These days it’s an undisputed fact that a successful digital transformation requires a cultural shift across the organisation. A holistic approach spanning across process and people, infusing itself not only in the highly disrupted areas of the business, but in the more stable, unchanged areas too. Changing the way people think is a key priority.
But what about the point, six months or maybe a year in, where the process has begun and people are keen to start thinking digitally? First of all, congratulations – as many as 84% of companies fail at digital transformation. You now have an organisation of keen and enthusiastic people ready to champion the digital approach. Excellent.
Have you ever heard the quote which originates with a German General from the 1930’s which is now much repeated in the business world? People can be categorised into clever or stupid and industrious or lazy – they’ll be one of the first two and one of the second two. The most dangerous kind of people are the stupid, industrious ones. While the world of business is maybe not so black and white, it’s certainly true that without the correct knowledge and understanding, your team’s new found enthusiasm for all things digital may at best go to waste and at worst do more harm than good.
At the centre of a successful digital transformation strategy is the customer; but the foundation is built on data. After customer centricity, data is at the heart of good decision making. Whatever the goal of your digital transformation, it almost certainly has a basis in data, whether you’re creating efficiencies, optimising processes, growing revenues, or anything else. It’s likely too that measurements of success will require some kind of data analysis. Your people will be in contact with data, day in, day out. After all, information is data with context, and is required for everything.
Data literacy, the ability to read, understand, create, and communicate data as information, is therefore one of the most sought after skills in modern businesses, and one of the things which separates a truly excellent digital transformation strategy from a merely ‘fine’ one is the focus on using the momentum to upskill staff so they are best equipped to operate in their new digital world.
This might feel like another hurdle on your road to digital realisation, but it doesn’t have to be. If you’ve successfully gained buy-in from your organisation and started the cultural change, your people should be primed for engaging with a data literacy program. The first step is identifying the different levels of existing data literacy within your organisation. For example, it’s likely that business analysts have a high level of data literacy already and are in positions to act as ambassadors of your data literacy program.
Even if you don’t engage with a full organisation-scale education program, there are certain key tenets that you should make sure are well understood across the company. Being digitally literate will empower your employees to feel at home in their new digital world. I’ve outlined some of the most important pointers to get up to speed in data literacy here.
The First Step
Former BBC Head of Statistics Anthony Reuben has shared some powerful advice about how he approaches his current role as a member of their fact checking team, and it is just as relevant for business people as it is for journalists. The first thing you should ask when presented with any piece of data is “Is this reasonably likely to be true?” A little bit of upfront reality checking could save a real headache down the line. Take for instance a report that the UK uses 42 billion plastic straws per year. Setting aside the obvious importance of the message, let’s examine the number. The population of the UK is around 66 million people. If that report is accurate, every person in the UK is using over 630 plastic straws a year. That’s nearly two every day. I don’t know about you but I don’t know when I last used a plastic straw.
Suss out your sources
Where has the data come from, and what does the source tell us? The most obvious example is that of a tobacco company presenting findings which suggest smoking isn’t as bad as everyone claims. All it takes is a bit of awareness that thinking critically about the source can prevent simple errors. Just like a journalist wouldn’t trust “my mate down the gym told me” as a valid source, you too should be a discerning consumer of data. If you’re not a chemist, it could be easy to be alarmed by the website www.dhmo.org, but a little bit of scrutiny will reveal what dihydrogen monoxide really is – water.
When someone tells you something is on average, it might indicate the need to ask a few more questions. First of all, this could be referring to three different things; if you remember your school maths, you’ll know these are the mean, the mode, and the median. Not all averages are created equal. One of my favourite examples of this is that people in the UK have on average less than two legs. This is because nobody has more than two, and some people have fewer. While this claim is not wrong, it’s also not very helpful. That’s using the mean (adding up all the results and dividing by the number of results) – both the mode (the most frequently occurring) and the median (the result in the middle of the highest and lowest) give the much more expected answer of 2.
Careful of Cost
Every winter, there will be an unexpected snow storm, roads will grind to a halt, and the headlines will explode with cries of how much it costs the economy. Using another Anthony Reuben example, let’s consider some snow we had in 2009. One such headline claimed the disruption cost the UK economy £3bn – a figure they got from the Federation of Small Businesses. Unpacking this, they split it down to £1.2bn each for a snowy Monday and Tuesday, with the rest of the smaller costs being incurred throughout the week.
The first thing you should be wondering is where this number came from. Helpfully the newspaper in question tells us: the FSB estimated that 20% of people would be off work due to snow on each of Monday and Tuesday, and therefore took 20% of what they have estimated a bank holiday to cost the economy (£6bn, apparently). This raises more questions than it answers, frankly. Anyone who has ever worked in retail or catering will be astonished to learn that the whole economy grinds to a halt on a bank holiday. There are more esoteric considerations too, like the occasional three day weekend might actually make people more productive the rest of the time. It simply doesn’t make sense to see a bank holiday as a complete loss of output for the economy.
If you’ve been looking closely you might have spotted another, just as problematic assumption: that snow is bad for the economy at all. When someone describes something as a cost to the economy they usually, as in this case, are referring to a loss of GDP. However the GDP of a country is a very complex measure. While some businesses, for example, manufacturing plants that depend on having people on site, and getting deliveries of materials, will experience a loss due to the snow, there are a growing number of people who can simply work from home. People probably have their heating on if it’s snowing and they are stuck in the house, as well as using other appliances they might not normally be using, and they might be using the time stuck at home to do some online retail therapy. And then there are cases such as a person whose haircut or dentist appointment gets cancelled due to the snow. That person still needs a haircut or a check-up, so the economy isn’t losing out on that spending – it’s just being delayed. And then consider that sometimes bad things are good for the economy – if some people crash their cars due to the snowy weather, that’s actually good for GDP.
The same can be said of costing estimates across the board – they are notoriously difficult to compute and almost always contain some suspect estimations.
If a figure is quoted as a percentage and only as a percentage, you should be wondering whether it’s because giving the actual numbers would tell a different story. For instance, if you look at the sales of vinyl records, they show a trend of decreasing for years, until recently when there has been somewhat of a resurgence. If you want to paint this in a certain light, you need only give the percentage increase in sales during this renaissance, which is around 1900% - but in their decline their sales fell from around 90 million per year to less than 1 million – that’s about a 90% drop. After the new wave of vinyl buying, the annual sales are around 4 million. The actual numbers tell a very different story to the percentages.
Big Numbers Overwhelm
It is incredibly different to really get to grips with big numbers. The annual output of the UK’s economy is a little over £2 trillion – two trillion! How can a person really have any concept of how much money that is? It means that it can be quite easy to obfuscate the reality behind the numbers when they get larger than a person can get to grips with – you see this a lot in politics. The most helpful way to approach big numbers is to contextualise them, which is why you see a lot of things being compared to the size of Wales – it probably means more to someone than saying 21,000 square kilometres. Dividing something by some other relevant number may also help, which is why government spending is often expressed as ‘for every person in the country’ or similar. Remembering a few key ‘big numbers’ can also be useful, such as the population of the UK, or the distance from Earth to the Moon.
Correlation Not Causation
Getting caught out by the assumption that correlation implies causation can be very costly. It can be very instinctive to see two things increasing in line with each other and to go on and assume one of them caused the other. However, this link simply can’t be assumed. If nothing else, this article should have demonstrated how figures are often not what they seem on the surface, and often have a much more complex story to tell when you dig a little deeper. This is the case here too. To start with, sometimes a correlation is just a coincidence.
Secondly, sometimes there are confounding factors – if you’ve got an ailing pot plant and go with the dual method of playing it some jazz music every day and adding some plant food to the water, it would be spurious to claim that jazz music makes plants recover their health.
All or Nothing
Cherry picking data can be at best misleading and at worst extremely dangerous. If someone is picking and choosing from a collection of data and presenting only that which supports their stance, only part of the story is being told. This is particularly noticeable in politics, where it always seems that every side can provide data that upholds their points while contradicting their opponent’s. Be aware of what you’re not being told.
When evaluating whether one thing causes the other, check whether the numbers you’re looking at are weirdly specific: if they are, they might have been cherry picked to demonstrate a point rather than giving a full overview of some information. If you wanted to look at whether one thing causes another, you’re likely to have a time period in mind, e.g. 10 or 15 years, before analysing the data. If the numbers are weirdly specific, for example something being more likely to happen to a certain group of people over a period of 13 years, ask what happens to the data if you round it to a more sensible number. It is likely that the data with the most prominent correlation has been deliberately chosen, and the conclusion starts to break down if you change the range. A good example of cherry picking is the UK debt – you could say that public sector debt has nearly tripled since 2002 – note the strangely specific date. If you expand the range you’ll see that the debt as a % of GDP is less than half of what it was in the 1950’s.
Words of Wisdom
Before I leave you to go and evangelise these key data literacy pointers, one last warning. There are some words and phrases that should raise more questions about the data. ‘Up to’ is a common one and is essentially meaningless if you don’t care about the maximum. If you were offered a job with a salary of ‘up to £40,000 per year’ you’d immediately be on guard – that caveat could be held to if you were given £0. It just means you won’t be given more than the specified number.
Anything claiming to be a record number, especially when it comes to things like spending. The world is a growing place, and such is the way of the western economy that population growth, economic growth, and inflation means that things like government budgets are almost always at a record high. Comparisons over a long period of time are very tricky beasts too. Metrics, methods of measurement, rounding conventions… these things change over time, and it can have a big impact on how the numbers look. If you’re comparing something over a 500 year period of time, how certain are you about the veracity of those 500 year old figures?
Whatever approach you take, the key is to provide your staff with the skills required to make the most of the exciting new frontiers you’re soon to be exploring with your digitally transformed organisation. These ‘bare minimum’ tenets of digital literacy should be a good jumping off point, and the internet is full of interesting numbers you can use to illustrate your points.