In the last fortnight, I had two interactions with customer care teams. The first was Kotak, for a bank account. There were failures at multiple levels, but the biggest takeaway was how difficult it was to get through to a human. Chatbots have their uses but this was extreme! Twitter DMs weren’t a help either. The second was Amazon, and that surprisingly included being put on hold for about 15 minutes and transferred 5 times! Thankfully, the problem was resolved in 30. But given the famed Amazon customer-centricity, this was a disappointment, and I wondered if at scale, it was inevitable.
What causes it? The best frame I have seen is entropy. (via) Yes, that physics and thermodynamics thing – the degree of disorder/randomness in the system. Though the meaning remains the same, the application changes. In the context of a business, it is the tax applied by the system between input and output. Just imagine the effort that is required to get things done in a corporate structure as against a startup. While it is theoretically possible for a complex system to have lesser entropy than a simple system, I have not seen it in practice. I can imagine why it is so – the randomness that can be generated by different touchpoints, and their optimization for following rules as opposed to providing solutions. What happened to me with Kotak and Amazon were examples of that.
What can be done about it? There is some inevitability about systems tending towards maximum entropy, can it be slowed? Technology has provided a bunch of means by which the system’s tax can be reduced if not nullified. “Frictionless design” is a mantra and several businesses have made billions of it. Finding the best restaurant nearby is a search away, reaching a loved one is a message away, and buying pretty much anything is just a click away – including uranium ore! It has also enabled friction reduction in processes and communication within organizations. But just as they have created problems as companies optimized for their own growth, efforts to minimise friction within organisations can also lead to misuse and instant fixes that might cause more problems in the long term.
What are possible checks and balances? The friction in the system is most likely a regressed form of checks and balances from earlier timeframes. But these were also probably made by smart minds and worked well at a time. How does an organisation ensure systems evolve well?
a. Adapting: The definition of friction to a consumer, internal or external, is constantly changing. A regular evaluation of the system, the external environment and expectations, and applying changes is one possible approach.
b. Planning: Changes are important, but so is the framework for deciding and making those. Secondary effects, deciding goals, evaluation timeframes and metrics are some of the components that easily come to mind.
c. Avoiding the God complex: Businesses operate in complex environments. Delivering simplicity should therefore be a prime objective. In this context, emergent solutions could work better than command and control models. If still unconvinced, find a middle ground.
d. Communicating “values”: Making sure that consumers and employees know the values of the organisation is an important first step. On this note, trying to please everyone might be sub-optimal. Setting expectations and meeting them might be a wiser strategy. What is the behaviour you want from the consumer, how do you facilitate it? Where does that behaviour lie in the larger framework of society and humanity?
e. Ensuring feedback loops: A thought just struck me – are aggregates of humans similar to algorithms? Even if not, setting feedback loops is important. The behaviour that gets encouraged is a signal. What signals are you sending as an organisation?