« January 2016 | Main | March 2016 »

7 posts from February 2016

25 February 2016

Blessed are the MOOC

If you’re interested in developing your data skills or just have a desire to know more about anything from Parkinsons Disease to Hans Christian Andersen fairytales, the MOOC might be for you.

The term MOOC stands for "Massive Open Online Course". MOOCs are designed to let people learn from anywhere in the world, and study at their own pace. This lets you structure learning around your free time, providing flexibility as well as high quality teaching. The content can take the form of videos, articles, discussions or quizzes and often involves tasks which test your understanding of that section.

Supporters of MOOCs have argued that they open up new opportunities for people to learn with a minimal cost of joining. One of the founders of Coursera stated that “It will allow people who lack access to world-class learning to have an opportunity to make a better life for themselves and their families”. Future Learn celebrated its 5 millionth course enrolment recently, and is regularly updated with new courses, partnered with universities all over the world. The effect of MOOCs has even been compared to the way in which Spotify revolutionised the music industry, forcing traditional companies to change the way they operate whilst also allowing people shut out of the current system access to such opportunities.

It’s difficult to draw conclusions about the impact of MOOCs without examining the data behind the signups and completion rates. FutureLearn released basic statistics about their audience after the first eight courses were launched last year. They found that the majority of learners already had some kind of degree and many were returning to education or learning again after a period of time.

This demonstrates that the use of MOOCs is potentially a new method of learning and is encouraging those who may not be otherwise in the education system to learn something new. However, the people doing so are those who have mainly been educated to a high level already. The view of the academic director at the University of Nottingham is that institutions have now moved beyond trying to widen participation and MOOCs are now intended for use as professional development tools, with increasing focus on business.

A group of students from Harvard looked in more detail at the relationship between socioeconomic status and MOOC enrolment, focusing on the Harvard specific platform, HarvardX. They found that a seventeen year-old whose most educated parent has a bachelor's degree is more than five times as likely to register than a seventeen year-old whose most educated parent has a high school diploma.

Back in the UK, MOOCs have come under criticism in recent weeks as the Open University announced it was going to close regional centres. The Open University owns Future Learn as a commercial subsidiary, and is having to adapt to this change in technology and the way in which its courses are delivered.

I had the experience of completing a MOOC on an Introduction to Cyber Security, hosted on FutureLearn. I found this to be interesting and the content was varied enough that you could work on the course for a section at a time, without becoming bored with articles the length of a novel or a video as long as an average TV program. Although I studied Computer Science at University, there was enough new information and interviews with people in the field that the course remained engaging throughout the 8 weeks that it covered, but I also think that anyone new to the field would have been able to pick up the basic information easily. The self-paced nature of the course meant I could easily fit it in around daily life and the content is still accessible now if I decide I want to go back and refer to something. On the other hand, I've lost track of the courses which I signed up to on a whim, read one article and never looked at again, so it can go either way!

If you’re interested in signing up for a MOOC yourself - there are lots of different sites out there– in particular Coursera, EdX and Future Learn, where the courses are free to sign up and there is new content to focus on each week of the course. Certificates are often available to prove you’ve completed the course and can help to boost your CV.

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.

24 February 2016

Drawing False Conclusions

Accurate interpretation of data is a difficult task. Despite this, even the smallest companies are attempting to apply analysis to any sources of information they can get their hands on, looking for a pattern that explains their performance. An added challenge to any business taking this approach, is human tendency to perceive meaningful patterns where they do not exist. This tendency can be exhibited by simply flipping a coin. Below are two sequences of coin tosses:

TTTTT HTTHT

We intrinsically view the second sequence as more likely, as landing five tails in a row just doesn’t feel random. Of course, those well versed in probability know that the probability of either sequence taking place is exactly the same, ½ * ½ * ½ * ½ * ½ = 0.015625. To add to our inability to accurately process probability we find patterns in random sequences, such as the sequence of forty coin tosses below:

HTTTHHHHHHHHTHHTTHTTTTTTTTTHHTTTTHHHHTTT

We all see distinct clusters of heads and tails, leading us to believe there are patterns in the data, when in fact there are none. The technical term for this phenomenon is Apophenia, the human tendency to perceive meaningful patterns from random data. It is the reason behind many of our everyday assumptions, from our fear of flying, to why it seems to be 11:11 so often when we glance at the clock. Why do we hold these beliefs?  (Answers: Disastrous events are remembered more vividly, whilst uneventful flights are mainly ignored.  11:11 is a memorable number and you look at a clock more often than you think)

Apophenia can create inaccurate conclusions from our client’s data. It is tempting to create meaningful connections from meaningless figures, to find the trend that explains poor performance or record sales. With the ever increasing volumes of data available, the opportunity to discover these non-existent patterns is only becoming more pronounced.

Approach your conclusions with an awareness of how we are hard wired to see what is not there. Remember that our brains are pre-disposed to find patterns, so do the hard math before relying on your intuition to give you the answer.

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.

22 February 2016

How not to build your social media presence

Have you ever received a tweet or email offering you a quick and easy way of increasing your followers, fans or positive reviews?  My guess is that you have, as I get these all of the time. But, have you ever been tempted to give it a go?  For a seemingly small price, you’re offered 1,000s of extra followers or reviews. What could go possibly go wrong?

Amazon has been in the media recently as it plans to take over 1,000 fake reviewers to court.  Most of the reviewers had advertised through freelance website Fivrr and offered to write reviews of products and services for a price.  Likewise, if you've spent any time on Twitter, it's likely you'll have received offers from other users offering to increase your followers for you.

This recent crackdown indicates that Amazon is serious about maintaining the integrity of its review and rating system.  

Here are 3 reasons why you should stay clear of paid-for followers and reviews.

1) It's pretty obvious what you're doing

It's often pretty easy to check if an account or product is legitimate.  If your account has thousands of followers, but has made very few tweets or many spam-like tweets, most will realise there's something fishy going on.  Likewise, many people look at a selection of 5 star reviews and 1 star reviews to get a feel for how good a product or services is.

2) You risk having your account suspended

Buying followers or reviews is nearly always against the terms and conditions of the network or online platform.  So, buying followers or reviews only increases the chances that your account will be suspended.  Now, that would be embarrassing!

3) It's just plain wrong

Trust is such a valuable commodity.  Once lost, it can take a long time to rebuild.  Don't deceive your potential followers or customers by buying fake followers or reviews.  While it may make you or your products look popular initially, it's not a long terms strategy for success.  It’s a risky tactic that can harm your reputation.

So, how can you increase your followers and reviews legitimately?

Creating a strong, powerful brand that people trust takes time.  Deception will always harm customers' trust.  Some of the most powerful consumer brands have strong customer advocates.  Those brands didn't win their customers' trust through deception.  They deliver on their promises, and their customers come back for more.  Those customers willingly post positive reviews about the brand's products and services and retweet or share its content.  These endorsements will help you build a strong online presence and a genuine following.  

And don't get blinded by the numbers.  Having a smaller number of engaged followers is way better than a large, unengaged following.

Create good products which you're proud of.  Create great content which will engage your followers, communities and future customers.  Don't take shortcuts.  In the long term, a shortcut made now might turn into a long and winding road full of obstacles later.

Have you made sure that others in your organisation understand this too?  Your policies should address these practices and set out how your staff and colleagues should engage on digital platforms.

How are you engaging with your followers?  And are your staff onboard?

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 February 2016

Can you trust your organisation’s data analysis?

Mistakes happen, no matter how experienced you are at doing data analytics. You might know somebody who claims that they don’t make mistakes, or somebody with so much experience that their analysis is always trusted and never checked. But whoever does the data analysis, there will always be errors that aren’t spotted unless it’s properly independently reviewed. A single mistake can also have grave consequences, as we saw with the Intercity West Coast fiasco when a simple modelling error was one of the factors that cost the government more than £40m.

I recently reviewed a model that an organisation was using to support their strategic planning. The model was built by another large consulting firm, and while it looked very professional on the surface, I uncovered a catalogue of errors that were completely corrupting the model outputs. These errors would have been very difficult for the client to spot, and if the model had been reviewed properly in the first place they would not have been there.

However it’s not just complicated data analysis models that need to be reviewed. Simple Excel spreadsheets that organisations or departments rely on can often contain difficult to spot mistakes. If the professionally built model that I reviewed could contain so many errors, then I sometimes wonder how many errors are contained in the quickly developed spreadsheets that can frequently become ingrained in organisations.

Investing the time to review modelling tools is often overlooked. It’s not a glamorous task, and it can be low on people’s priorities. Organisations also don’t always have the right people available with the right capabilities to conduct these independent reviews. However, without an independent, systematic review of your data analysis there will always be a risk that the results are inaccurate or even just plain wrong. Don’t make a wrong choice with your strategy or become the next Intercity West Coast just because somebody has put a bracket in the wrong place!

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 Joe Stacey - Manager, Strategy&

12 February 2016

Using data to predict the future: the Holy Grail

I was recently in a very insightful presentation from Qlik (the writers of Qlikview) and they touched on the topic of Predictive analytics. As a practitioner and consultant in the area of data, I am always intrigued about this topic as it is a popular topic for discussion, but a complex thing to get right.

I was listening with my usual amount of professional scepticism, but my ears pricked up when I heard a subtly different take on this area... the presenter used the word...."anticipation". This was a very interesting moment as I felt that suddenly someone had captured the essence of "predictive" analytics, but in a much more practical and accessible way.

The more I thought about this I realised that it linked nicely with my own views. People are often very concerned about doing something "big" with their data, like implementing predictive analytics, but this is not easy.

The idea that you can predict the future is obviously the Holy Grail for any business, but simply being able to anticipate something is often all you need to be one step ahead of the competition.

So what does this mean practically? I think the use of techniques to “anticipate” have applications in a number of areas.

Social media immediately sprang to mind and the ability to read and understand the mood and views of the public. This sentiment analysis would not normally fall into the category of predictive analytics, but it can certainly allow you to spot a trend before it turns into a nightmare... or maybe even spot a trend that you can profit from.

In a more traditional, pure data analytics context, this may be about simply ranking customers or suppliers using risk factors or highlighting the occurrence of an event... maybe missing a payment... or making a bad delivery.

Suddenly all these things fitted into the context of "predictive"... but more accurately... "anticipation".

So where does this little blog leave me? Well, I think I will be talking about the practicalities of predictive analytics in a different way. I will be talking about taking small steps. I will be talking about analytics and monitoring that can be used as ways of "anticipating" interesting and important events that might influence your business.

... a major step in the right direction... on the road to predictive analytics. 

…and Maybe we all need to ask ourselves a question…

“Could I take that first small step?”

It might be the beginning of something amazing!

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.

09 February 2016

Personal data: Who do you trust?

Trust is an essential commodity in business – you only have to look at the experience of Volkswagen over the past year to see what happens to your reputation (and your profits) if you damage your customers’ trust. Companies know this, and it is perfectly normal to hear large corporations say loud and often how hard they work to maintain a trusting relationship with consumers.

In the data-driven world, this becomes even more important but even more difficult. It is practically impossible for any of us to use the internet or social media without handing over personal information in some form or another. Personal data is a new currency – companies want it, need it and use it, and that makes it valuable.

In this world, trust is everything. A report by DMA Group and Acxiom found that 40% of consumers choose trust in an organisation as the most important factor when deciding to share personal information – four times more than any other factor.

But what do we mean by ‘deciding to share information’? How often is the decision to hand over your data conscious and thoughtful? If we’re honest, in most cases it means ticking the box that says you’ve read the terms and conditions attached to a website. But how often do you read that endless stream of legal jargon that pops up? Do you really read the T&Cs? If you have, congratulations, but you’re in the minority. A survey by Skandia found that only 7% of adults read online T&Cs when signing up for new services or products.

In other words, while companies, big and small, say that their customers have agreed to let them use their data – and believe that it’s because their customers trust them to use it well and wisely – they’re deceiving themselves, and us (albeit with our collusion), on a gargantuan scale.

The truth is that most of us aren’t comfortable with sharing our data – a survey by Digital Catapult in 2015 found that 60% of consumers were uncomfortable with sharing personal data (65% because they were worried about it being shared without their consent) and 14% refused to share any at all. The younger generation, though, are more amenable; 84% of 18 to 34-year-olds say they’re comfortable with sharing personal data if they get something in return (which may simply mean access to a particular site). Our own research into the use of wearable technology in the workplace had similar findings with 70% of millennial workers saying they would trade their personal data from a wearable device for a better work deal.

The more cases we see of companies losing data, through hacking or a lapse in internal control, the more likely it is that we’ll start to refuse to share our data (or choose to provide false information). So if companies want to secure access to personal data in the future, and protect their relationship with their customers, I think they need to consider changing their approach to data consent.

I think the answer lies in simplicity. Why not replace the unfathomable, jargon-laden T&Cs with a one-line consent question – jargon-free, strictly no lawyers involved – and ask website users to reconfirm their permission if the context of use changes? It’s a risk, but isn’t it more likely to win trust than an endless stream of legalese that no-one reads? In the long-term it could even create a stronger, more proactive and dynamic foundation on which trust is built. Digital business is all about innovation, after all – so let’s take an innovate approach to data consent.

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.

02 February 2016

Energy suppliers or data companies? A new world of opportunity in the Utilities sector

by Andy Howe Senior Manager

Utilities continues to be an incredibly exciting industry, with rapid change and whole industry projects.

Those projects and their scale provide great opportunity but also great risk and disruption, with the potential to directly impact the whole of the UK.

The 'Smart' meter roll-out programme has the potential to significantly reduce the burden on our national electricity infrastructure by ultimately enabling consumers to flex their demand in response to dynamic tariffs that are intended to smooth out the peak demand periods, providing a more stable generation profile without the need to be able to manage the existing extreme spikes in load and demand.

Some of the other benefits will include: more accurate billing based upon actual usage and real-time information to customers through home display units that means they can pro-actively manage their usage to reduce bills and environmental impact.

This programme though requires the replacement of existing meters across the country with a current target for completion by 2020. This is a major challenge for the utilities companies, not just managing the physical replacement but also the resulting changes to underlying systems, data and managing the customer experience during the transition period. All those challenges involve data, whether it be assessing the quality of the underlying meter data used across the industry, the accuracy of the bills issued to individual customers or informing the most optimum way to roll-out the meter assets across the customer base.

Data is being used increasingly to provide information to customers about their bills and their options. This can be seen in the continued rise in price comparison sites that take detailed data about a customer, their property and consumption to identify better deals. The regulatory drive by OFGEM to focus on the customer and to keep them informed has gone so far as to now require suppliers to include details of what customers could have paid had they shopped around for different tariffs and present that on the actual bills issued to them. There really is no excuse for not knowing if we could get better deals elsewhere!

We see companies trying to make more of their data assets by linking disparate legacy and new systems to try to identify the impact of issues on the customer experience and ultimately their retention and profitability. For example the linking of customer complaint systems with billing and debt recovery to see the impact of poor customer service.

Those legacy systems have provided a stable platform for providing various services to customers but have not proven to be flexible to the changes in the way customers want to interact digitally with their suppliers or indeed move into the world of Smart metering. Recent press reports have shown the risk involved for utility companies moving from those legacy systems, with significant financial and reputational impact of getting it wrong.

'Challenger' suppliers have taken the market by storm. Unencumbered by these legacy systems, and with smaller populations of customers, they have been able to identify and embrace customer wants, along with new technologies to provide a much more digital and dynamic service. This is having a real impact on the Big 6 electricity suppliers, for example, who are continuing to see significant numbers of customers moving to these challengers.

And the possibilities for linking data sets retained by the energy companies with third party data sets open up a whole new world of opportunity: improving billing, driving the customer experience, targeted marketing. A question to ask is 'are we talking about energy suppliers, or are we actually considering data companies?' It feels like the latter to me.

These are all major, exciting challenges and opportunities for our clients....and for PwC to support, assure and challenge them using data to provide the real evidence that they are looking for.

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 Andy Howe Senior Manager