Below we note several trends we believe will become more central to the technology investing framework in 2020.
Interestingly, we classed some of these trends as more long-rage in nature just two years ago. That we are now expecting them to see the light of day in 2020 is testament to how quickly technology adoption can inflect.
#1 Technology becomes politicised
It is clear the unregulated ‘wild west’ days of tech are coming to an end, with more and more big tech companies now trying to balance time-to-market with accountability. The socioeconomic risks brought about by the threats of AI and Big Data to employment and privacy, as well as cybersecurity becoming a new battleground, leave the door open to regulatory creep. This is a trend the technology industry has avoided for years. However, stricter regulations are inevitable, given other productivity-enhancing industries (from transport to pharma) have already faced this reality. We have witnessed a tightening of privacy laws under GDPR but there are many other technology domains, including AI (eg regulation around ‘explainable AI’) and social-media (eg controlling childhood addiction), which might need further regulatory framework. The tipping point for such a regulatory cycle tends to be the centrality of the topic for political discourse. As 2020 is a key election year, it is likely to represent the tipping point that politicises technology.
#2 FOMO drives widespread experimentation
Many of the hyped tech domains, from IoT (eg home automation, industrial sensors) to AI (eg intelligent process automation, image and language processing) are already being commercialised. This has led to many companies embarking on ‘proof of concept’ projects to on-board such technologies. The net effect, despite many failed projects, is for companies to put teams and workflows in place to experiment with and adopt new technologies. This is a major change from the past when many waited to see the net impact on early adopters before industry-wide experimentation. A FOMO (fear of missing out) attitude has now changed this. As experimentation puts strains on available skills within IT departments, it forces better collaboration between IT and line teams. In essence, this means vendors in these hyped domains experience quicker growth curves. Nothing speeds up adoption (aka people signing cheques for IT) faster than when there isa line-team buy-in alongside FOMO from a competitive perspective. Clearly this is not applicable to all hyped domains, with some specialised areas (eg VR/AR, Drones, etc), enabling technologies (eg blockchain), and extremely early stage tech (eg quantum computing) seeing limited adoption benefit from this trend.
#3 Digital transformation projects will accelerate IoT adoption
Digital transformation (DX) has been a trend for a number of years, which has generated significant services revenue. We expect it to last a while (see p10 of our Softcat Initiation). In essence, digitisation takes businesses’ existing workflows and supercharges them with technology to enable them to become more data driven and productive (eg via automation/AI). The resulting upgrades to networking structures, storage options, data policies, and security decisions makes it easier for these DXed companies to deploy more points of data gathering and actioning. This makes the greatest impact when it enables easier deployment of the Internet of Things (IoT). Projects such as better fleet monitoring for logistics companies (eg predicting when a delivery is likely to fail due to amechanical issue with the delivery-truck) to linking a shops’ shelf availability and foot traffic to the automated just-in-time warehouse are examples of incremental IoT projects that are easier to deploy once a company has already undergone significant DX. In the former examples, real-time delivery notifications and warehouse automation were the DX projects enabling the deployment of the IoT (sensors for shelves, foot traffic and engine management).
#4 Demand for explainable AI will gather momentum
Much of the modern AI revolution centres on pattern recognition. Some of these pattern-based classification techniques, mostly statistical, have been used in some industries for decades. Therefore, people have already experienced frustrations concerning the biases in such classifications systems (eg unreasonable increases in insurance premiums, invalid mortgage rejections) and the inability of customer-facing staff to explain the underlying actuarial models. But at least many of these models are statistically explainable. New AI technologies such as Deep Neural Networks (see page p7 of Generation 3.0) are not explainable by design. This is made worse by the fact these models are trained on data which can introduce significant biases (eg an African American may have higher rejection rates for a loan vs a White American despite all other attributes, education, age, income, location etc being equal). With an increasing number of decisions being made by such hard-to-explain algorithms, it is becoming difficult to contest them in the courts, and worse, they are starting to have unintended consequences. GDPR already factors in the need for safeguards (eg human-in-the-loop and a non-binding right to an explanation).
New York City has already passed the first algorithmic accountability bill. We believe this is just the tip of the iceberg.
#5 Becoming a ‘tech company’ is not always a panacea
With technology becoming ubiquitous, companies of all sizes in all industry verticals have been able to transform themselves. But some industry verticals, such as transportation, healthcare and real-estate have sticky unit economics and regulatory frameworks. This makes it difficult to scale new models, and therefore materially alter intrinsic profit margins. An example of this is the struggles of local takeaway platform champions to stay away from delivery models and to replicate the model in other countries. Another example comes from the social media space where new unforeseen costs (eg compliance staff) are dampening the margin expansion. This also applies to large incumbent companies that are more averse to breaking the ‘rules of engagement’ in an existing vertical. The rise and fall of the ride-hailing industry is an example of this. At the other end, smaller companies are susceptible to buzzwords (eg omni-channel communication, gamification). They then deploy capital to on-board these new technologies without understanding how existing workflows would be affected. Dramatic changes for smaller companies can change results in the under utilisation of the new whizzbang thing at best, and at worse, could negatively impact the culture.
#6 ESG becomes a bigger theme
From adopting software to better monitor diversity and pay gaps, to on-boarding technology that lengthens the useful life of existing technology, ESG has moved from being a tick-box exercise to a theme that can unlock sales conversion for technology companies. It is fast becoming a ‘no one gets fired for buying IBM’ sort of theme, and we are likely to find more technology companies repositioning (using marketing) their offers as ESG enablers.
#7 hyped structural trends gather real momentum
1. New infrastructure The ‘new’ infrastructure:
• Cloud computing allows greater flexibility in the use (and payment for) of IT.
• Edge computing will address the negative aspects of centralisation that cloud computing will bring about by moving some of the ‘decisioning’ power to devices that sit closest to the data (smartphones, cars, smart meters, etc).
• 5G unlocks the speed and bandwidth necessary to deal with the enormous amount of data that will be gathered, generated, and used at the edges.
2. Ambient computing
Since being commonplace, people have either ‘operated’ computers (eg back in the era of mainframes) or ‘used them’ (eg what we do with smartphones). However, we believe computers will become an ‘invisible’ part of our life in the future, augmenting our capabilities without the need for conscious engagement. This is unlocked by innovations in:
• The internet of things. These connected computational devices are typically limited to a very niche function. Smart meters, LiDAR systems in cars, connected heart-rate sensors are some examples. These devices tend to have communication capabilities (eg WiFi, 5G), energy efficiency (eg smart wearables that last weeks), sensory capabilities (eg 3D environmental sensors), and fully-fledged computation capabilities powered by Edge computing.
• Artificial Intelligence. While people seem to have talked about AI forever, some recent innovations have unlocked a revolution. For example, see page 5 of our Generation 3.0 note where we talk about the impact of machine learning on things such as face recognition and autonomous driving. Not only has AI enabled better interfaces (eg voice assistants like Alexa), it has handed over some decision making to computers. From as complicated as a self-driving car to as simple as a smart thermostat, the impact is being felt more.
3. Distributed computing coming of age
Solving distributed computing is what has brought about recent computational leaps, including the internet. However, it took cryptocurrencies to bring distributed computer science to the attention of investors. The one that won the buzzword race is the distributed ledger technology, blockchain. Underneath the hood, it is just another distributed consensus mechanism, like Paxos and Raft.
The initial excitement centred on disintermediation in the most centralised of sectors, finance. We believe the concept of decentralisation of trustis likely to find disruption in areas where ‘trust’ is not something well thought through (See page 23 of our Generation 3.0 note). That said, it still feels like we are on the hype part of the hype-curve, with start-ups focusing on transaction speeds and consensus algorithms to come up with a solution in search of a problem. Nevertheless, this is exactly how major revolutions begin
4. Modularisation of technology
Turning technology solutions into Lego-like building blocks, which can be assembled as you wish, has been a trend for decades. However, having had many false starts where proprietary modularisation standards vied for dominance, we are finally at an age where modularisation can be achieved using open standards. We are likely to hear more and more technology companies mention words like Microservices and Application Programming Interfaces (APIs).Why does this modularisation matter? ERPs are the biggest tech solutions out there. It is well known that many new implementations fail completely (50% according to technologyevaluation.com) and, of those that work, many fail to deliver measurable ROI (90% according to erpfocus.com). This is because requirements are miscommunicated, or changed during implementations. The more modularised the system, the easier it is to address such issues around requirements. Modularisation also allows technology companies to better co-exist with competition and increase the chances of a land-and-expand strategy.
5. Pendulum will swing back to Vertical Industry Specialisation
For some time now, more generic tech giants have been dislodging vertically-specialised technology providers (eg Google’s entry into areas of healthcare and Microsoft’s progress in retail). Specialist vertical offers are also being consolidated into broader offers (eg Siemens’ software buying spree). Clearly, these trends allow verticals that have seen little technology innovation to embrace innovation from other verticals, thereby disrupt an existing status quo.After a period of euphoria, customers will inevitably want the technology vendors to speak their language, and understand their DNA. In a world where consumers demand ‘hyper-personalisation’, without understanding the ins and outs of an industry, it becomes difficult for technology vendors to provide ROI beyond the buzzwords. An example of this evolution is the transition of Salesforce from a standard CRM to a platform on which others are building vertical-specific solutions.