Insights

It is great to see a different take on cloud cost optimisation in this ITProToday article from last week[1] – the author has some great ideas on how you can amplify the results of your cloud cost optimisation programme by looking beyond the typical rightsizing techniques. 

Don’t get me wrong- there’s nothing wrong with rightsizing. You need to be cognizant of what technology you are using for each of your workloads and whether it’s fit for purpose. It’s a great place to start, you’ll get some good gains, and it starts building the right mentality around spending wisely. 

The problem is that it’s not enough on its own – not if you really want to tackle today’s economic conditions and definitely not if you’re a fast-growing organisation with a commitment to the end-user experience and/or a highly peaky demand profile – it’s a blunt technique that doesn’t really consider these factors. 

Consider this analogy – as we’ve all noticed, food prices at supermarkets in the UK are a bit ridiculous at the moment (all of my recent shopping trips result in a “Oh my gosh £4 for a strawberry”?) 

So, some may have adopted an approach of changing their shopping habits, for example, to use lower-cost supermarkets for the our store-cupboard essentials but higher-end supermarkets for special items (e.g. switch to Lidl for baked beans but splurge on M&S for biscuits). That’s a bit like rightsizing – pick the right option for your workload. It’s going to save you some money, and it’s perfectly ok, but if you’re honest you could probably go a bit further.

The article suggests some bigger and bolder measures – multicloud for example. Again – this is a great idea – split your workloads by need and go to the right provider for each of those needs. So in our shopping example maybe you stick with Lidl for baked beans but you got to Costco for toilet rolls, which you buy more of in terms of quantity. You might need to make a few more trips to get your shopping, but it should be more cost-effective and you’re really taking advantage of the market.

Another great idea from the article is enterprise agreements – which leverage the further discounts that cloud providers would offer due to volume. So in our shopping example, maybe you buy 2 months worth of baked beans so that your bill triggers the £4 off £40 coupon. A great idea and simple to implement. 

But you could be even bolder

What if you knew exactly how many beans you would need, for every day/week/month over the next 3 years? 

What if you could anticipate high-demand events – Sunday brunch or summer barbeques?

What if you could build in the people you’d need to feed – and their demand – the bean lovers vs the bean haters – and you made sure you had the right number of beans for everyone? 

Surely that would be the best way to guarantee you’re not overspending?

What if you could ALSO ensure that every bean was absolutely as satisfying at the more expensive beans that you used to buy? So no one was disappointed?

What if you knew, and factored into your equation, the fact that bean tins were actually going to get slightly bigger year on year, and so actually you’d get more beans for your £ in a year’s time than you do today, so you need to buy less?

Suddenly, you’re not just optimising your shopping bill and catering for peaks but you’re also getting better performance too. 

These are the gains that you get when you dig even deeper, get down to the application layer and understand the profile of your workload – what drives it and how to optimise it. Analysis shows that 80% of your cloud savings are there – with the added benefit that you’ll improve the experience for your customers and be able to scale to meet your peaks in demand too.