Professional Documents
Culture Documents
3
SHIFT 22 Editorial
4
‘DIGITAL IS 5
STILL IN ITS
INFANCY, Welcome to Shift, the annual
FRESH AND Shift is the brainchild of In The Pocket’s experts, who work in
the proverbial trenches of product and platform develop-
VIBRANT’
ment. This document is meant for you, the professional that
wants to stay ahead of the curve.
From inclusive design to low code, from crypto to the
metaverse: if it needs to be on your radar, you will read
about it here. More than ever, the digital landscape is in
flux. On your own, it is impossible to get a handle on every
important trend in the digital realm. You need guides to
navigate you through the trends that matter and that’s the
role we hope to play.
If I personally had to draw one conclusion from the
10 articles in this report, it would be that digital is still in its
infancy. It is young, fresh and vibrant. We have so much left
to learn. For companies that find the agility to adapt and
the courage to explore, the future is bright. This is only the
beginning.
08 48
6 7
Europe’s product Hey ML, meet UX
challenge
56
18 Living on the edge
Data in healthcare:
time for action 66
01 9
EUROPEʼS
Koen Boncquet
Tribe Lead
PRODUCT
CHALLENGE
SHIFT 22 Europeʼs product challenge
take a serious look at the United and Mathematics) and houses some excellent Computer
Science departments in top universities such as ETH Zurich or
States, as the epicentre of digital Cambridge University. If that’s not enough, Europe tops the
charts when it comes to the number of smartphones in use.
entrepreneurship in our current Still, the important product companies that put their
stamp on our modern way of living originate almost solely
information age, and compare it in the US. Companies like Google, Facebook, Snapchat and
against Europe. Netflix have been immensely successful and even managed
to influence contemporary culture worldwide. If you look at
the 100 largest global companies, it’s striking that only three
European tech companies show up on the list.
It’s also remarkable that since 2010 the early-stage
startup investments in California alone double that of all of
Europe. One can only look at this and conclude something
doesn’t add up… How come the EU doesn’t make better
digital products?
A scattered landscape
TLDR Europe has a product problem. Although we have a
One of the main reasons is Europe’s diversity. The EU
bigger market than the US, we’re trailing behind in
the worldwide digital product race. In this article, we zoom in on the consists of many different markets that product companies
reasons for this product challenge and look for possible solutions and entrepreneurs cannot treat as one. Scale-ups struggle
to put Europe on top of the charts. because the European landscape is too scattered. Europe
SHIFT 22 Europeʼs product challenge
12
Tech startup 13
investors
may have more people than the US, but its economy is
essentially a patchwork of different markets. Every other
in Europe
market comes with custom internationalisation or even
different regulations which mean your products have to
change. Consider the real estate and fintech domains: both
have a lot of country-specific regulations, creating a hard
border, not only physically, but also digitally.
Making matters worse, compared to the Anglo-Saxon
model, the economic model used in European societies is
much less flexible and less suited for rapid technological
change. The American economy and the Anglo-Saxon ap-
proach, in general, have been more successful in coping
with this technological change. Not surprisingly, the UK
consistently tops the European charts for investments and
startup successes.
14 15
Europe. The United Nations forecasts the emergence of We strongly believe that the best product people should
10 additional megacities by 2030, of which more than half progress from engineering, design or other roles that
will be in Asia. have actually been hands-on in building successful digital
So, looking at the numbers, you can see how Europe is products.
actually a scattered digital landscape, with a lot of variation So it seems that we mostly have a ‘quality product
in maturity. The northern markets are leading the way, but problem’, not a lack of great people or a willingness to go
aside from some urban talent concentration in the main there. It’s an underdeveloped and scattered ecosystem,
cities, a lot of continental Europe is three to five years be- undervaluing core roles that excel in building great software
hind. The fact that too few product companies arise is first products. Mature markets simply value product manage-
and foremost a geographical ecosystem problem. ment and design a lot more, leading to better products.
16
Better Products,
are interested in exploring product development opportuni-
ties. Understanding product strategy, analytics, tactics, and
complex technical challenges should be the goal.
Bigger Futures.
The business world is well-positioned to take the lead
since it’s not a discipline you can typically learn during your
university years. We need to develop more people with these
highly sought-after skills and make them the entrepreneurs
that help build the next European unicorns.
Once we acknowledge what is holding Europe back
and why it’s holding us back, we can move forward with a
We help you create top-tier digital products
new force and maximise the talent and professional energy to ensure a bright digital future.
that is abundant in our beloved continent.
18
02 19
DATA IN
HEALTHCARE
Bo Smet
Product Strategist
TIME FOR
ACTION
SHIFT 22 Data in healthcare: time for action
correctly and fail to process what it a truckload of data since the Covid pandemic began. But
that doesn’t mean this has influenced our behaviour sig-
means for users. nificantly. We have seen first hand — and on an unprece-
dented scale — that too much data risks drowning out the
message. This is not a pandemic-specific issue, but rather
Covid has highlighted a bigger problem: it’s no longer the
question of how to collect the data, but more than ever
about what to do with data.
22 23
no longer a conscious choice. The massive data footprint the current states and suggest the next action to take. For
this leaves behind makes gaining any valuable insights a example: when showing a positive virus test result, suggest
challenge. The signal-to-noise ratio becomes substandard going to the doctor to discuss treatment.
and so we look at algorithms to tell us what is important
and noteworthy. And therein lies the biggest weakness: the 4 Communicate insights with visuals, numbers and
reliance on opaque mechanisms that interpret data and words. Make sure your message is accessible to all. Com-
make decisions for us. bine numbers, text, colours and graphs. Colour coding can
make values instantly readable and graphs are ideal for
comparing with averages.
Designing data-rich applications
How should we deal with the abundance of data? 5 Protect the privacy of your user. When dealing with
From our own experience in building data-rich applica- sensitive medical data, don’t assume the user’s device is
tions, we’ve put together a few pointers on how to design always used in a private setting. Try to put as little private
an effective interface around numbers and insights that information as possible on your app’s launch screen.
offer clarity and real value for users.
24
Better Products,
night of course. The Fogg model teaches us that there are
different ways of changing behaviour. From starting a new
behaviour, all the way to stopping unwanted behaviour.
Bigger Futures.
Whatever the pattern, change starts with taking small steps;
reminding someone to take a run at a time when you typi-
cally see very low movement is a good example.
Once you succeed in triggering users to implement
these one-time changes, you can focus on creating patterns
of behaviour, such as going running every day for a month.
Andthat’s the main goal: to implement lasting change in
All companies are transforming into digital
a person’s behaviour and to offer a truly meaningful expe- versions of themselves. Some fast, some slow,
rience. To make data actionable.
but all have entered the arena of digitisation,
automation and UX. And so have you.
26 27
STOP
HIRING
FOR
03 Emma Braeye
VP People CULTURAL FIT
SHIFT 22 Stop hiring for cultural fit
30 31
needs nourishment. The future talent of your organisation look beyond the safe environment of what they already
should bring that nourishment to the work floor. Instead of know. Automatically, you’ll create a much more diverse and
looking for the perfect cultural fit, it pays off to find some- open-minded workplace.
one who doesn’t fit in seamlessly. Someone who brings new The fertile soil for an evolving company culture comes
ideas, or someone who dares to challenge current mindsets from new hires. Your new rule of thumb when recruiting new
to break the status quo. Those people will benefit your com- employees should be to assess their cultural contribution
pany as they bring diversity and innovation to your offices. capabilities. Try hiring people with diverse backgrounds,
They break the cycle. Don’t go looking for the egg that fits skills and personalities and avoid the pitfall of picking the
the dozen, but look a bit further for the odd one out. usual suspects. Don’t start the hiring process with a set of
boxes to tick to see whether the person fits in your company,
but rather prepare questions that inquire about their indi-
Keeping culture alive vidual potential and personal views on certain processes or
So how should you break the cycle? Although we say strategies. You’ll reap the benefits later.
that you need to stop hiring for cultural fit, you don’t want to
stop assessing, defining and articulating your culture along
Talent nourishes culture
with its values, goals and practices. Recognising your current
culture is the first and most important Company culture is a cornerstone of every modern
step towards starting to challenge it. business. And it’s essential to keep the noses in the same
Recognising your
Don’t be afraid to welcome new influ- direction. Hiring for that cultural fit instinctively seems like
ences, and accept that your culture will
current culture is the the right thing to do, but it carries a huge risk of smothering
change over time. It’s normal, it’s healthy first and most impor- the culture that keeps your company on edge. Today, it’s
and it allows for your company to dis- tant step towards about cultivating your culture and giving it room to grow.
cover new pathways and set up a path starting to challenge Hiring new people with new insights is essential to keep
for growth. it. Don’t be afraid to your culture alive and growing. The next time you’re hunting
Next, promote diversity and en- for talent, instead of focusing mainly on a cultural fit, ask
welcome new influ-
courage people to disagree. Avoid be- yourself how that person will contribute to the success of
ences, and accept
coming a monolithic block by creating your organisation. What’s that saying again? Contribution
that your culture will
a safe space for new ideas and don’t be eats culture for breakfast?
afraid to mix things up a little. If a new change over time.
joiner or colleague feels free to speak
their mind and introduce new ideas, it opens the way for
workplace culture to evolve and encourages teams to
SHIFT 22 Inclusive design: beyond regulations
32 33
INCLUSIVE
DESIGN
Stephanie Depuydt
Product Management Lead
BEYOND
REGULATIONS 04
SHIFT 22 Inclusive design: beyond regulations
36 37
medical problem that needs to be prevented, cured or con- mantra: inclusive design isn’t about creating a specific de-
tained. The problem is the disabled person. But is that really sign for disabled people; it’s about creating a flexible design
the case? What if we looked at disability from a totally differ- that needs to work for everyone in any given situation.
ent angle? Instead of focusing on the disabled person, try to
view them as people who are disabled by barriers in society
It’s not a cost, it’s an opportunity
and not by their impairment or difference. Does the prob-
lem lie within the person in a wheelchair, or does it lie within Let’s crunch some numbers. According to The Click-
the buildings without ramps? In this societal model, exclu- Away Pound Report, 83% of users with accessibility needs
sionis not an inevitable consequence limit their shopping to sites they know are barrier-free. 85%
of having an impairment. Disability is of the respondents say businesses lose out when they don’t
A solution that
not a personal medical condition, but a cater to digital access needs, according to a survey by
mismatched human interaction. aims to overcome Inviqa. Research by Q42 shows that a staggering 43% of
It makes more sense to consider in- a certain barrier Dutch interviewees admit using one or more accessibility
clusiveness as taking away as many bar- often helps a much features on their phones.
riers as possible. Why? Because everyone larger group of We have a considerable group of people with a per-
encounters barriers one way or another. people than you manent disability, but the actual audience that appreci-
So a solution that aims to overcome a ates inclusivity efforts goes way beyond that. Embedding
would expect.
certain barrier often helps a much larger inclusive design, and going beyond the obligatory checklist,
group of people than you would expect. shows real and massive potential for growth. It’s one of
Let’s take visual impairments as an example. You’re those rare win-win situations. It increases your audience
considering adding audio descriptions in your app for the reach as your features include more and more people. It im-
visually impaired, but you’re afraid it’s a big investment for proves your product’s usability, not only for the unfortunate
only a small number of people. That might be right if you few but for everyone. It leads to an increase in customer
believe it would only benefit permanently blind people. loyalty as users can use their favourite app, no matter the
But that’s simply not the case. Broaden your view. Consider specific barrier they experience at any given time or place.
the huge number of people who are starting to experi- Last but not least, it definitely improves your overall brand
ence problems with their eyesight. Or think about situations perception. It’s cool to care.
where we’re not able to look at our screens, such as when
we’re driving the car or riding a bike.
Integrating inclusive design
Everyone experiences disabilities or barriers, whether
permanently, temporarily or situationally, so we should We hope that by now you have realised that acces-
embrace inclusivity in the products we ship. Try this for a sibility and inclusive design isn’t merely about complying
SHIFT 22 Inclusive design: beyond regulations
38 39
with some new set of rules and regulations, but that it is, in 2 Build and spread knowledge internally. Start an
fact, both an opportunity for you and for your users. User ‘Inclusive Design Domain’. Look for a group of
outcome beats compliance. However, make no mistake: cross-functional employees who feel passionate about in-
integrating inclusive design is not a one-off, nor is it an add- clusive design and are willing to make a change in your com-
on. You won’t achieve it by checking off a list of things and pany. They can be the starting point for sharing new insights
carrying on as usual. It needs to become fully embedded and the go-to team for any accessibility-related questions.
in your company, your processes and your culture. No easy And get them to document their learnings. Let them partici-
task, but here’s something to get you started. pate in meet-ups or talks in your company to spread aware-
ness. And, perhaps most importantly, include relevant users
1 Determine the status quo and look for the way as quickly as possible. Interviewing users always leads to the
forward. To get a first impression of where you’re at best, most honest and specific feedback on your products.
as a company, try selecting some flagship sections of your
product and start user testing these for accessibility. Don’t 3 Process integration. Make sure your inclusivity efforts
trust your own assumptions, but invite impaired users to put are durable by integrating them into processes. There’s
your inclusiveness to the test. Exclusion is a proactive choice, no exact way to do it; the most important part is to just start.
so ask yourself who you are currently excluding and who you It’s different for every organisation. Having self-empowered
are willing to exclude. This question isn’t as loaded as you or autonomous delivery teams, however, will make it eas-
might think. Banking apps don’t want children to use their ierto adapt processes within teams and let them spread
products, right? So they exclude them by choice. throughout your whole organisation. You’ll see how fast
it changes colleagues’ mindsets if you make them aware
of the topic. They’ll perceive it as a new challenge, which
will have a positive impact on the world around them.
Knowledge building not just hype; it will become the new standard, enforced by
European law. Instead of acting out of fear of becoming
non-compliant, you should look beyond the legislation and
Knowledge building isn’t easy and it shouldn’t
tap into the wide-open space of opportunities this creates
be underestimated. Start small and try picking for you and your company. Better today than tomorrow.
a focus, like visually impaired users. It should
get you on the right track and you’ll see
concrete results quicker than you think.
SHIFT 22 Shifting left on security
40
05 41
SHIFTING
LEFT ON
Louis De Jaeger
Security & Privacy Officer SECURITY
SHIFT 22 Shifting left on security
44 45
STATE
is a successful consequence of the shift-left approach, but
it doesn’t encapsulate security. So, from now on, let’s try to
OF
think in a DevSecOps mindset: we don’t only want to shift se-
curity to an earlier stage in the cycle; we want to integrate it DEVOPS, 2021:
in every step along the way in order to accelerate the cycle.
The main goal of shifting left on security is preventing
failure by performing security checks earlier on. In prac- Consistent with previous
tice, engineers will catch and solve potential security issues
reports, we found that elite
before they can become a real problem. With DevSecOps,
teams can continuously take security into account when performers excel at imple-
working on their products. Inevitably, the shift-left approach menting security practices.
will result in cost-saving (the sooner you detect, the cheaper
the fix), a faster time to market (eliminating security bot-
tlenecks), the mitigation of risks and the creation of a solid This year, elite performers
security culture in your company. who met or exceeded their
reliability targets were twice
Automation and security: it’s a match as likely to have security
Apart from the obvious benefits a shift-left approach integrated in their software
brings about, there’s more to gain below the surface. When
development process. This
embedding security early on, and at each stage of devel-
opment and testing, your team will save tons of time. The
suggests that teams who
key to that time-saving lies in automation. It may sound have accelerated delivery
counterintuitive, but automated security testing does a while maintaining their
splendid job on condition that it’s applied early enough in
the development lifecycle.
reliability standards have
The market offers lots of tools that can help you shift found a way to integrate
left and implement automation through your pipeline. security checks and prac-
Static or Dynamic Application Security Testing (SAST/DAST)
tices without compromising
automatically checks for vulnerabilities in your application.
Interactive Application Security Testing (IAST) combines their ability to deliver soft-
these two and is typically employed as some kind of agent ware quickly or reliably.
SHIFT 22 Shifting left on security
46 47
into the runtime environment that observes operations and team will still need guidelines, best practices and new pro-
attacks. Dependency Scanning, on the other hand, scans cedures that are developed and maintained by the core
for flaws in dependencies — a must-have to check your security team.
open-source libraries — and Secrets Detection guarantees Downsizing your InfoSec team, failing to collaborate
that you have no compromised secrets such as database with them or engaging them too late are the most common
credentials or API keys. pitfalls that prevent teams from shifting left on security. If
In essence, automation means less room for human your information security team is understaffed, they are
errors, increased code coverage and no issues in your pro- not able to address all risks and oversee all the processes
duction environment. If you want to shift left on security and people that they need to. It leads to situations where
without automation, you will not succeed. they are not able to collaborate with your software teams,
something that is vitally important in this era of continuous
cyber-attacks.
Don’t move responsibilities, share them
Shifting left is not about moving responsibilities, it’s
Assess, learn and implement
about sharing responsibilities. By aligning development,
operations and security, we can start talking about a true Starting with shifting left on security is not always easy
DevSecOps approach. The cross-pollination between these and we can’t give you a one-size-fits-all solution. What’s
branches inevitably leads to more awareness and educa- most important is that you define a clear strategy. Decide
tion among the technical teams in your organisation. If where you intend to go and try to do this in a one-page
certain security processes become embedded in your way strategy that you can share with the company. Your strate-
of working, developers, for instance, will automatically gain gy will help every colleague to understand what successful
a better understanding of how security protocols work and implementation looks like for your organisation. Include
how they positively impact the outcome. vision, ownership structure and metrics into the document
Does that mean that security teams We don’t only want and follow a ‘plan-do-check-act methodology’ to mature
become obsolete in the long term? Ab- to shift security to your strategy over time. It doesn’t have to be perfect from
solutely not. Empowering DevOps to the beginning as long as you’re willing to stick with it.
an earlier stage in
perform security testing doesn’t equal Over time, you’ll see that shifting left on security will
the cycle; we want to
replacing the core security team. To make help your organisation reduce risk, improve security and gain
sure your shift-left process is functioning
integrate it in every efficiency in resolving potential issues. What’s not to love?
properly, you still need to perform pen- step along the way
etration tests or audit access rights and in order to acceler-
logs. On top of that, your development ate the cycle.
SHIFT 22 Hey ML, meet UX
48 49
06
Rémi Delanghe
ML Engineer
Kenny Helsens
ML Lead
52 53
All too often we blame products and companies for To empower the user, instead of infuriating them with
using AI in order to make money, but we think too little about a system that will never be perfect, we have to identify what
the unintended consequences AI has on the user experience the user is willing to offload to an ML system and what the
side. So, instead of cursing the past, we should focus on what user wants to control. Repetitive, boring, data-intensive or
we’ve already learned and improved in recent years and time-intensive tasks, producing only intermediary results,
start looking to the future. are often good candidates to offload to ML. Tasks or actions
with a high impact, a high cost of being wrong, or nuanced
results are better off staying in the user’s control.
ML For Users An excellent example of this balance is Google Maps.
When we think about how ML in products affects the After entering the start and end-point, the user offloads the
user, two considerations take centre stage: ‘Expectation task of finding the best route to the ML system. The system
Management’ and ‘User Control.’ responds by offering multiple possible paths, information
In traditional digital product design, we assume about the trajectories and the possibility of adding more
deterministic behaviour from the components we use. When waypoints to keep the user in control of the final result.
we press ‘send’, we expect our mail to travel the web. We
learned over the years that pressing ‘control+z’ will undo
ML For Product Teams
our last action the same way that flicking a switch turns on
a light. And when it doesn’t, we freak out. The point is that Creating ML-enabled digital products that integrate
we have built up trust and expectations around particular ML in a meaningful way for the end-user is a team effort. It
digital objects, and breaking these expectations causes a involves both the engineers, the designers and the team’s
lot of friction for the user. product manager.
Machine Learning is the opposite of deterministic. The product manager needs to ensure that user needs
The predictions are probabilistic, and their quality depends are captured correctly and know what parts of the process
both on the data the model learns from are worth automating or augmenting using ML.
and the context we use the model in. To empower the user The engineering team needs to master a wide range
These properties lead to opinionated of technologies within and outside the realm of ML, under-
we have to identify
results, and when designing interactions, standing the powers and drawbacks of these technologies
what the user is
we have to frame the results along with in the context of the specific user problem. The resulting
willing to offload to
their uncertainty. The results should be design space enables designers to create a wide range of
presented not as an objective fact or an ML system and interactions that solve real user needs and give them the
magical solution, but as a prediction that what the user wants freedom to frame the AI predictions and capture feedback
can be wrong and should be challenged. to control. correctly. Ultimately, this feedback updates the product
SHIFT 22
54
talent@inthepocket.com
www.inthepocket.com/careers
SHIFT 22 Living on the edge
56
07 57
LIVING
ON
THE
EDGE
Ilja Strobbe
Software Engineering Lead
Jonas Goderis
Cloud & API Lead
Jonathan De Moor
Cloud Architect
SHIFT 22 Living on the edge
60 61
EDGE
COMPUTING back towards more decentralised structures. What does
edge computing have to offer that’s attractive for both
Edge computing brings data industries as users?
processing closer to the data source. Less latency. Edge computing finds its origin in the
early CDNs (Content Distribution Networks) where caching
servers would be placed in places geographically closer to
the end-user to reduce latency. The physical limitation of
signal latency is not acceptable in mission-critical applica-
tions expected to give near real-time performance.
More autonomy. Bringing computing closer to end-
users reduces the chance of network problems in a distant
location affecting local customers. And it can bring connec-
tivity to places where there’s no connection at all. Even in the
event of a nearby outage, the edge devices will continue to
operate effectively on their own.
Savings-friendly. By filtering out irrelevant data on
the edge, one can reduce network bandwidth, data storage
and processing in the centralised cloud. Spreading this
‘pressure’ could significantly affect the running cost of a
cloud platform.
62 63
A win for privacy. Applications that require private Truck platooning is the linking of two or more trucks
data can benefit from edge computing as well. Instead in a convoy, using connectivity technology and automat-
of analysing all private data in one centralised cloud, eddriving support systems. These vehicles automatically
advanced machine learning models could be deployed on synchronise their speeds and routes and maintain a close
the edge. The private data of end-users can be processed distance between each other. The truck at the head of the
much closer and quicker, limiting the risk of a breach in the platoon acts as the leader while the trucks behind it will
cloud exposing data. react and adapt to changes in the leader’s movement.
This new way of transportation, working on the principles
of edge computing, offers quite a few benefits:
The edge in practice
Leaving theory to one side, let’s look at some interest- Ecological. Given that trucks can drive closer together,
ing case uses for edge computing. Take autonomous vehi- the air resistance force is reduced significantly and will result
cles, for instance. When we’re thinking about autonomous in lower fuel consumption and CO2 emissions.
vehicles, Tesla and their self-driving cars will almost certainly Safety. IoT sensor data is analysed on the vehicle
pop into our heads first. But most likely the first use cases of edge to provide automated and immediate responses,
real autonomous driving will be truck platooning. like breaking. Vehicles following the leading truck only need
DRIVER,
container truck to the platoon vehicle detection, de-coupling to
leading 3* driver- via wireless anti-collision and allow other road
less trucks communications lateral control users to cross
technologies for between platoon
MULTIPLE
TRUCKS
*Number of trucks in each
platoon may vary according
to trial results.
SHIFT 22 Living on the edge
64 65
a fifth of the time a human would need to react, thus im- to assist safely, quickly and accurately. If these devices rely
proving safety. on submitting data to the cloud and waiting for a response
Efficiency. Platooning optimises transport by using before making a decision, it could be fatal for the patient.
roads more effectively, delivering goods faster and reducing
traffic congestion.
On the brink of a breakthrough?
With healthcare and autonomous vehicles, we are
Changing entire industries only scratching the surface. Smart cities, for example, aim to
Edge computing isn’t limited to very specific use cases; improve the quality of life for residents with smart parking,
it’s likely to affect entire industries, such as the life sciences traffic management or improving security by proactively
sector. Before 2020, digital transformation in healthcare monitoring public spaces and law and order.
was pretty slow, but over the past year, Likewise for the oil and gas industry; if there are any
the industry has speeded up efforts to Rich Bird: ‘We’re not problems or unusual activity, this can be detected instantly.
embrace digital transformation. The replacing doctors or Edge computing allows us to analyse the data immediately,
sky’s the limit when it comes to the op- identify the cause of the issue and take the appropriate
care teams with this
portunities for using edge computing in actions to rectify the situation long before significant dam-
technology, we’re
healthcare, says Paul Savill, Senior Vice age occurs.
President of Product Services at tech helping them make Even in the financial sector detection of fraudulent be-
company Lumen. quicker and better- haviour often happens post-event. Banks detect suspicious
Many devices currently out there informed clinical transactions and take steps to prevent further losses by free
(for example health monitors, sensors decisions by gener- zing cards or accounts. Running AI-enabled analytics at the
and wearables) are either not connec ating insights from edge could detect fraudulent patterns and proactively pre-
ted,or are collecting a large amount vent fraud — benefitting both the banks and their customers.
providing data that
of unprocessed data that needs to be Let’s be clear, countless applications will leverage the
helps improve the
stored in the cloud. As healthcare often possibilities offered by edge computing to create solutions
deals with very sensitive information, this
outcomes for their that were previously deemed infeasible or even science fic-
creates privacy and security concerns patients.’ tion. As always, adoption will not happen overnight, but one
that require a lot of attention. An edge step at a time, offering real-world improvements to already
on the hospital could process data locally, protecting data existing cloud solutions. The real question is: are you ready
privacy and instantly alerting caregivers when unusual to start living on the edge?
trends occur with a patient and help is required. Also, robots
assisting in surgery must be able to analyse data instantly
SHIFT 22
66
08
Bram Vandenholen
Solution Architect
HOW
LOW-CODE
CAN YOU GO?
SHIFT 22 How low-code can you go?
70 71
HOW BIG IS THE
OPPORTUNITY? 4,45
These figures back up why every
IT and digital professional should
BILLION
2020
have low-code on their radar.
65%
of app development will be
low-code by 2024
(Gartner)
PROJECTED
EXPANSION
75%
The enterprise low-code
application platform mar-
14,38
ket is projected to expand
from $4.45 billion to $14.38
of development shops will use billion between 2020 and
some form of low-code by end
of 2021 BILLION 2025, with a compound
annual growth of 26.4%.
72 73
streamline workflows or even automate processes. In a way, poorly-secured application in the stack can cause security
it’s a democratisation of IT, allowing every employee to or privacy issues all over the place.
become a ‘citizen developer’. Stability/testability. Since ‘works for me’ is never
The benefit is obvious: ‘business’ no longer needs to good enough for enterprise-grade software, always be
lean fully on ‘IT’ to get digital solutions in the hands of cus- mindful that unit testing and end-to-end testing for low-
tomers or staff, thus saving costs and speeding things up. code isn’t as well established as in more traditional software
To give one obvious example, low-code solutions proved engineering.
tremendously useful at allowing organisations to quickly Debuggability. No-code tools can sometimes be dif-
respond to the pandemic. Customer-facing experiences ficult to debug if something is amiss somewhere in the flow,
were set up in a matter of days, not months. Internal work- especially if code is distributed over snippets of scripts.
flows were adapted without a complete rebuild of the Maintainability. Software is a living, breathing thing,
enterprise resource planning. meaning things will change. API’s will change, the data
With such promise, it’s no wonder the low-code model will change, and logic will change. You will have to
space is booming. But is this truly the future of software maintain a helicopter view and update the various compo-
development? nents without impacting users.
Scalability. If your digital solution proves to be a hit,
can you scale it up using the low-code stack or will you have
There are always limitations to start from scratch?
As everyone in software development will tell you, Supportability. When (not if) things break, who do
there are always trade-offs. Low-code has speed and you call? How are the SLA’s of the low-code tools you use?
democratisation as two huge aces up its sleeve, but they Extendibility. Given you’re depending on integrations
can create some cons as well. Whether the pros outweigh offered by low-code tools, do these have a future-proof
the cons depends on each specific use case and context; roadmap so they too will evolve over time?
what’s crucial is that these quality attributes are discussed Usability. The UX will probably be good enough, but
upfront before you make your technology choice. is that really good enough? What about accessibility re-
To get you started with your pro and con list, we’ve quirements, localisation? Sublime UX is what makes you
jotted down some typical quality attributes and trade-offs stand out after all.
(all ending with ‘y’) that are inherent to low-code platforms.
Just to show that some nuance will always be justified:
To no-code or not to no-code?
Security. Does not have to be a problem in itself,
but with ‘everyone being a developer’, you need to avoid Nuance is the only way to answer this question and
a Wild West of applications and shadow IT, where one end this article. The low-code movement really does have
SHIFT 22
74
User Research
to make.
Are you looking for speed in very straightforward
use cases, such as some business-line apps or very specific
at In The Pocket.
customer flows? Or do you want to kickstart a new busi-
ness without making huge customer software investments
upfront? Then let low-code help you on your way.
For business-critical digital products where user expe-
rience is a differentiator and where factors such as security
and scalability are fundamental, the trade-offs inherent to
low-code suites will probably come at too high a cost. The
There’s no way around it. When designing
future has many paths, choose wisely. and building digital products the user should
always be the focal point of attention. User
research is of paramount importance here at
In The Pocket, and we’re always eager to learn.
THROUGH
76 77
THE WEB3
LOOKING GLASS
09 Christophe Rosseel
Chief Service Delivery Officer
SHIFT 22 Through the web3 looking glass
G
CLIENT
DATABASE
SERVER
CLIENT
DATABASE
80 CLIENT 81
A
DATABASE
CLIENT
THE
SERVER
DATABASE
CLIENT
one’s head around because it challenges our assumptions
INTERNET
F
CLIENT
about how the Internet should work. Let’s start there.
DATABASE
While the Internet started as a grassroots movement
SERVER
TODAY
CLIENT
DATABASE and an open architecture, it slowly evolved into a set of
CLIENT closed silos. Under the dominant paradigm of web2 apps,
A
DATABASE
context does not travel from one app to another. The im-
CLIENT
Client-server architectures
SERVER plicit assumption is that a single economic actor controls the
DATABASE business logic on reading from or writing to its applications.
CLIENT
on centralised platforms. This architecture has many impli-
cations. You need an account for each
entity and your data is not interoperable Tradeoffs/
with other applications in the system. As client-server
CLIENT
DATABASE
a user, you have to trust many different architecture on
SERVER companies to ensure the integrity of your centralised
CLIENT
DATABASE data. platforms
CLIENT Then there are economic conse-
CLIENT
SERVER
DATABASE quences. When you upload a picture Pros
to a social network, you transfer some
DATABASE • Network effects
CLIENT rights to the network. The social network
• Scalability
is entitled to aggregate the data of their
CLIENT
DATABASE
users and monetise it. Usually, the plat- • Fast iteration
SERVER
CLIENT form doesn’t share. Why is that?
Cons
DATABASE
A lot of value and power accrues
CLIENT
DATABASE
to a few centralised platforms that dic- • No shared state
CLIENT
SERVER
tate the rules of the game. When these • Single point of fail-
platforms are small and competing for
CLIENT
DATABASE
ure, data integrity
users, their relationship with other net-
is not guaranteed
work participants is positive-sum. As
they move up the adoption curve, their
• Economic incen-
power increases and the relationships tives lead to mo-
change to zero-sum. Examples include nopolistic behavior
Facebook and social game developers • Platform risk
SHIFT 22 Through the web3 looking glass
82 WHY 83
DECENTRALISATION
MATTERS
like Zynga or Twitter and third-party clients like Tweetdeck.
The latter parties invested a lot of time and money in their
ventures to find themselves de-platformed when their host
system no longer needed them.
We landed on this set of trade-offs when we were up-
loading cat photos to these platforms. The question is: do
we want or need to make different trade-offs when we are
talking about money or identity?
01
What does web3 have to offer?
The key feature of web3 is codified trust. In 2008,
Platform’s
Satoshi Nakamoto famously resolved an age-old compu- relationship to
tational issue known as the Byzantine Generals Problem: complements
how do you get nodes in a peer-to-peer network to coor-
dinate and reach consensus when nobody knows or trusts (developers, creators,...)
each other? The answer, it turns out, is through a consensus
algorithm that incentivises the participants to cooperate.
Bitcoin miners, for example, are rewarded with the BTC
token to secure the network and avoid double-spending.
Public blockchains also have disadvantages. Current-
ly, they consume a lot of energy, and it’s tough to scale writ-
ing to them. Ethereum can only handle seven transactions
per second right now. Many problems remain to be solved.
And yet, a lot of smart people are very excited about this
tech. Why? 02
One obvious extrapolation of codified trust is the
Platform’s
so-called creator economy. If groups of strangers can come
relationship
84 85
FAT
PROTOCOL YouTube, Patreon or OnlyFans control the rules of distribu-
tion and allocation of proceeds. In web3, creators could set
these rules and codify them into immutable smart contracts.
WEB 2 Furthermore, in web3, context can travel. Instead of an
account per silo, you now have a wallet. The metaphor isn’t
perfect; your web3 wallet does not store any funds — it holds
the keys to your funds. Banking on asymmetric encryption,
VALUE CAPTURED
APPLICATIONS you can share your public key with others to receive trans-
LAYER
actions. You must never share your private key as it can be
VALUE CAPTURED
WEB 3
APPLICATION LAYER
are a native feature of web3. This new ability of software
to channel money is an exciting prospect for creators of
digital products.
VALUE CAPTURED
APPLICATION LAYER
It’s not just your payment data that becomes owned
and portable; blockchains such as Ethereum can store
VALUE CAPTURED
PROTOCOL
LAYER and execute any combination of data and business logic.
Ethereum is quite literally a world computer. This is what trips
PROTOCOL
LAYER up a lot of people. Our existing mental models have not
prepared us to think about the possibilities of a distributed,
global computer with a shared state.
Another way to approach it is to think of web2 as hav-
ing a ‘fat’ application layer on top of a ‘thin’ protocol layer.
The Internet stack, in terms of how value is distributed, is composed of ‘thin’ protocols Facebook, Google and the like, capture and re-aggregate
and ‘fat’ applications. This relationship between protocols and applications is reversed most of the data and resulting value in their applications.
in the blockchain application stack. Value concentrates at the shared protocol layer Web3 reverses this. It stores user data on the protocol level.
and only a fraction of that value is distributed along at the applications layer. It’s a One of the consequences is that applications will not have
stack with ‘fat’ protocols and ‘thin’ applications. the same moat they have in web2; barriers to entry are
Source: Joel Monegro, Fat protocols lower and apps are more interchangeable.
SHIFT 22 Through the web3 looking glass
86 87
88 89
excited about composability: the ability to take existing When you combine a permissionless ecosystem with com-
programmes and adapt or build on top of them. Let’s take munity standards and open source, composability emerges.
a quick look at DeFi to understand how composability emer Suddenly anyone can take existing programmes or
gesand how it gathers pace. primitives and build on them without having to reinvent
In a few short years, the world of DeFi has managed the wheel. Once a problem is solved, its solution becomes
feature parity with traditional finance: yield generation, a digital Lego block, ready for the next developer to use or
lending, borrowing, hedging, futures and derivatives etc. change at their bidding. To illustrate the point; at one web3
Every financial primitive has seen a web3 counterpart arise university you can take a beginner course called: ‘Building
since 2017. Even more impressive, DeFi is creating new fi- a bank that isn’t a toy’.
nancial products, like lossless lotteries, self-repaying loans We are still in the early stages of this innovation; time
and savings accounts that generate actual yield. Caveat will tell how transformative all this turns out to be. This
emptor. How is this possible? First of all, web3 is permission- author never bets against the Internet but as they say in
less. You don’t need anybody’s permission to build an app web3: do your own research.
on a public blockchain; you just do it.
Ethereum, for example, issued standards
to facilitate this. ERC-20 is an important Web3 is permission-
standard, enabling smart contract pro- less. You don’t need
grammability and interoperability be- anybody’s permis-
tween (fungible) tokens. sion to build an app
Another quality attribute of web3 on a public block-
is that it is immutable. A smart contract,
chain, you just do it.
once deployed, cannot be updated. It
can only be forked. The sword of immu-
tability cuts both ways. In the creator example, it reassures
our creators that the network won’t change the algorithm
that redistributes the proceeds. On the other hand, immuta-
ble smart contracts do not let you go back to fix a bug or
a vulnerability.
This permissionless and immutable nature is an incen-
tive for developers to open source their code. If you were
going to play around in DeFi, would you rather lock your
funds in an open or a closed source, smart contract? Exactly.
SHIFT 22 Meditations on the Metaverse
90 91
MEDITATIONS
ON THE
METAVERSE Kenny Deriemaeker
XR Tech Lead 10
SHIFT 22 Meditations on the Metaverse
94 95
creator economy and break down the walled gardens of what the true meaning of ‘Metaverse’will ultimately end up
the tech giants. Microsoft, on the other hand, sees it as the being. That would be like someone in the early nineties try-
future of work, with a population now accustomed to work- ing to predict what the Internet would look like in 2022. But
ing remotely thanks to Covid-19 ready to fully embrace the we do think there are some parallels to be drawn between
virtual workplace. The crypto/web3 enthusiasts see a pros- the mobile revolution of the 2010s and this next Big Thing.
perous green field for blockchain, NFTs and decentralisation
of all kinds. And the list goes on and on...
Riding the curve
Big technological shifts tend to follow an S-curve: very
Zuckerberg’s gamble slow adoption and growth in the beginning, followed by
A lot has been said about why Facebook/Meta specif- a steep acceleration towards mass adoption and greater
ically is betting so heavily on the Metaverse as a vision and value, and finally a tapering-off as the technology becomes
as a brand. Recent scandals, increasing regulatory scrutiny fully mature and integrated. At that point, the technology
and the fact that their main product is being largely ig- may become the staging ground for the next S-curve, just
nored by Gen Z, can be interpreted as good reasons for the like the VR and AR devices we have today, which only be-
company to rebrand and refocus with an eye on the future. came possible thanks to displays and sensors developed
The other reason is Oculus, which was acquired by Face- and commoditised in mobile’s acceleration phase.
book in 2014. Their Quest 2, a consumer VR headset that has Having said that, better hardware alone is not enough
sold surprisingly well, is the closest thing to put your foot to the floor on the Metaverse S-curve.
we have to a mass-market XR device to- Interoperability standards, content creation and delivery,
‘That virtual concert
day, and it is about to become a promis- payments, networking, privacy, accessibility, responsive-
in VR with a crowd of
ing playground for Meta to start fleshing ness and many other aspects still pose many big unsolved
thousands will hap-
out the Metaverse concept with millions challenges.
of actual users. pen someday, but Where will the content come from? Concept videos
No coin is one-sided. The ‘coming not anytime soon.’ from companies like Meta and Microsoft show a world of
out’ of Meta also shows how far we’re still realistic avatars, holograms and other rich interactable 3D
removed from some of the concepts out there. For example, content, but these assets are not easy or cheap to produce.
virtual meetings and events on the Quest work fine but are Despite the ongoing consumerisation of 3D content crea-
limited to just a dozen participants for performance rea- tion with technologies like LiDAR and photogrammetry, the
sons. That virtual concert in VR with a crowd of thousands vast majority of digital media today is still flat and screen-
will happen someday, but not anytime soon. At In The Pock- based. How do we manifest all that valuable 2D content in
et, we are not quite ready to make any predictions about a spatial 3D context?
SHIFT 22 Meditations on the Metaverse
96 97
How do we deal with the less photogenic aspects of are unknown. Whoever is behind the steering wheel — Mark
online life in a spatial environment? The concept videos Zuckerberg, Tim Cook or company X — is likely to matter a
don’t show any ads, toxic online interactions, breaches of lot in the short term and less in the long term. It’s going to
privacy or technical issues; all things that are very much be an interesting ride up the S-curve either way.
present on today’s Internet. These will need to be factored
in and dealt with, preferably more effectively and trans-
parently than the way companies like Facebook have dealt
with them in the past.
Can the Metaverse be made as responsive as the
Web? As supporting hardware and infrastructure evolve, it
will take time to become globally available and accessible.
There will likely be levels, with a low-fi Metaverse experience
on less-capable hardware and in bandwidth-constrained
environments, and the full Pixar treatment on expensive
headsets with ultra high-bandwidth connections. How do
we organise interactions between users with a $199 Meta
headset and a $1,299 Apple Glass (hypothetically) without
compromising on features and UX?
98
Written by
Koen Boncquet, Bo Smet, Emma Braeye,
Stephanie Depuydt, Thijs Van de Broek,
Louis De Jaeger, Rémi Delanghe, Kenny Helsens,
Ilja Strobbe, Jonas Goderis, Jonathan De Moor,
Bram Vandenholen, Christophe Rosseel,
Kenny Deriemaeker
Contributors
Niels Dewelde, Louis Ghekiere, Hannes Van de
Velde, Jeroen Lemaire, Michiel Dheedene
Graphic design
Koen Surmont
www.windwaait.be
In The Pocket creates digital
products that make people
happy and businesses
grow. We work on the inter-
section of AI, AR, mobile,
web, cloud and world class
product design. If your
company is serious about its
digital future, partner with
In The Pocket.