At the height of Tulipmania, a flower could cost the same as a house. Are we in a similar bubble with banks investing in certain types of technology?
I’m standing in front of the florist waiting to buy flowers for our wedding anniversary. The fact that I was a month early is another matter, which will haunt me for some time.
While standing there, I noticed a vast selection of tulip bunches on display – a wave of pinks, purples, whites, yellows, and reds. This sight reminded me of Tulipmania, a period of Dutch history when everyone went crazy for tulip bulbs. At that time, people considered tulips bulbs a better investment opportunity than gold. Tulipmania is one of the earliest recorded examples of an irrational speculative bubble. The prices bloomed (pun intended) to extraordinary levels until the inevitable collapse in 1637.
Bubbles or Hypes
There have also been other similar situations in recent memory: The rhodium bubble in 2008, where prices rose 1800% over two years; the dot-com boom of the ‘90s, which could well repeat itself with the now hundreds of $1bn-valued “unicorns” in Silicon Valley; and the growth of the cryptocurrency market over the last few years, to name a few. Some of these are easier to understand than others – but the most bizarre of them all must surely be the tulip bulbs.
My mind turned to banks and how they are investing in technologies such as blockchain as if they were tulip bulbs. I’ve seen countless of these investments pitched as credible solutions and then failing to deliver, yet we continue to throw money at the next one that comes around the corner.
On the topic of Blockchain. The main feature with Blockchain is that the transaction is immutable but did anyone ask the GDPR team about their view on this functionality? If you can not remove information about the individual the transaction what do you then do with the transaction? See my blog https://transanalytics.co.uk/blockchain-vs-gdpr-overlooked/
The abnormal has become the Norm
I remember a conversation in which a stakeholder was telling a group: “After a year of hard work, we’ve succeeded in running three of our risk calculations on Hadoop on a bit more than 2500 servers. How amazing is that?” Someone in the meeting room couldn’t resist asking how many remaining calculations that were needed to convert before the legacy risk system could be decommissioned. “close to 3000” was the reply.
In other words, it took 12 months to develop three calculations, and close to 3000 remained. Let’s look at the maths: Three months multiplied with 5 developers multiplied with 3000 calculations equals 900,000 man days/4,500 years Hmmm, was there ever a business case?
All technologies have what you call a sweet spot; this was clearly not it.
The Difference
This reminded me of a similar project with an in-memory database using in-database processing. The work to convert the complex calculations into SQL took a very short time. The cluster for the database consisted of a bit more than 10 nodes and the calculations ran with execution times in seconds. Although there were a smaller number of transactions, the technology and design provided linear performance i.e. the same solution on the bank’s transaction volume would have required around 60 nodes and still provided runtimes in seconds; 4,500 man-years vs months, what had the best business case and they bank evaluate all options or was the decision driven by a hype?
At the height of Tulipmania, some bulbs would reportedly sell for as much as a good-size house. In the end, it was just not feasible for prices to continue to rise in something that has no intrinsic value. I wonder if the same is true of banks investing in blockchain?
Do not Forget
This trend is a serious problem for us to discuss within the industry, but nothing on the scale of what I experienced when I returned home that evening with an anniversary card and a bunch of flowers; “just a month too early” my wife said with a smile. And, no, they weren’t tulips.