The tree and the iceberg offer some of the most profound lessons for both personal and professional life. The roots of the tree bear its weight allowing it to grow to greater heights; the deeper the roots the more resilient the tree is to disruptive forces. Personal and professional life rooted in strong foundational values helps us withstand the test of times. The same applies to businesses and technology as well. On the other hand, the tip of the iceberg, what is only visible, hides the perils that lurk beneath the waters. Narrowly focusing on only a few things that are often convenient and ignoring or simply staying ignorant of what is more important, not doing the right things, not having the right people for the right roles, are all anti-patterns we see everyday. As businesses try to get onto the Digital bandwagon to survive, thrive, and reimagine/rediscover themselves, we see these metaphors often play out in digital transformations in many ways. A few (random) thoughts follow.
Deployment Architecture… ignored
More and more businesses across industries are experimenting with AI and Machine Learning use cases, which is great as this invariably brings in fresh thinking and innovation into the traditional way of doing business. That said, while the focus seems to be very much on building and training the models, we don’t see the firms laying a strong foundation in terms of the technical architecture and deployment architecture. As a result, ML model training cycle runs days at a stretch, demands much more than originally desired level of computing resources (mainly the number of CPU’s, GPU’s, RAM size) thus driving up cost of production setup, and teams struggle to achieve acceptable levels of data processing throughput and response times. As use cases increasingly tend be online and real-time rather than offline or batch, this poses a serious challenge to the realization of the returns at a reasonable investment. Another common problem we come across in digital transformation is in architecting omni-channel core transaction processing systems. The weakest link can put the whole organizational reputation at risk when incorrect architectural assumptions are made on which channels are not likely to get high volumes and inadequate stress testing is done to validate architectures end-to-end.
Is there such a thing as too much encapsulation?
The concept of encapsulation is a beautiful one – put in layman terms, hide complexity while giving easy access to the value something offers. I am all for this. API’s give access to pre-packaged functionality, so you don’t have to build everything. All the popular Cloud providers give you a plethora of powerful features (such as speech synthesis and speech recognition) that you can access without having to know what is under the hood. So, is there such a thing as too much encapsulation? Unfortunately yes. We have seen over-enthusiasm in using third-party API’s, open source frameworks, no-code and low-code platforms without having a sound architectural and design foundation, and without validating the impact of such software on the non-functional behavior of the system (performance, scalability, resilience, etc.). The more such usage is in critical parts of the system higher is the risk that needs to be addressed upfront rather than put off for later.
Perils of Fragmentation
Take any new digital technology – today there are many flavors and options available for businesses to pick from. While this choice is great and it helps avoid vendor lock-in, without a proper governance to ensure alignment with what the business needs, the scenario could easily become uncontrolled chaos. Even with just a few standard web technologies in use we have seen IT teams struggle with integrating the UI and the backend, such integrations fraught with bugs and delays. Agile methodologies have many benefits not the least allowing development teams to get more functionality to production more often. However, an overly complex system that does not have clearly defined integration architecture will quickly go from bad to worse when you have multiple moving parts each in their own technology and proprietary frameworks – negating the benefits of Agile. Higher fragmentation also means potentially higher number of hot spots within the system where security could be compromised.
It is always a question of mindset and culture
In any digital transformation, it is easy to get blindsided by the next shiny new technology and assume everyone will easily take it on within the organization. We have seen that underestimating the need to train the employees on digital mindset and culture, which is based on being data-driven, customer-centric, and process-oriented, is the biggest reason for failure of the transformation. It is highly imperative that the transformation takes an evolutionary approach, with many incremental initiatives jointly embraced by the business and IT staff.
Client relationships – acquisition versus retention
The focus on digital has only given further momentum to digital marketing, with businesses spending significant budgets on wooing new clients through digital channels. We have seen many firms allocate significant time, effort, and money on new client acquisition while giving only secondary importance to servicing existing clients and building stronger client relationships. A firm might feel good about saying they have bots engaging with customers to service their needs but many of them do a shoddy job of this. Measuring how many questions were asked of/responded to by bots isn’t very useful if you are not measuring what impact it had on client satisfaction. We can forgive a bot for not “knowing” enough “yet” to answer a customer’s question, but if it does not then connect the customer to a human service representative then you are going to have an unhappy customer. Have seen so many bots not doing this basic thing. That said, many firms don’t even collect and analyze client feedback. They are at the worst end of the spectrum and any digital initiatives that are solely aimed at trying to woo new clients are so shortsighted. It remains true that happy customers are your best brand ambassadors and do the selling for you.
Role of empathy and client involvement
You simply cannot digitally transform your business if you do not involve your clients in the process. Once again while it has become fashionable to throw around buzz words such as customer journey mapping, empathy maps, co-creation, and agile methodologies, we still see many firms not really involving clients in the transformation or not getting the right persona involved. In a B2B context, for example, we have seen cases where leadership level within the end client organization buy-in into a digital solution but the actual end-users are not involved in shaping the digital solution, resulting in a failed solution. Also, when you involve the right end user persona, you get an opportunity to test your assumptions on sensitive areas such as what kind of end user data you can or cannot collect without explicit user consent. With businesses enamored by what they can do with AI and ML, this creates an insatiable appetite to collect as much data as possible of customers and potential customers. Taking an empathetic approach on this is of paramount importance to not only stay on the right side of any upcoming regulations but also to consider every customer as a long term relationship and not a transaction.
A final word
With the democratization of advanced technologies that are now easily made available and accessible on the Cloud, now is the best time for businesses to innovate using all this power at its disposal. That said, as the saying goes, with great power comes great responsibility. Not focusing on foundational attributes some of which I have highlighted above, will for sure lead to encountering icebergs of many forms with deleterious impact to the business.
Ramki Sethuraman
