Cloud Migrations and the Sunk Cost Fallacy

Nick McDonald, Senior Account Director at Fujitsu, says a growing number of organisations are turning their backs on the cloud to repatriate their infrastructure despite having invested heavily in migrations

Have you heard of the sunk cost fallacy? It refers to the tendency to continue a losing or failed endeavour because of the resources – time, money, etc – that have already been invested into it.

Those who fall victim to it believe it would be a waste if they change tack, even if that would be a much better route moving forward.

The sunk cost fallacy always comes to mind when I speak with someone about cloud migrations.

In recent years, companies have been pushing ahead with these in a bid to cut infrastructure costs and unlock the flexibility it promises.

But these streamlined costs and claimed flexibility have not necessarily materialised to the extent expected, and now, a growing number of companies are dropping their plans for the cloud and returning to on-premise and local data centres.

This shift is being dubbed cloud repatriation and sees companies move data, applications and workloads from a public cloud environment back to on-premise or a local data centre.

This trend was predicted by Morgan Stanley in a 2018 financial report that anticipated substantial revenue growth from hardware despite companies embracing the cloud at scale.

The report said that while hardware sales had slowed, it was because companies were putting purchases on hold while they took time to better understand the cloud and migration process, or to make the jump.

Today, companies are now reconsidering these migrations, which can take years to complete, while those already transitioned to the cloud are looking to return to on-premise or a local data centre.

So, what’s been the issue with the cloud?

Cost is the most prevalent. The companies hit hardest in this regard moved to the cloud with a lift and shift approach, which means existing applications were migrated without any modifications.

This makes integration easy but in the long-term, it prevents the company from being able to maximise the flexibility and scalability of cloud infrastructure.

Why? Because these applications are not optimised for the cloud environment so consume more resources. Given the pay-by-usage model of cloud solutions, this makes it expensive.

Other issues have emerged, such as limited infrastructure customisation. Compliance has also become a headache, especially when it comes to data sovereignty and residency.

In this industry as in others, businesses must have complete control over their data, and this can be hard to achieve when data is hosted in a public cloud.

The final issue, and looping back to cost, is that online gambling companies generate vast amounts of data and often find they need to keep taking on more capacity which increases costs.

Anyone with iCloud storage will know exactly what I mean.

So how does cloud repatriation overcome these challenges?

First, it can significantly reduce costs. An initial hardware investment is required, of course, and so is an investment in maintenance and updates/upgrades. But these costs can be a lot less than hosting data and applications in the cloud.

Performance can also be boosted because only the company’s data and applications are being hosted on the hardware, unlike public clouds which can support hundreds of organisations.

Compliance is an absolute must for online gambling businesses, especially those in regulated markets. Repatriation gives companies total control over their data storage and security, and they can ensure infrastructure is fully compliant with local regulations in each market they target.

Other benefits include better disaster recovery and business continuity and, ultimately, greater control and flexibility.

Of course, there are some things to think about before opting to repatriate from the cloud. This is a fairly complex undertaking that requires careful planning.

You’ll also find that most cloud platforms have not been designed to integrate easily with other solutions, especially on-premise data centres.

It also adds further complexity to the IT infrastructure, which, for international companies operating in this industry, is likely to be pretty complex already.

Then there are the higher operational and maintenance costs (including things like the cost of electricity and hiring specialist engineers) and the need to install and adapt to new infrastructure.

But these can be overcome with the right approach and so long as the company does not get sucked into the sunk cost fallacy.

How? It really comes down to planning which must include a full evaluation of the company’s needs and resources, both current and future, plus the applications and data sets that would benefit from being hosted in-house.

It’s then important to map out the hardware of software required, and the investment needed to build and run the data centre on-premise.

The actual repatriation process must also be mapped out with accurate timelines set so that business continuity isn’t compromised during what will always be a disruptive process.

By doing this, organisations can bring their data and applications back down to earth while ensuring they don’t fall into the trap of the sunk cost fallacy.