Blog 1: Post-mortem of a cloud-only approach
For those organisations that have decided that cloud adoption is the path to successful digital transformation, the terms cloud-first and cloud-only are familiar ones. However, while bearing many similarities, the two cloud strategies are not interchangeable, and the outcomes can differ considerably.
In this blog, we first briefly examine the difference between the two approaches.
Then we deep dive into the risks of a cloud-only approach, using our reference organisation, Organisation X (OX), as an example. We’ll join OX on their journey and examine the outcomes of their cloud-only decision, and the roadblocks they experienced on the way.
The anatomy of success: Choosing a cloud-first vs cloud-only strategy?
Organisations that opt to embrace a cloud-first strategy, make a conscious decision to consider the cloud before all other technologies. So, when developing new processes or adapting old ones, they always consider cloud solutions before non-cloud options.
As a result, over time, most or all their infrastructure is moved onto a cloud-computing platform. The important words here are ‘consider’ (rather than mandate) and ‘most’ (rather than all). Cloud-first is not an end-all or at-all-costs strategy. By comparison, a cloud-only strategy is a top-down directive that compels an organisation to implement cloud solutions only – to all services and systems – and frequently at scale. Regardless of perfectly viable alternatives or more cautious approaches.
Why does approach matter so much?
According to McKinsey & Company, digital transformation has become harder, not easier.
only 30% of organisations that have undertaken a digital transformation achieved success
While reporting in previous years that only 30% of organisations that have undertaken a digital transformation achieved success, more recent survey results suggest that the odds of realising success are now even greater.
Even just four years ago, McKinsey & Company reported that only 16% of respondents agreed that their organisations’ digital transformations successfully improved performance and also equipped them to sustain changes in the long term. Meanwhile, an additional 7% said that while their performance improved, those improvements were not sustainable.
While McKinsey & Company also concluded that those organisations that reported more successful transformations had also deployed more technologies than the rest. But along with this, they cited 21 critical keys to success, in categories ranging from digital-savvy leadership, empowering people to work in new ways, to communicating change using both digital and traditional channels.
With these key factors in place, say McKinsey & Company, respondents are up to three times more likely to report a successful digital transformation.
And you’ll see over the course of Organisation X’s journey, why these points as well as ‘technical’ judgement calls mattered.
In the beginning
In its ‘The State of Cloud-Driven Transformation’ report, Harvard Business Review Analytic Services surveyed 260 respondents familiar with their organisations’ cloud adoption. As was expected, 83% said that cloud is very or extremely important to the future strategy and growth of their organisation.
Obviously, when done well, a successful transformation creates value in an organisation.
But caution is urged by Rodney Zemmel (global leader of McKinsey Digital) in McKinsey & Company’s ‘Digital transformation on the CEO agenda’. And we quote: “There’s often a temptation for companies to say, “Let’s make the technology investment, and then we’ll figure out all the great things we’ll be able to do in the business.” If you’re not working backward from a business case, then you’re unlikely to get the value from it that you’re intending to get.”
And perhaps OX should have taken more notice of this warning.
However, the compelling business logic of moving lock, stock, and barrel to the cloud to achieve greater agility, support for collaboration, reduced costs, and increased flexibility, insight and innovation faced little if any internal opposition when mooted at OX executive and board level. And the proposal even garnered a round of applause.
In its enthusiasm to accelerate the realisation of those benefits, especially the financial ones, OX opted for a cloud-only strategy. The board signed off on a project which would rapidly migrate their entire technology infrastructure using a “lift and shift” approach to a consumption-based model.
Those first few steps and quick wins
For OX, the initial wins were fast to roll in the door. Replacing some of the organisation’s legacy on-premises applications with cloud-based ones won fast favour with users empowered by anywhere, anytime, any device accessibility. And a range of departments rushed – in isolation from one another – to find suitable alternatives to supplant their old applications and processes with shiny new cloud versions.
But despite the initial enthusiasm and quick successes, OX’s cloud-only project started to drag well beyond the original delivery timeline – and then grind to a shuddering halt. But why?
The workload consideration (and why it boils down to money)
Organisation X’s initial decision to lift and shift ‘everything workload’ proved to be ill-considered.
To quote McKinsey Digital’s article on ‘Cloud economics and the six most damaging mistakes to avoid’: “The elasticity and scalability of cloud is economically ideal for workloads with variable cloud-consumption patterns.”
However, organisations like OX, which fail to examine and differentiate between workloads that deliver economic benefits from on-demand scaling and those that don’t, typically see their costs increase. In not aligning their compute needs to the demand patterns of the organisation (users and customers), OX was unable to predict its cloud consumption costs with any degree of accuracy, or to subsequently optimise those costs.
50% of organisations are already calling out optimising their cloud costs to improve ROI as a transformational goal for 2022
In its determination to be cloud-only rather than cloud-first and migrate every workload (regardless of scale or type) to the cloud – while ignoring viable alternatives like hybrid on-premises, private cloud and hyperscale solutions – OX missed a trick. A big one. And at a time when 50% of organisations are already calling out optimising their cloud costs to improve ROI as a transformational goal for 2022, it was a backward financial step.
The people-first equation
Organisation X’s failure to invest sufficiently in either staff training, user experience, or organisational change management (OCM) played its part in the project coming to a stop.
While the OX stakeholder enthusiasm for ‘everything cloud’ was significant, the race for change often outpaced the organisation’s ability or level of preparation to take its employees along with it. With poor user adoption already playing a critical role in the failure of technology implementations across the board, OX’s focus on cloud solutions as an outcome rather than an impressive employee experience, saw its people become disheartened, exhausted, and disengaged.
And while automation was touted as the cure-all for repetitive manual processes, designing and implementing these workflows was often a painful and time-consuming process for users expecting a slick consumer app experience.
Despite a well-publicised global shortage of IT skilled resources, Organisation X, tried optimistically but unsuccessfully to recruit an internal cloud expert to manage their cloud environment through and after its adoption. And any resources they did initially attract, demanded salaries that wiped out potential staff savings they’d make elsewhere. All of which added to a difficult situation.
To quote Gartner’s report, 4 Trends Impacting Cloud Adoption in 2020: “Through 2022, insufficient cloud IaaS skills will delay half of enterprise IT organisations’ migration to the cloud by two years or more.” And despite all efforts, say Deloitte in their 2021 ACS Australia’s Digital Pulse, it’s not going to get any easier.
With just 7000 students graduating with an IT degree (in 2019), and an estimated need for 60,000 extra workers per annum over the next five years, the challenge to close the ICT resource gap seems insurmountable.
Unexpected cloud computing costs, a nail in the coffin?
What often seems like small decisions at the beginning of a cloud adoption project, can, as OX subsequently found out, dramatically impact expected monthly costs. According to Spot by NetApp, there are three main components offered by cloud providers that determine the cost of cloud computing services.
There’s compute which is generally offered in a range of instance types, each with a specific amount of CPU resources, memory, and sometimes specialised hardware (like fast networking or graphics acceleration). Customers pay based on the number and types of instances and for how long they’re used.
Then there’s networking, which is billed by the volume of data transferred into the cloud service (ingress), out of the cloud service (egress), or both.
And lastly, there’s storage as a service, where customers can opt for elastic storage services and pay by GB-month of storage used, or managed storage services, where they pay for an entire storage volume, regardless of how much they utilise.
But it doesn’t stop there.
As well as the cloud migration and adoption costs that are usually calculated at the beginning of a project, says NetApp, there are a range of direct and indirect expenses. These include the relatively simple to calculate direct costs like hardware, software, management, maintenance, and staff, as well as any physical facility. And indirect costs like loss of productivity (for example, server downtime), loss of customer trust, support costs and timely responses from vendor, and reputational damage – which are considerably harder to predict.
On top of this are the decisions to be made on cost models, which include pay-as-you-go, fixed-term subscriptions, reserved instances, savings plans and spot instances.
For OX, poorly informed decision-making and a lack of financial rigour soon escalated their cloud costs beyond control. And on top of that, they completely forgot to allow for items such as data backups and third-party copies.
Post-mortem findings
If OX was to have a board-level EOY dissection of their decisions, we believe these would be their findings:
- Question a cloud-only approach when aspiring to a successful digital transformation. Cloud at all costs is not necessarily an affordable, effective or sustainable decision.
- Don’t put all workloads in the cloud just because you can. Make discerning, well-informed decisions about existing workloads and the best way to manage them.
- To achieve value, work backwards from your business case. Put outcomes and people before technology. And make sure that employees are on the same journey as the organisation.
- Invest time in understanding the underlying fixed and variable costs of the cloud models on offer. With 50% of organisations already focusing on optimising their cloud costs to improve ROI as a transformational goal this year alone, you won’t be alone.
- Factor in resource availability and costs. Or better still, engage a proven cloud services partner (who can help you optimise costs and make informed decisions based on your individual requirements instead of a ‘standard’ package).
