When the first cut is the hardest, however shallow

Remember when I wrote about mainframes on Mars?

A lot of hard decisions fail to get made because of politics, habits and very basic human fears

Imagine just how smug I felt when, sitting at home one day not so long ago watching ‘Expanse’ on TV… set in a distant, intergalactic future… two space shuttle pilots are having to deal with an issue while fleeing the bad guys, and one of them says: “The ship’s mainframe must be down.”

HA! They were literally zooming back to Mars and there was the mainframe.

I CALLED IT.

I was well pleased with myself.

I will (grudgingly) admit that the presence of mainframes in a space-worthy ship in the year 3,000 probably has more to do with the technological illiteracy of the screenwriters on this occasion than my prophetic abilities… but you cannot take this away from me.

It is ironic.

And it is apt.

Because mainframes are still… everywhere.

Often lurking behind shiny, new things.

A few years back, a bank I worked at decided to do an AI pilot in partnership with a tech firm. It was in every sense of the term… ‘jazz hands’ innovation. But it was still a first foray into being data-curious so we could become data-driven.

It was a start, however timid.

And as first steps always are… it was hard to take.

When we finally picked the use case… after all the political battles of first nobody wanting to go first and then everyone wanting to go first… we finally picked a division as the area that would do the honours of ushering us into the future. It was a political decision, before you ask.

We selected the problem statements and data sets.

Worked with risk and compliance and infosec.

Went back a bunch of times to revisit assumptions and make sure everyone was comfortable.

And finally… several weeks on… we had everything approved, signed, sealed and ready to go. Which is when the actual work could start.

Then we sat down, extracted the data from the mainframe… onto a spreadsheet.

Stop giggling at the back.

Data. Spreadsheet. Then we anonymised it by hand. BY HAND.

Then we burned it onto a CD and I personally handed it over to the representative of the tech firm we were working with.

It felt like it should have ‘Summer Hits 1999’ written in stencil on the cover. But no. It was a bunch of anonymised customer data sets and a lot of curiosity.

A week later, they emailed us the AI report.

And it didn’t tell us anything we didn’t know. But that is a story for another time.

This was ancient history, you may say.

This was only a few years ago, I would answer.

A few years is a long time in tech, you may say.

And you wouldn’t be wrong.

But.

A couple of months ago at a conference, a principle from a major PE house mentioned that, during due diligence, a lot of AI companies were revealed to have small armies of offshore teams… not developing but rather ‘supporting the AI’… ahem… and large organisations admitted on the very same stage that they had had to abandon over 90% of their own proposed AI use cases because the data (that, for the avoidance of doubt, existed in their own systems) was not accessible for analysis. Because… mainframes.

So you could argue that mainframes hold back 90% of your ambition… and arguably nobody would care about the remaining 10% anyway if, say, their credit cards are not working, because the basics always matter more and the basics are, for now, still married to the mainframes.

And this is where you tell me that mainframes are resilient. They don’t fall over. They don’t die. So the credit cards are largely safe.

And you would be largely right.

Until you are not.

But even if that was enough… they also don’t evolve.

And that leaves us in a strange position.

Observe:

In some upcoming Mechanical Orchard research that will be released in the next few weeks (so watch this space because I am bound to be talking more about it)… 74% of respondents said they see legacy modernisation as a way of embracing innovation and 84% as a way of acting on a visionary mindset. But a quarter also admitted that they won’t fix what ain’t broke… and they will always look over their shoulder for any decision that is too visionary by half and may end up being career limiting.

So if the reasons for doing are aspirational… and the reasons for not doing are existential… that is exactly how we end up with mainframes on Mars.

I am not belittling the innovation of the last few decades.

We have done amazing things in our societies and in our industry. But we have done those things in a largely additive way. New technologies living side by side at best or precariously balanced on top of the old stuff.

Is that sensible?

No.

Does it create resilience and stability issues? Yes.

Will it be a nightmare now that DORA is coming into our lives?

Why… yes.

So why are we doing it to ourselves?

The Mechanical Orchard research indicated something very interesting: 100% of respondents said that any tech modernisation they have undertaken in their careers so far yielded results across many verticals. Stability and performance improvements, productivity gains, cost reductions. All of the respondents also confirmed that they got results in more than one vertical even if they were originally motivated by just the one.

Yet, and unsurprisingly perhaps, 71% also admitted that the idea of engaging in such projects (even with the knowledge that they bring results) is a source of anxiety.

Given the choice, decision makers would rather leave well alone.

Once the decision is made, the respondents showed confidence and the ability to execute and navigate the multitude of decisions that are made after the decision to engage is made. And they do deliver results.

But the early part of decision-making? From a standing start to ‘let’s go’?

There be dragons.

Delegation feels inappropriate and influence flimsy.

The political capital required feels daunting and the budget often proves elusive.

Given everything we know from years of doing this work, and the research the Mechanical Orchard team carried out, it really feels that the early part of the decision-making is a different beast altogether to the rest of the work – and that is hard enough as it is.

It feels like the curve from ‘leave it for the next guy’ to ‘let’s deal with this once and for all’ is an insurmountable obstacle for many and an exercise in bravery for the few.

But shouldn’t it be an exercise in risk management instead?

What would you do if the way to modernisation was packaged in a way that project deadlines are genuinely a line item in a Gantt chart and the unit cost of each project phase no larger than a rounding error? I exaggerate, of course. But the question remains: what would you do if taking action did not need to involve bravery?

Not because you have magic dust to sprinkle all over the problem, but because you know how to sequence the work and chop it up in the smallest and most risk-managed constituent parts.

What would you do then?

If the decision to change wasn’t existential, but procedural?

I have no doubt that the world is moving relentlessly forward.

And I have no doubt that hard decisions are being made every day.

And yet I also know a lot of hard decisions fail to get made because of politics, habits and very basic human fears. That is the real battleground. And we need to fight.

And we need to win.

Or we won’t even get to start with the real work at hand.

#LedaWrites

Leda Glyptis is FinTech Futures’ resident thought provocateur – she leads, writes on, lives and breathes transformation and digital disruption.

She is a recovering banker, lapsed academic and long-term resident of the banking ecosystem. 

Leda is also a published author – her first book, Bankers Like Us: Dispatches from an Industry in Transition, is available to order here.

All opinions are her own. You can’t have them – but you are welcome to debate and comment!

Follow Leda on X @LedaGlyptis and LinkedIn.