“Cloud first” isn’t an IT strategy
This post isn't as niche as it sounds
One of my standout memories from the height of the pandemic (aside from apocalyptic fires here in Australia, and the toilet paper shortages) was working with a multinational resources company on their Cloud strategy.
The thrust of the document was simple: cloud first, cloud everything.
This didn’t sit well with me, but my role in this instance wasn’t actually to facilitate strategy, but simply to write it up.
The trouble with a “cloud-first” IT strategy—or in fact an “anything first” corporate strategy of any kind—is that the strategy is attempting to guess what the right decision is for all future software choices, and it also assumes that the answer is the same in every case.
This is not strategy
This is a rule of thumb at best.
And strategy should be the solution to ill-fitted rules of thumb that lead to expensive decisions with low-value outcomes.
The point of strategy is not to pre-make all decisions. Instead, it is to define a goal, diagnose what is happening in your business and market that will affect that goal, set guiding principles and create a coherent action plan. (Yes, I am yet again talking about Richard Rumelt’s Strategy Kernel.)
In reality, “cloud-first” was an internal marketing campaign to drive more system owners to choose cloud under the assumption that cloud was the optimal long-term cost decision for the organisation, and some other nebulous benefits. Just quietly, the numbers did not support this decision.
Cloud computing may have absolutely been the right decision for may of the new applications the organisation was implementing, and even for some of the old ones. But it wasn’t the right decision for the entire network of systems they used.
The hallmarks of bad strategy.
What a strategy looks like
What would a strategy have looked like in this situation?
Define the goal.
Perhaps the goal was to reducing computing costs. Perhaps it was to gain more visibility over the full system landscape. Perhaps it was to improve security.
“Going cloud” isn’t a good goal because it does not define value to the business. Cloud may provide benefits but it is not itself value. However, cost, visibility or security could be benefits and therefore could be good goals. (Although, even with visibility I would ask: why does the organisation need more visibility over it’s systems? Hidden costs? Vulnerabilities in shadow IT? Too many system options creating complexity for employees and waste in processes? But I digress)
Diagnose the situation.
This should have included internal and external diagnoses. Let’s say the objective was to reduce cost for the organisation. On paper, moving everything to cloud would actually be more expensive. However, let’s say there was a diganosis that several departments within the organisation were going to increase their computing needs due to a surge in customer sales and a separate data strategy resulting in 1,000s of employees receiving training and needing temporary compute space to practice their analytics. But then they expected this surge to slow down as new analytics users moved to BAU usage levels (low) and sales returned to normal.
Maybe here cloud could be useful because of the temporary surge in computing needs.
Set the guiding principles.
These principles help your team navigate the diagnosed context to reach the goal. Some principles that might have worked in our example:
where the demand is steady and an existing hardware stack exists, keep using on-premises computing.
where demand is fluctuating or it is an emerging tool and we can’t forecast our demand, use cloud computing for flexibility.
These principles would keep the costs down by using cheaper self-managed infrastructure where demand was steady, and more flexible cloud infrastructure that we can switch on and off where demand is highly variable.
Create a coherent action plan.
Based on the goal, diagnosis and principles, the strategy team could have put together a coherent plan for their technology team to execute on, including steps like:
Find or develop an evaluation framework to help decide on cloud use-cases.
Evaluate the top 5 new computing use cases, and top 15 existing systems.
Identify any computing use cases that need to move to cloud.
Execute on the cloud transition for those use cases.
I’m not saying this would have been the right strategy for this organisation. But reflecting on the conversations I had with the technology experts at the time, this would have come a lot closer than the rule of thumb approach.
I’m not in IT, what has this got to do with me?
Now, IT strategy is not the usual topic of this blog. However, I realised this story gives a really excellent illustration of the difference between most strategies and what good strategy actually takes.
Most corporate strategies seem to be either:
A catch-all sentiment, or
A rule of thumb.
And none of these provide enough information to corporate teams to enable them to create and deliver on a coherent plan that moves the organisation in the direction it’s stakeholders want.
If you are responsible for strategy setting, or if you’ve just been handed a fresh new strategy from your leadership team, run the “rule of thumb” test on it. If it sounds like a rule of thumb attempting to pre-empt all the decisions in your department, it is probably not a good strategy.
Same goes if it is a wish or a catch-all sentiment.
What can I do about it?
If you have been handed a bad strategy, don’t worry, you don’t just have to live with it. You can create a sub-strategy for your team or department that delivers on the expectations the leadership have on your team, but provides a lot more guidance than a rule of thumb.
Use the steps I’ve shared here (adapted from Richard Rumelt’s book, Good Strategy Bad Strategy):
Define your goal.
Diagnose your internal and external context.
Set guiding principles.
Create a coherent action plan.
I wish you best of luck battling bad strategy to get great results, good leaders. You’ll be upgrading your team’s work as you do it.
And feel free to share any examples of rule of thumb, wishful thinking or catch-all sentiment strategies you’ve seen out there in the wild.