things.
The point is to show that mobile is much more than just phones, and the
implications for innovation provided by this technology is huge.
Passports for example, use RFID. Airport employees for example can use
RFID readers to locate anyone in the airport by the tag in their passport.
High end retail brands can also use them to prevent theft.
Industry 4.0 gives the ability to control robots wirelessly through the Cyber-
Physical Systems (CPS) and the (IoT)
To wrap things up, in the future there will be many billions of devices that
will all communicate on one network. This will facilitate unprecedented
automation, innovation and especially transformation.
The App Economy
Money in apps is made primarily in advertising. A whole industry has
emerged out of nowhere since the advent of the mobile phone (which is
only 15 years old).
Flurry handles serving you ads, targeting ads, and handling the micro-
income that happens when you watch someone’s ad.
This revenue did not exist 15 years ago.
The amount of change is mind-boggling,
especially with the speed of change.
This shows how they actually make money off of apps.
More money is made
in smartphone games
than any of the
massive pc or console
game markets.
Before smartphones,
companies thought
that perhaps other
consoles or pc would
be the most
competitive offerings,
because you own the
platform. But now
there is a platform
that is universally available, and this changes the dynamics of the industry.
The platform becomes important.
Facebook knew, and it had the data about who the user was, and they did
absolutely did nothing, and went out of their way to purposefully do
nothing.
High fixed costs, low
variable cost. You
have costs making the
game, not much in
delivering the game. It
keep generating
money by itself without needing much input. This is an example of a “hit
economy”
Attributes of Mobile
Mobility
Personalisation
Localisation
Convenience
Interactivity
Ubiquity
For every business there are pain points. The paint points of shops is time at
the checkout line. This is where technology can help.
Cyber Security
Certain cyber-criminal groups hide within certain nation states, with the
benevolence of the nation state. The state does not stop their actions, and in
certain cases, the nation state employs the cyber criminals to act on their
behalf. This can be nation against nation, or nations against individuals.
What you really need to be worried about as an individual is cyber
criminality, sometimes linked to nation states. As an organisation you also
need to be aware of the insider, who is already inside.
This shows you who you’re facing.
Ransomware and identity theft is the most prevalent kind of cyber attack in
recent times.
A virus that has spread to you whole system, and you have no easy way to
remove it.
It’s not common on an individual level, but on a company level yes. You’re
only one click away from disaster. You click on a link, any link, you may
have no going back from inviting a cyber criminal into your system.
Malware follows where users are
Social engineering is conning you into believing something else than what’s
actually happening. Getting tricked into believing who you’re talking to is
different than who you’re actually talking to.
People are the weakest link in cyber security
They are looking for your credentials first and foremost.
The key to all of these is the credentials or verification that is being sought.
What is blockchain?
(Wired blockchain video)
As each transaction occurs, it’s put into a block.
Each block is connected to the one before and after it.
Each block is connected to the previous and proceeding block.
Transactions are blocked together, creating an irreversible chain; a
blockchain.
The chain shows you time and sequence as well as individual events.
Why is it such a big discussion point?
Cyber currency as a technology uses blockchain.
But it is not blockchain; it’s a use of blockchain.
- It is a distributed leger. A few people have the ledger of how the
blocks are distributed.
- You have to have permission to access and add to the blockchain.
- It’s secure, but can still have credentials stolen
There isn’t a single ledger - the key benefit of crypto currency.
You can’t change a single ledger that will change all the accounts.
Blockchain has multiple uses (insert slide).
- Blockchain does not equal cryptocurrency
- Smart contracts
- Identity registry
- Database registry etc.
The examples show the diversity of how blockchain is used.
Allianz uses it for claims settlements, as it reduces processing time. It’s
about efficiency and security
JP Morgan uses it for smart contracts. It’s essentially a secure, permissioned
contract space. Better than the smart contracts they had before as it gives
them more security.
Boeing do it for digital aircraft and airplane parts. Blockchains are being
used in traditional companies for companies to know who is changing what,
who is accessing what etc.
De Beers diamonds create an image of each diamond and then every
transaction on that diamond goes on that blockchain. From the point of
origin to the final sale is all on the blockchain at each step.
We could blockchain cattle. How do we authenticate that the beef coming
out of Ireland is grass fed? We could create authentication documents
digitally attached to physical goods.
The point is the implementation of the technology doesn’t have to be for
sophisticated stuff.
NFT - Nonfungible tokens.
The image on the slide was sold for $69.3 million.
A NFT is essentially a blockchain that authenticates something. It’s
authenticating that there’s only one original digital image for that image on
the slide. It’s a digital artifact and the NFT proves that the owner has
copyright and ownership of the artifact.
Blockchain hasn’t taken over everything, but it has specific uses.
Developing blockchains is incredibly intensive on computer resources
(building out new encryption bits). Some people think it’s a very inefficient
use of computer resources.
Need to know what it’s uses are and what it is.
- Permissioned
- Encrypted
- Distributed
Botnet- robot network- someone controls a lot of devices and can turn
them on at will and direct them at a target system.
If your computer is infected by malware it is sometimes done to turn it into
a botnet. The might not use your system until they need it for an attack.
1 million machines you can turn on at any time and “recruit” for an attack
If you want to actually attack a system.
A logic bomb - stuxnet is an example. It only deploys in certain situations.
Stuxnet only deploys with nuclear reactors.
It is no longer to think that technology solutions are sufficient to protect
data.
IS Development & Acquisition
How do we build and buy systems?
A report just out from CEOs what are their top priorities.
- Transforming company IT is important, and very real for companies.
They see the potential of technology to do things better, faster, more
competitive. It’s higher than new products, and sustainability strategies.
Where are they going to focus?
It’s all about organisational capability, and how we get IT systems to give us
that organisational capability
How does a new system come to be in a company? Who decides?
Your IT system should come from the strategic plan, not vice versa. If
you’re a low cost strategic plan, you spend money on a system that saves
you costs. If you’re a niche strategic plan, you spend money on a system
that supports the strategy. The business plan drives the planning.
How does a company decide between systems? It decides on the best return
to shareholders. If we invest a million, how much do we get back? It has to
be more than a million. It has to be substantially more than a million
otherwise why not put the money in a cash deposits?
An IT system has to bring back money and profitability for the money
invested.
- Intangible benefits - goodwill for example is intangible. If you build a
customer support system that makes customers like you much more,
you can’t put a figure on that. “Customer rating” ??? It’s not a hard set
number. But IT can impact this greatly.
If you’re a telecoms company, they have spend a fortune on the old
systems. They’d have to look at 4 different systems to talk to you.
The sense is that this is high money, there’s a real challenge. The challenge
is, “Change Management”. Most IT systems aren’t about bringing in IT
systems, but bringing in change. The problem is not bringing in technology,
the problem doesn’t lie with the technology, but with how difficult the
process of change is.
All 3 of these were developed separately but are now integrated seamlessly,
and act together.
