« February 2016 | Main | April 2016 »

8 posts from March 2016

23 March 2016

More information per pixel

‘I didn’t have time to write a short letter, so I wrote a long one instead’ – Mark Twain

For effective data visualisation, the ability to say more with less is crucial. Just as writing in a verbose manner to communicate an idea often backfires, over complicating what is displayed on screen often fails to communicate a message. Charts and graphs should be clear, concise and easily understood. In many situations we need to allow for further exploration of the data, but the headline is what we need to convey, and we do this by displaying only the essential detail.

Striking visualisation does not require further detail or explanation. It’s intuitive, the image is the explanation.

The need to add more detail to every visualisation seems to be a remnant from formal education. But we no longer have to expand our sentences to meet word counts. Applied to our role, there is no requirement to add more analysis on screen. By elaborating less we can explain more.

Practically, what we are trying to achieve is more information per pixel. This is a process of simplification, not addition.

Look for what can be removed, not what can be added to ensure you get your message across.

If you would like to discuss these issues, or the impact of emerging technology or data and analytics on your industry, then contact our Data & Analytics team.

21 March 2016

8 ways to perfect your visualisations

In a world where data visualisations are being demanded by board members, clients and the media as a way of translating complex data into beautiful images, it's easy to get caught up in a frenzy of artistic creativity. But, you always need to be conscious of the quality of the work you're doing. These steps will help you remember to focus on the client, and not just the fantastic picture you've created.

1 - The Brief

The first thing to understand about the visualisation(s) you’re about to produce is the brief. The brief will be set out to you by your client (both internal and external). They’ve hired you for a reason, for a specific purpose, so it’s critical that you actually understand what that purpose is, and ensure that you constantly check back to it at every stage. This way, you’ll have a better chance of delivering something of value to your client.

2 - The Data

The next consideration is of course the data. Where is it? Who owns it? What format is it in? How do I access it? What does it mean? At this stage, you’ll need to build up a good relationship with one or more Subject Matter Experts – they’ll be your guide through the maelstrom that will undoubtedly be their data. Once you get access to it and understand it better, it’s time to explore. Keep your brief in mind at this stage, ensure you’re looking in the right areas. However, don’t let that stop you from taking note of any issues you spot outside of the brief – this could come in handy later on.

3 - The Tools

Depending on your brief, you may or may not have a choice in which tool, or tools, you will be using. At this stage it’s best to keep an open mind as each visualisation tool comes with its own pros and cons – no perfect tool exists, and even if you do have your own particular favourite, it may not be the best option 100% of the time. Additionally, keep an open mind on other tools that could supplement the visualisation, such as ETL (Extract, Transform, Load) software, mapping software etc.

4 - The Audience

Who will be your consumer? Whilst all of the things featured within this list are important, your audience is perhaps the most important one. These are the people who will actually rely on your visualisation to assist them in making business decisions, so it will need to be tailored to meet their needs. It’s good practice to have regular contact with your audience, or at least a sample of them, to understand what works for them and what doesn’t. As with any software development process, your visualisation will only be accepted if it provides the end-user with something they can actually use!

5 - The Charts

Selecting the right charts will ensure that the message contained within your visualisation is conveyed in a way that your audience will understand. There has to be a trade-off at this stage between your technical capabilities, and the comprehension levels of your audience. For example: If they’ve never seen a box-plot before, ask yourself ‘Is a box-plot really the best way of showing this?’ If there are alternatives, look into them. If it turns out that the box-plot really is the best of getting the message across, think about how you can aid your audience in understanding it.

6 - The Story

Once you’ve interrogated the data, inspected it, played around with it, understood it and produced some insightful charts, the next step is to pull all of that insight together in a compelling way. At this stage, it might be useful to convey the journey you’ve just embarked on in order to understand what the data is showing. It’s likely that your audience will not fully understand their data either, so pull your dashboard together in a way that delivers the message in line with the original brief, but also in a way that captivates the end-user.

7 - The Interactivity

One way of captivating your audience is to include some level of interactivity into your visualisation. It’s important to ensure that this interactivity is relevant – think drill-downs and filters. These types of interactivity serve a couple of important purposes: firstly, they allow the end user to conduct some investigation of their own, and secondly it should allow you to answer any new questions that may arise from the visualisation. For example, if your visualisation shows your clients that their stores in the north of England are performing worse than those in the south, think ahead to the obvious question: 'Why?' This way, when the client sees your dashboard, they can click a few buttons and see the answers for themselves!

8 - The Added Value

At every stage of creating your visualisation, there are opportunities to go above and beyond for your client with minimal additional effort. Remember, during your visualisation process, you’ll be rooting through the client’s data trying to find everything you need to fulfil the brief. You will find plenty of data that may not be relevant for this current project, but within this you will also find a lot of interesting stories that your client may not necessarily be aware of. Think about how you can share these with the client; can they be incorporated into this project, or is this a conversation for another day?

In summary

It’s an easy trap to fall in to, but remember you’re not producing a visualisation for yourself, you’re producing it for someone else. Building and maintaining a good working relationship with the client and audience (not always one and the same) will ensure you stay close to the brief, allowing you to create something sufficiently tailored to their needs and deliver a visualisation that will help them to achieve their business goals.

A good visualisation will deliver a win-win situation. The client will win because you've produced something incredible that's answered previously unanswerable questions, and you will win because you've delivered an excellent piece of work, built up some good relationships and enhanced the reputations of both yourself and your firm.

If you would like to discuss these issues, or the impact of emerging technology or data and analytics on your industry, then contact our Data & Analytics team.

18 March 2016

Big bang data - the impact of data on society in the Digital Age

by Will Leighton Senior Manager

A broader perspective

When fifteen esteemed Data Assurance colleagues visited the Big Bang Data exhibition at Somerset House, we discovered some startling facts on the impact of data on humans and society as a whole. Believe it or not, the first transatlantic cable was laid down over a hundred and fifty years ago. Since this date in 1858, the use of data gradually crept into relevance in human civilisation, until the introduction of three key inventions - the transistor, followed by the microprocessor and the internet. These applications of producing, storing and transferring data led to an explosion of data being produced, stored and transferred to all corners of the globe.

At our fingertips

Our attention spans are shorter than ever, but with the ease that one can find information nowadays it’s easy to imagine why. Take for example the ability to find out the address, contact information or background of virtually any business or individual in the developed world within a single click or Google search. Twenty years ago this would have been unfathomable, but data has made the world a small place. The benefits are huge – not only keeping in touch with friends and family is easier than ever, but the global economy is completely interconnected, with imports and exports of goods and services connecting people all over the planet.

Beautifying data

Data is also being presented in new and insightful ways, shrinking previously unreadable volumes of data into formats that can be viewed in powerful ways by people and society. Below is a single snapshot of all tweets captured in one second across the world – the information fills an entire 1000 page book.

Willblog

Into the unknown

The exhibition asks more questions than it answers, and rightly so - how are we to know what future lies ahead for us living with all this data, and what are the downsides and dangers of having access to all of it? Dangers such as mass surveillance and lack of data privacy are proliferating, governments are tapping into our data for their purposes, and cyber criminals are gaining access to our personal information and bank accounts more than ever. All of this is occurring from our willingness to create data on publicly-accessible platforms and the “Cloud”. Lastly, we should never ignore the power of human intuition and creativity - something that data will never fully replace.

If you would like to discuss these issues, or the impact of emerging technology or data and analytics on your industry, then contact our Data & Analytics team.

by Will Leighton Senior Manager

14 March 2016

Price elasticity - better ways of predicting insurance customer behaviours

Price elasticity is an area that insures have spent a lot of time analysing, not least in the context of understanding customer lifetime value and ensuring that customers deliver acceptable returns through a sufficiently long tenure. With the proliferation of price comparison websites, aggressive marketing campaigns and the promise of lower premiums by shopping around, ensuring that customers remain loyal enough for insurers to overcome new business strain and deliver target levels of profits has become increasingly more difficult. As a consequence, ensuring that the renewal premium quote offered is sufficiently attractive to be taken up needs to be balanced against the need to deliver an acceptable return.

Whether a given customer will accept the renewal quote depends on a number of issues, including:

  • Whether the customer made a claim over the policy year. A positive claims service experience will be in the insurer's favour, while a bad experience is likely to result in the loss of the customer, unless the customer can be swayed - assuming the insurer wants to retain the customer of course.
  • If the customer is cash rich, but time poor, the convenience of simply renewing with the insurer for a premium that looks broadly in line with the expiring price (and perhaps only marginally different terms and conditions) may be very attractive. Determining whether a customer fits into this category has historically involved considering the customer’s credit score, and any financial health, socio-economic or demographic information that is available. Other indicators include the time of day when interactions with the customer occur, as a proxy for their time poverty. Paradoxically, many cash rich, time poor individuals are not as price inelastic as insurers might hope – as Michael Douglas’ character in the movie Wall Street, Gordon Gekko, famously said, “a fool and his money are lucky to get together in the first place”.
  • If the customer is motivated by paying the lowest price, and does not care about the quality of service or the value proposition on offer, the renewal premium would need to be pitched aggressively (especially if the customer has already visited price comparison websites to assess the renewal quote) in order to have any chance of retaining the customer. Customers as “promiscuous” as this are unlikely to deliver the target customer life value sought, unless they are non-standard risks, say, for which the market is smaller in the first place, and so competition is relatively less fierce.

As with many other areas within the field of Analytics, a number of advances have been made in recent times:

  • Social media has been increasingly analysed for brand trust or at least positive net promotion by policyholders, especially claimants.
  • The wealth of data around quotes and their conversion rates (accessed from in-house data or through price comparison websites) and renewal acceptances and lapses have been mined extensively, often using machine learning, to try to divine the characteristics or motivations of customers in order to better predict their behaviours. 
  • Data enrichment, utilising data from grocery buying habits or mobile phone usage, for example, has provided deeper insights into price elasticity. For example, customers who purchase premium products at supermarkets or who purchase additional mobile data over and above the levels offered by their tariffs/bundled packages, might be deemed to be more price inelastic, or are at least willing to pay for a better offering. 
  • An alternative approach to predicting customer behaviour utilises innovative behavioural simulation models that aim to provide insights into decision making at an individual consumer level. While insurers have an increasing appetite and need to understand their customers – which is hampered by the relatively few interactions that insurers typically have with their customers – insurance customer preferences and behaviours are changing at the same time. Attempts to understand customer behaviours are better informed by their financial commitments at any given point in time, which are in turn fuelled by life events such as marriage, children and the payment of school fees or mortgages. Based on household profiles built from augmenting an insurer's own data at the customer level in this way, real customer decisions impacting their insurance products can be simulated. These models use typical decision biases derived from behavioural economics to provide added realism to those decisions. Using this approach provides insurers with real insights into customer segments that cannot be obtained from traditional analytics (especially when there has been a dramatic change in legislation, say, which means that the past is not a useful guide to the future). The results of individual customer simulations also build up to an understanding of portfolio level trends that differs from those assessed more traditionally.

All of the above relates to the traditional annually renewable insurance policy. As the demand for a different type of insurance coverage intensifies, most notably from the millennial generation, the concept of renewal may become more blurred. For example:

  • Progressive in the US has for some time been offering customers the option to say what they can afford to pay for motor or household insurance, and coverage is then offered to meet that price point (mostly by flexing terms and conditions);
  • For usage-based insurance coverages, insurers could make suggestions for how monthly premiums could be reduced, such as reducing mileage or the times of day when the car is driven, from a tailored analysis of the customer’s driving behaviours;
  • In the sharing economy, top up insurance or one-off insurance to drive a friend’s car becomes a very transactional offering, whereby the premium rate offered needs to reflect the fact that acquisition costs are unlikely to be recouped over a longer period.The potentially rapid pace of product evolution that digital insurance offerings could bring may well mean that understanding – let alone predicting – customer behaviours is about to become a lot more difficult.

If you would like to discuss these issues, or the impact of emerging technology or data and analytics on your industry, then contact our Data & Analytics team.

08 March 2016

Can investing in data and analytics actually save capex?

In capex-intensive industry, the time is right to think about how a data-driven approach can transform your capex planning for the better.

In a capitally-intensive business, how you spend money on key infrastructure can make all the difference when it comes to investor returns. For industries such as Telecoms or Utilities, where assets are both costly and highly complex, getting to grips with capex can prove a nightmare.

The way companies have coped with this complexity historically has been to hire vast armies of specialist engineers to determine which assets to buy and how to deploy them, while commercial teams stick to running the business and managing the products and tariffs. As a result of this arrangement, all of the knowledge and data around the drivers of capex resides with the technical teams, and the business rules governing spend exist in a black box that Finance and Commercial do not understand.

This makes it very difficult to plan. When information is not transparent, and capex is not linked to performance, everybody in the business can feel the pain.

  • When the technical teams report that they have overspent, it is easy to shoot the messenger, but often the causes of the overspend are outside their control, caused by capex-intensive sales.
  • Commercial teams suffer when they exceed their sales targets but the infrastructure is not in place to meet the demand.
  • Without knowledge of what drives capex, finance teams effectively have to treat it as a fixed cost and they cannot incentivise efficient behaviour.

In this world, capex planning happens in the technical teams. They use sparse financial forecasts and operational models to build very capable capex plans, but these plans are not linked in any meaningful way to commercial drivers such as product, sector, bandwidth or region. The end result is that financial planners are unable to see the impact of changes in product mix or other commercial scenarios on capex or returns.

However the recent advances in analytics and planning tools are disrupting the status quo. The old argument that the only tools fit for purpose are the engineers’ dedicated operational models no longer holds true. Many businesses, particularly in Telecoms, are starting to realise that combining technical asset planning with commercial models is not only possible, but enables them to fundamentally change the way they think about planning.

The value of an integrated driver-based planning approach should not be underestimated. By linking the cost of each asset to the products and sectors that drive them, a business can understand the returns across the entire asset base. Furthermore the business can start to differentiate good capex from bad and focus its commercial strategy on products with a higher return on capex. A PwC survey in 2012 identified that the average Telecom operator wastes 15% of total capex on assets that do not create value for shareholders. Driver-based planning provides a mechanism to unlock and re-invest this capex.

There now exist sophisticated data products that can link data from across the business to produce driver-based capex plans that are sophisticated but don’t sacrifice the transparency and auditability of an old fashioned commercial plan built in Excel. The historic trade-off between simple but flexible and transparent vs. complex but opaque and with a high IT overhead is beginning to disappear, with tools that are self-serve and code free, built from scratch in a few months.

Maybe now is the time to have another think about capex.

If you would like to discuss these issues, or the impact of emerging technology or data and analytics on your industry, then contact our Data & Analytics team.

07 March 2016

Connected Devices – Does bringing us closer make an easier target?

Is our heightened connectivity bringing with it heightened risks to our security? Could malice organisations, criminals, hackers or even other governments be able to collect our data, record our phone calls and see through our cameras?

As the internet and all technology has grown, so has the debate surrounding its security. Much of the focus has been around the fact that with billions of new systems, devices and sensors connecting each year, the attack surface for hackers continues to widen, but when coupled with a lack of security in many of these devices and the potential threat is mind boggling. U.S. director of national intelligence, James Clapper, delivered a presentation earlier this month to the Senate Armed Services Committee about various threats and the Internet of Things (IoT) was treated to its very own private section.

With our growing reliance on technology and connected devices, data privacy, integrity and continuity services become all the more important. Clapper stated that not only does this incredibly heightened level of connectivity serve as easy targets for hackers and criminals but it is highly conceivable that foreign countries could gain access to information through our least protected devices. Whilst we believe our vital information is secure, with banks and governments etc, it is the same concept as having state of the art home-security on your front and back door, only to leave your bedroom window open. It isn’t just cyber espionage that is a cause for concern, multiple terrorist groups are well known for their strong media campaigns for recruitment and organisation through social media and messaging platforms such as BBM. However, it isn’t all doom and gloom, as pointed out by Clapper.

"Devices, designed and fielded with minimal security requirements and testing, and an ever-increasing complexity of networks could lead to widespread vulnerabilities in civilian infrastructures and U.S. government systems. These developments will pose challenges to our cyber defences and operational tradecraft, but also create new opportunities for our own intelligence collectors."

We’ve seen very recently, with the Apple Letter to Customers link going viral, how the FBI are particularly keen to use our connectivity through devices to monitor and obtain information. Additionally, a study by the Berkman Centre for Internet and Society at Harvard University disputed the FBI's contention that data encryption on the Internet was significantly hindering efforts to track terrorists and other criminals. The study instead found that the rapid growth of connected devices and systems gives the federal government a widening array of opportunities for surveillance and data gathering.

It is becoming more commonly argued by the technology sector that security in the age of connected devices and the collection of personal data by companies and service providers is a complex issue made more challenging by the lack of attention given to it by consumers and device makers alike.

Consumers are more interested in the convenience that connected devices give them, while system and device makers tend to worry more about the features in their products than the security. Think how often we are seeing big data breaches in the news with customer data being lost in the 1000’s. Features are built with the customer’s usability in mind, security regarding those features is rarely a thought during design. Data is not only a security problem through theft – manipulation or “false data” can have devastating effects for systems, such as automated trading systems causing significant stock market fluctuations. With the growing reliance on artificial intelligence this is an additional concern for data security.

With increased connectivity and data sharing comes greater risks. As a result, you need to know your data is safe and guarded. How protected is your most easily accessible data?

If you would like to discuss these issues, or the impact of emerging technology or data and analytics on your industry, then contact our Data & Analytics team.

04 March 2016

Social Media - The Hacking Revolution

When it comes to social media and networking, it’s all a bit of fun, right? Connecting with friends, family, business colleagues, strangers, dangerous criminals…Wait, what was that last one? Yes, you read that correctly. Social networking sites are becoming an increasingly used tool by ‘cyber-criminals’ to target organisations or people, for the means of getting your vital information and data. You are probably wondering, how can someone use the information I post across social media to commit a crime against either myself or my organisation? In all honesty, it is quite scary just how straightforward some of these attacks can be.

Think first about what you share over social media on a personal level. You have your age, photos, job role and company posted across most networking sites. Often personal details to a more shocking level are disclosed on social media, ranging from medical problems, travelling, relationship troubles or frustrations with work. Social media is international and here to stay. We share and connect with people across the world every single day. In fact, we spend on average a quarter of our time online on social media sites, and these websites account for an astonishing one third of all internet usage. It helps to build relationships but it can also tear them down just as quickly. You have instances of burglaries taking place due to information obtained on social media; employees losing their jobs due to inappropriate posts or comments they have made on various networks as well as many other examples. When you bring that now to a professional level, it is easy to see the links and how the information and interactions organisations have across networking sites can be just as damaging.

Most news stories these days break on Twitter; organisations are facing an increasing demand to engage more effectively with customers across such channels. This is now one of the primary methods that organisations use to communicate with their customers, which means it is now more than ever susceptible to attack. Over the past few years there has been a notable rise in the amount of accounts being hijacked and attacked by hackers. Corporate accounts have been infiltrated and damaging messages have been posted. Not only this, but attackers can also use the data found across social networks to monitor the effects of current attacks by updates posted by organisations, to then alter their attack – as seen by the recent JANET hack. Furthermore, targeted attacks on organisations can be conducted such as spear-phishing to obtain information about company employees, which is then used to send seemingly legitimate emails across an organisations network to trick employees into providing information such as passwords or credit card details, or even to install malware onto company machines to conduct further monitoring for additional attacks.

The list appears to be endless, so the question now is, how can you protect yourself?

Be Vigilant.

Have you noticed any strange requests, or unknown users trying to connect with you? Have you noticed any colleague accounts or organisational accounts posting things that aren’t generally seen as ‘the norm’ for them, or could be damaging? Point it out for them, let your friend, colleague or your company know if you have seen something that may not be quite right.

Be Aware.

Be aware of what information you are divulging about yourself. If you are on holiday and away from your home, be cautious about who or what you are saying. Also, be aware of what social network you are using and what is appropriate given the platform you are using.  Be aware of who is talking to you, and what information they are trying to get from you.

Be Proactive.

You can take proactive measures to protect yourself across social networking sites.  For example, ensure that you are using two factor authentication when logging into your accounts. By doing so, this will alert you as to whether someone logs into your account that isn’t you, which means your account credentials may have been stolen, and it adds an additional layer of security on your account. Also, remove any information, photos or posts that you don’t want to be known as public knowledge and in turn ensure privacy is at the maximum level on your profiles to prevent those you aren’t connected with from finding out your information.

We can help you to effectively mitigate and govern risks, such as those discussed above in relation to social media. Is this something you need help with, or would like to chat about? Please get in touch with our Data & Analytics team.

03 March 2016

Big Data = Big opportunity? How does it work?

Every forward thinking, technology-centric company in the coming years will have to face a difficult and daunting question –how do we make big data work for us?

Cost and complexity are normally the major factors at the forefront of consideration. Big data environments therefore must be stable, highly-integrated and scalable to allow organisations to begin the migration towards true data-and-analytics centricity.

Data and analytics centricity is a state-of-being, where the power of big data opportunities and big data analytics is available to all the parts of the organisation that need them. It is becoming more familiar now that data provides the ability to solve genuine business problems and discover valuable insights. However, the data streams and user toolsets required to do this must be applicable to organisations’ big data infrastructure. That’s how big data should work.

So what components make up big data infrastructure and how can we assemble them for maximum performance in a stable and sustainable way? The three main components consist of:

  • Data sources
  • Data Platforms, warehouses and discovery platforms
  • Big data analytics tools and apps

Data sources examples include, but are not limited to, operational and functional systems, machine logs and sensors. Data platforms enable data capture and data management, and then critically its conversion into customer insights for much better informed action. Big data analytics tools are at the front end and used by the organisations analysts to access customer insights, models, scenarios and provide insight. As you will probably be able to see by now, it is vital to have a big data strategy designed to create a calculated big data architecture to examine both current data streams and data repositories but remains specific to business objectives and longer-term market trends. So the key take away is that there is no one correct template to using big data.

Whilst the capital outlays are considerable in establishing a move towards big data many forward-thinking organisations and big data pioneers have been able to prove that designing the right big data environment can actually produce cost savings, counter-intuitively. To top it off, these are no measly savings…but big, harvestable savings within a relatively short time frame. Now the big difference between a well-orchestrated and implemented big data infrastructure that leads to efficiency increases and cost savings, and one that sits by the wayside after running into a continuous stream of issues is integration - the buzzword of Big Data.

Integration is arguably the pivotal variable of big data success, as reusability is virtually tied to it. With poor integration comes poor reusability, leading to poor efficiency and lower cost savings. Forrester Research estimates that up to 80% of big data value stems from integration. The entire concept of big data is that the highest-value data is readily available to the appropriate users, with robust and clear governance structures. This is where big data starts to make its money. Data scientists are able to review and model long-tail customer histories or transactional data stored in deeper data sets, as and when necessary. This means they only need reliable storage and robust data management.

So we appreciate integration is key, but how can we take action to improve such a subjective variable? Big data integration is all about the big picture. Data islands must be bridged, dots must be connected and gaps filled in to create an inclusive and multi-dimensional environment. Good integration, well designed environments, uniform and compatible data architectures – all focused around data and analytics – might not be the definitive list to make big data function. In fact, much less than this is needed to simply operate a big data environment, but they certainly stand out as the difference between the average and the exceptional data management programmes. Could big data be the future for your organisation and its ever growing digital needs?

If you would like to discuss these issues, or the impact of emerging technology or data and analytics on your industry, then contact our Data & Analytics team.