Coming to terms with terms: digitization, digitalization and digital transformation

By | July 15, 2020
Photos as memories

Photos as memories

Singapore, July 15, 2020 – There’s lots of grey when it comes to three terms that as a journalist I used rarely because they were such turn-offs to readers and editors alike. But companies like them and they’re useful, up to a point, to help us understand this process we’re going through.
The terms are digitisation, digitalisation and digital transformation.

Yes, they’re horrible.

One reason they’re horrible is they aren’t exciting. No way would an editor of mine have OKed a story with those words anywhere in it.

Another reason is they’re very similar, in both sound, and definition. Nobody seems to agree on what they are, which is usually a good sign you’re veering into marketing/consultant-speak. When a term is not one that people use in public confident that everyone in the room understands it and agrees with everyone else what it means, you

a) shouldn’t use it and

b) should assume it’s dreamt up by some fella to sell more widgets (or consulting time.)

As a consultant I’m offended by that so I’m going to take a stab at defining it. It’s not that the concept is hard, it’s that the terms, I feel, aren’t particularly helpful.

So here’s my stab at defining them, in the hope that they actually demonstrate something useful, which is presumably why we have them. (And yes, arguably we should ditch them.)

The world is still largely analog

The natural world is analog. In the words of Peter Kinget, Department Chair of Electrical Engineering at Columbia University:

The world we live in is analog. We are analog. Any inputs we can perceive are analog. For example, sounds are analog signals; they are continuous time and continuous value. Our ears listen to analog signals and we speak with analog signals. Images, pictures, and video are all analog at the source and our eyes are analog sensors. Measuring our heartbeat, tracking our activity, all requires processing analog sensor information.

For most of us this is an ongoing process. We are forever asking our devices to convert the analog world to digital. But let’s keep it simple: suppose we have a cupboard full of old photographs. Or slides. Or negatives. They are, obviously, analog. We can still look at them, hang them on the wall, put them in an album, print off the negatives, show the slides on a projector, but all of those are analog processes. The data has not been changed.

So we are still keeping both the data — the photos, slides, negatives — and the process analog.

Think of it like this:

Analog to analog: images

Analog to analog: images

You may be quite happy with that arrangement (I am — I can never throw out analog photos, they seem to be quite a durable medium) but the pressure is on to digitise. So we go about converting that data to digital by scanning them. This is an analog-digital process:

Analog to digital: images

Analog to digital: images

This all might seem rather basic, and it is. But it gets more complicated when we talk about more complicated digitisation. When a library digitises its books that is obviously using a similar process. But when it digitises its catalog, but not its books, it gets more complicated, as we shall see. How useful are the terms when an enterprise only digitises half of the process?


And there’s another problem. As I mentioned, we’re living in an analog world. And so a lot of our supposedly digital tools are actually largely performing the same task as the scanner in our photos example. But in real time, all the time. Take our cellphone, for example. More of the chips in there are actually ‘analog chips’ than not digital ones. An analog chip will handle power supply to produce a well-regulated supply of power to other chips, wideband signals, and sensors. Or they may combine with digital chips to convert analog to digital (a temperature sensor, say) or from digital to analog — for making sound, for example.

Digitisation in a phone

Digitisation in a phone

Here’s another way of looking at the same thing:

According to Cricket Semiconductor, from which I’ve borrowed the above and below images, there are more than twice as many analog chips as digital ones. (Cricket itself might no longer be with us, and the iPhone model is a very old one, so this proportion might be out of date.)

These chips are forever converting real world data into digital data, from where you are, to how you’re holding the device, to where you’re touching the screen, what you’re watching, listening to, taking photos of, as well as some of the actual communication between your device and the outside world. Digitisation, in other words, is not necessarily managing a step but managing a continuous process. This bit, I believe, is why we run into problems with the next two terms.

If it’s digitised, it can be digitalised

Yes, an ugly aphorism, but the idea is a simple one: Unless you’ve gone through the digitisation stage, outlined above, you can’t start to reap the benefits of digitisation. Which is what we call digitalisation. Not all of us, but let’s for the moment leave them out of it.

Going back to the pile of photos. You’ve scanned them into the computer and they’re all now bits, noughts and zeroes. And you can look at them on your computer, or phone, or whatever you used to scan them. But they’re not digitalised, as it were. Once again, this is both a data and a process.

  • First you would be renaming the photo files to something useful — usually a date, perhaps with some idea of who is in the photo.
  • Then you might be adding some metadata — data about the data (in this case a photo).
    • You might do this manually — adding details to the file itself (i.e. not the filename, but the fields that accompany the JPEG format, or whatever format you’ve chosen to store the file in.) These could include location (geolocational data, usually in the form or coordinates), type of camera, date the photo was taken, subject matter. Anything you like.
    • Some of this process might be automated — for example, dumping the photos in Apple Photos, and letting it scan the photos for faces, and then grouping those files together when it recognises your Aunt Maude is in them. (In more complex examples, the digital images can be explored using something called computer vision, which is essentially training a computer to see a digital image and work out what it contains — whether it’s a dog, or a traffic light etc.)

Now this is, in my view, part of digitalisation, not digitisation, although you can see how this might be argued either way. To me you’re now already into the process of adding value to digital data by adding metadata to the photos, which is to me the key element of digitalisation. We’re adding data to the data so it can use, and be used by, other data and processes (what we call applications.) We can now search for photos of Aunt Maude and find her without having to remember when we last saw her, and so which box of photos or albums to hunt through, or if the photos were digitised but still lacking metadata, trawling through hundreds of thumbnails until we spotted her glistening red beehive.

Digitalisation of images 1

Digitalisation of images 1

Going back to the iPhone, this process of digitalisation is tightly woven into the process of digitisation. When our phone is busy converting real world, analog, information and signal into digits, that is just a conversion process. When that process is finished (which of course it never is, but I’m referring to individual sessions of conversion) then the digitalisation — the digital dividend — kicks in. For the iPhone that is seamless and largely expected — after all that’s the point of the device, a pocket full of real-world tools and applications — but the digitisation is still a process that has to happen. It’s just so quick and seamless we don’t realise that it’s two processes: digitisation and digitalisation. The capturing of real world data and converting it to digits, and then adding value to those digits by turning them into usable data. (The computer vision process mentioned above could also be compressed — photos and video are shot and analysed in almost real time, because they may well need to be. The automated or connected car needs to know whether it’s about to hit a dog in the road, as the below GIF shows.)

Digitalisation is a multi-step process

Now digitalisation doesn’t stop there. When data is digital it can now start talking to other digital data. Other applications can understand that data, combine it with other data, and create new data, and thereby add value. In our photos example, the photos — or usually the underlying metadata — can be connected with other applications, such as search engines, or databases, or virtual reality games.

Digitalisation of images 2

Digitalisation of images 2

In the case of the phone, all that real world data about heat, position, moisture, sound etc can be used by dozens of applications on your phone. Without that real world data the phone is surprisingly dumb. (And even wifi and GPS signals require some amount of analog to digital conversion.)
The apps that depend on analog data
The apps that depend on analog data

Now some would argue this is also ‘digital transformation’ because, when it comes to business, processes are being transformed by the digitisation dividend. By converting analog to digital and using that data it’s argued that digitalisation is synonymous with digital transformation. I don’t buy that, it strikes me as lazy shorthand and not properly looking at the stages involved:

Digitisation has converted atoms to bits;

Digitalisation has converted those bits to data that can be interpreted and used by the rest of the digital world (within the device, the house, the company, the world).

And yes, just as digitisation was also both data and process, part of that is also the process of making use of these digital assets. But it’s not ‘transformative’, at least in the sense I understand it.

Take the library: they digitised the catalog. Great.

But no biggie. People still have to go find the books on the shelves; they are just able to confirm its existence more readily — and in theory remotely.

Then the librarians converted the entire library to digits, scanning every book.

Better; now I can read the book on my iPad, in theory, and I don’t need to go to the library. Good. But. I would argue that’s digitalisation more than digital transformation. They may have transformed their own procedures, but not yet undergone digital transformation.

Let’s see why.

Digital transformation is, or should be, when processes and businesses are transformed

So let’s start with the library this time. It’s not going to take long before people realise that you don’t actually need a physical library (at least for storing books).

Or librarians.

Or even digital books. Why not just let people search the text and metadata of books digitally and put together whatever collection of reading, or notes, or insight they want?

Why not convert the librarians into curators, who develop systems to connect disparate subjects and disciplines together, training algorithms to think better than we humans about the links between subjects? Or to mine data from readers to better understand and recommend more books to them, or figure out how to encourage people to read more?

Librarian as curator of ideas

Librarian as curator of ideas

Whole new services could emerge from what we used to think of a staid environment wedded to slumber and the worship of dead trees. (And, yes, we could use the libraries for something else: poetry, education, talks, a post-prandial nap, advice.)

This is what I think is meant by digital transformation. It’s a long drawn-out process that we’re only beginning to touch the edges of. It embraces things like automation, Internet of Things, AI, biomimicry (because it’s about converting the real world into something we can use, and better understand, and biomimicry is exactly that).

Digital transformation is taking data that can now be connected to any other kind of (digital) data, and build new ideas, business models, industries, disciplines etc, that weren’t available or apparent to use before, so it makes sense that the real value is going to lie in places we haven’t dreamed of yet.

Real digital transformation

Real digital transformation

That’s the distinction I make between digitalisation and digital transformation. Digitalisation is the process of adding value to digitised data, improving business processes, making them more efficient. Digital transformation is the process of transforming how that data is used in innovative ways that change industries entirely.

In short:

  • You can’t digitalise any process until the data it spews out has been digitised.
  • You can’t transform a process until you’ve digitalised it — applying digital technologies to the data you’ve digitised.
  • When you transform a process you change it fundamentally, recognising and realising the opportunities digitalisation could unleash.

So, a final example to clarify what I think are the differences.

Let’s take a heat sensor (thermometer) attached to a machine.

The sensor readout itself could be digital, but if you’re writing down the readings in a book the data becomes analog. It needs to be digitised — entered into a tablet, and then into a spreadsheet, say. Or the data could be drawn straight from the sensor itself. That is, arguably, digitisation. I would argue it’s digitisation because it’s still part of the process of converting analog data into digital. I would say that unless you’ve got to the point where all your key data are digits, you’re not digitised.

Once the data is there, you can digitalise both it and the processes. The first step is converting it to a form that is intelligible to the rest of your processes. The data has now been digitalised. And then, the next step of digitalisation is to digitalise the process — where the sensor is read by a computer and an automated warning light goes off to signal when there’s a problem.

Digital transformation occurs when this process is overhauled so that the business itself is transformed. It might just be transforming the process — robots replacing workers, say — but that is just a step in a much longer process when you don’t just replace one sort of tool with another, but actually change the way the widget is made, or sold, or change the widget itself. In the case of the sensor, it would be first to automate not just the monitoring and warning process, but then automating the repair work, the replacement, or using AI to learn how to improve the lifetime of the machinery, or the optimal process for replacement in conjunction with other data about other machines, prices, time of day etc. The next step would be redesigning the machine itself based on the lessons drawn from the digitalisation. It could be to transform the business entirely, by using the data to improve the business model (XaaS), using different processes, or to get out of the business altogether.

Digital transformation is a journey (much as I hate the word, which business has rendered meaningless or obfuscatory, depending on the context), not a step.


I am well aware that I am not using the terms as some use them. And I am happy to be corrected by those who can show me I’m misunderstanding the underlying processes. But hopefully this will prompt a discussion, or at worst some brickbats.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.