Toxic technology

Part 1 of a series on Toxic Technology.

In 2018 I wrote about toxic technology, a short post explaining the threat organisations face from the legacy technology they accumulate. To explain the idea in more detail, I wanted to write more. This series of blogposts will cover a range of topics which contribute to toxic technology – the way teams work, the strategies we use, core operational processes, and market incentives. Later in the series, I will write about how to avoid, manage and mitigate the risks of toxic technology. This post is the first of many instalments, so if you’re interested, please do sign up for more.

Toxic technology is eating the world

In 2011 Marc Andreessen suggested that software is eating the world. He described the phenomena of new companies using internet-enabled business models to disrupt established markets. 

“we are in the middle of a dramatic and broad technological and economic shift in which software companies are poised to take over large swathes of the economy….all of the technology required to transform industries through software finally works and can be widely delivered at global scale”

Marc Andreessen, 2011

At the time he wrote this, in 2011, over two billion people used the internet, up from an estimated 50 million in 2000. Andreessen predicted that in the next ten years “at least five billion people worldwide [will] own smartphones [with] instant access to the full power of the Internet”. A decade later, there are an estimated 5.11 billion mobile users and 4.39 billion internet users globally. The majority of the world’s population are now internet users.

In the decade since, this pattern has continued as companies such as AirBnB, Uber and Snap inc. have disrupted markets. But a different pattern better characterises the most recent decade: not market disruption, but the accelerating use of digital technology in existing organisations. This has created an impact across diverse industry sectors such as government, finance, retail and transportation.  

New digital technology has been a trigger for widespread change in the public sector. Governments across the world are now transforming how they work using internet-era methods. The US Digital Service, UK Government Digital Service and e-Estonia movements led the way, and many more are following. In the US and UK, these changes represent a rebirth after an era of outsourcing, where investment in technology was principally done through procurement. Now governments and public sector bodies are building technically skilled workforces, and producing large volumes of their own technology. There are currently over 900 national and local governments and agencies contributing code on GitHub.

Disruption has also occurred across the financial sector. Large technology companies such as Apple, Google and Tencent have disrupted consumer-facing payment services. Fintech companies like Stripe, Square and Ant Financial have created innovative and popular products. But, at the heart of the financial system, established banks remain the dominant force. To compete with market entrants, conglomerates have invested large sums in digital transformation. BNP Paribas invested $3billion in 2017, HSBC $17billion in 2018, and JP Morgan $10.8billion in 2019. Challenger banks like Monzo and Starling represent a more direct challenge to established banks, but whilst their growth is rapid, they remain niche players in the global banking sector. 

Small to medium enterprises represent the majority employer in most countries and sectors. These small organisations collectively make a significant contribution to the software produced globally. Typical software produced by SMEs are systems to help with routine administrative tasks such as case management systems and customer records management systems. SMEs also produce millions of websites – whilst many are constructed using templated tools, many also involve writing bespoke code.

Outside professionalised software development communities, people use general purpose tools to create software. They may not identify as software developers, yet they create abundant software. Excel formulas, Microsoft Access databases and customised ‘low code’ platforms are examples. Millions more create software online, editing the markup code of their websites. Rudimentary knowledge of HTML (Hypertext Markup Language) and CSS (Cascading Style Sheets) can help them go beyond standard templates. Content building platforms like WordPress and Shopify democratise software development. This long tail could perhaps be the largest software sector – the user-generated content of the software industry.

A crisis in the sustainability of software

Software is growing in every sector, and within organisations large and small. Software is playing an increasing, vital economic and social role. But is it sustainable for software to keep eating the world? Do we have the resources to ensure all this software remains healthy, and effective? Many patterns exist to suggest this is not the case. There is a crisis in the sustainability of software.

When technology is not sustainable, basic cyber security and maintenance practices lapse. This causes organisations to experience data breaches with increasing frequency and at increasing scale. But many of these attacks and accidents are preventable. High-visibility outages are becoming more common as a result of neglected technology. Systematic records on service outages are not kept, making trends hard to observe. But, as technology in most organisations is ageing, it is reasonable to assume the trend is worsening.

The European Union introduced the General Data Protection Regulation (GDPR) in 2018 to strengthen pre-existing privacy legislation. Whilst remaining subjective, it has become harder to argue the compliance of a large legacy technology estate. Data protection is now a bigger challenge for organisations – it requires more investment in modernisation, and nurturing a culture of maintenance.

It is now expected that digital services are accessible – designed inclusively to make it simple and easy to use by all. Inclusivity affects everyone by including permanent, temporary and circumstantial needs (e.g. deafness, ear infections and noisy environments). Some countries are beginning to legislate in this area, adding new legal responsibilities. Yet, the ways in which existing services exclude are often trivial to identify. Away from mass consumer markets, niche software is often inaccessible for many — staff and specialist users must work-around the flaws. Low quality niche software can be the daily working experience for administrative staff in large organisations, made worse when the design excludes them.

High-growth tech companies provide many of the services that consumers experience daily. When consumers order a taxi, a takeaway or buy a book online, they use technology which has recently been renewed or replaced. High-growth gives the abundant resources to make this possible. Established organisations rarely experience high growth so accumulating technology becomes a maintenance burden. With limited investment in technology, organisations prioritise high profile services. Established airlines provide a good example of this. Buying a flight online feels like a modern internet-era service. A less-used service like changing your flight can be very challenging.  Lowest priority of all, office administrators will often use ageing, low quality ‘back end’ systems.

If software is eating the institutions which form the structure of our societies, it must not cause them to fail.

Unsustainable technology inhibits the agility and stability of organisations. It will become a threat to their existence. Businesses will not be able to compete with the agility of younger, leaner organisations. The role of institutions will erode through lack of trust, with citizens opting for market alternatives. The importance of sustainability goes beyond the impact on organisations. If software is eating the institutions which form the structure of our societies, it must not cause them to fail. We must find ways to make digital technology sustainable over decades if these institutions and public trust in them is to endure. 

Network and data centre energy consumption is already set to increase as a proportion of global energy consumption. If digital technology is not made sustainable, inefficiency will result in avoidable accumulating energy use. Sustainable digital technology is necessary to avoid the internet revolution being a key contributor to climate change.

Digital technology will not stop eating the world – the promise of automation is too great, and technology can have a positive transformative effects on people’s lives. If it cannot, and should not stop, it needs to become sustainable.

What is toxic technology?

Toxic technology describes the harmful characteristics caused by poor design, or neglect. Poor design is common, in an industry where outputs are often favoured over outcomes. Neglect is systemic, caused by short-termist cultures, processes and practices which inhibit sustainability. 

Whilst the impacts of toxic technology are significant, examples of toxic technology are mundane, everyday, and recognisable to most. It is: the broken kiosk at the local museum, the ageing computer-on-wheels trolleyed around the hospital, the unpatched web server that lead to the embarrassing data breach or the strange green-on-black interface from the 1990s used by the back office staff at a big bank. Toxic technology is around us all, powering our banks, care homes, warehouses and submarines. It’s pervasive.

The following are typical toxic characteristics in technology. Each is challenging, and subjective to measure – making toxicity hard to expose.

  • Insecure – unacceptable risks to breaches of confidentiality, loss of integrity or lack of availability
  • Unscalable – an inability to respond to change of scale, such as increased usage, number of users, or complexity of the domain
  • Unreliable – lacking durability, availability and predictability
  • Non-compliant – non-compliance with the law, standards or an organisation’s policies
  • Inaccessible – the design excludes users
  • Hard to support – cannot be maintained effectively and efficiently 
  • Hard to change – cannot be changed effectively and efficiently
  • Opaque – important information about the service cannot be obtained when needed
  • Overly expensive – the service isn’t value-for-money
  • Poorly understood – the service and its technology is poorly understood

Software in particular can move fast to toxicity, more so than physical technologies. Bridges can fail and buildings can decay but the patterns of neglect are reasonably well understood, and occur over decades. Software decay is faster, less predictable and subject to more complex external factors. Cyber security vulnerabilities can emerge in any component part. Open-source communities may become unreliable. Commercial suppliers may go out of business, or stop working in your interests. Even doing the basics like patches and upgrades is challenging due to the norms of culture, practice and process.  The software industry is not yet mature enough to match the risk-management rigour of civil engineering.

The term ‘toxic’ is intentionally evocative language to give a sense of active harm, worthy of attention. Terms like ‘legacy’, ‘technical risk’ and ‘technical debt’ are useful, but don’t give a sense of urgency.  For most organisations, toxic technology is a growing and ignored problem, so a change of language could help. 

Systemic issues are the principle cause of toxic technology, not individuals or teams. This is important to recognise when using the very negative term ‘toxic’. The assumption should be that historic creators and decision makers made decisions in good faith. Ageing technology accrues toxic characteristics which become more visible from a contemporary perspective. Historic code reveals the culture, language and decision making of the time. It should be valued as a form of communication from the past to the present – perhaps even aesthetically appreciated like historic buildings. Toxicity is avoided through understanding that it can emerge over time from even the most thoughtfully designed technology.

Continue to part 2…


Nick Rowlands (@rowlando) for the idea to publish as a series of blogposts, reviews, and general encouragement to write more.

Steve Marshall (@SteveMarshall) and James Stewart (@jystewart) for their many second opinions on my writing.

Giles Turnbull (@gilest) for timely advice to improve my writing.

Sign up for future instalments

Sign up for email updates below.

Success! You're on the list.


  1. Great stuff Dave. You have identified a lot of issues that are common to software some of which, I would suggest, are almost essential to an MVP product. It seems to be an aim to start with something insecure, inaccessible etc and pinky promise to fix all that later. Which is fine, if it happens 😁 Look forward to the next posts.

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s